Fault detection method and device for fan gearbox, electronic equipment and storage medium
The data of the fan gearbox is preprocessed and trained through the self-encoding model to determine the fault threshold, which solves the problem of low accuracy in the fan gearbox fault detection in the existing technology, and achieves fast and accurate fault judgments, ensuring the safety of the wind turbine. and efficient operation.
Patent Information
- Application Number
- CN202411278538.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-12
- Publication Date
- 2024-12-03
AI Technical Summary
In the prior art, the accuracy of the gearbox fault detection of fan fan is low, which affects the power generation efficiency and production safety of wind turbines.
The self-encoding model is used to preprocess and train the data samples of the gearbox. The fault threshold is smoothly determined through the abnormal score and exponential weighted moving average control chart to determine whether there is a fault in the gearbox.
It improves the accuracy and timeliness of fault detection, can quickly determine the gear box failure, and avoids fan shutdown caused by gear box damage.
Smart Images

Figure CN119063998A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of wind turbine detection, and in particular to a fault detection method, device, electronic equipment and storage medium for a wind turbine gearbox. Background Art
[0002] The gearbox is an important part of the transmission chain of a wind turbine. The health of its gears directly affects the power generation efficiency and production safety of the wind turbine. In the operation of modern large-scale wind turbines, the operating status data of the gearbox is collected online through the SCADA (Supervisory Control And Data Acquisition) system. The operating status data of the gearbox contains the working condition information and fault information of the gearbox. However, the current method of detecting whether the gearbox is faulty based on SCADA data has a low detection accuracy. Summary of the invention
[0003] In view of this, the present invention provides a method, device, electronic equipment and storage medium for detecting a fault of a wind turbine gearbox, so as to solve the problem of low accuracy in detecting a fault of a gearbox.
[0004] In a first aspect, the present invention provides a method for detecting a fault of a wind turbine gearbox, comprising:
[0005] Acquire data samples of the gearbox and preprocess the data samples; the data samples include test set data and training set data;
[0006] Inputting the preprocessed training set data into the autoencoder model for model training, and calculating the abnormality score of the training set data;
[0007] Based on the abnormality score of the training set data, smoothing the abnormality score using an exponentially weighted moving average control chart to determine a fault threshold of the gearbox;
[0008] Inputting the preprocessed test set data into the trained autoencoder model to obtain an abnormality score of the test set data;
[0009] Based on the abnormality score of the test set data and the fault threshold of the gearbox, it is determined whether the gearbox is faulty.
[0010] Beneficial effects: By obtaining the test set data and training set data of the gearbox and preprocessing the test set data and the training set data, the optimization process can be accelerated, and the accuracy of the model can be improved when training the model. The preprocessed training set data is input into the autoencoder model, and the autoencoder model is trained by the training set data. The abnormal score of the training set data can be further calculated by the trained model; then, according to the abnormal score of the training set data, the abnormal score is smoothed using the exponentially weighted moving average control chart to determine the fault threshold of the gearbox; by learning and analyzing the normal training set data, the standard for judging abnormalities is established; the preprocessed test set data is input into the trained autoencoder model, and the abnormal score of the test set data can be obtained. By comparing the abnormal score of the test set data with the fault threshold of the gearbox, when the abnormal score of the test set data is not within the fault threshold range of the gearbox, it can be quickly and accurately judged that the gearbox is faulty, and the staff can be reminded to repair the gearbox in time to avoid the wind turbine shutdown caused by gearbox damage.
[0011] In an optional implementation, the acquiring of the data sample of the gearbox and preprocessing the data sample includes:
[0012] Acquiring data samples of the gearbox based on a data acquisition and monitoring control system;
[0013] Using the quartile method to screen the data acquisition and monitoring control data samples, and removing abnormal values in the data acquisition and monitoring control data samples;
[0014] The data acquisition and monitoring control data samples are normalized after the outliers are removed.
[0015] In an optional implementation, the inputting the preprocessed training set data into the autoencoder model for model training includes:
[0016] Encoding the preprocessed training set data based on the input layer of the autoencoder model;
[0017] Inputting the encoded training set data into the hidden layer of the autoencoder model;
[0018] Decoding and reconstructing the preprocessed training set data in the hidden layer;
[0019] The reconstructed training set data is output based on the output layer of the autoencoder model.
[0020] In an optional implementation, the expression for calculating the abnormality score of the training set data is:
[0021]
[0022] Among them, score i is the abnormal score of the fan failure degree; X i,k is the input training set data, The training set data for the reconstruction output.
[0023] In an optional implementation, the abnormality score based on the training set data is smoothed using an exponentially weighted moving average control chart to determine the fault threshold of the gearbox:
[0024]
[0025] Among them, UCL(t) is the upper threshold limit; LCL(t) is the lower threshold limit; μ is the mean of the training set data score; and ε is the standard deviation of the training set data score; K is a constant; λ is the weight of the historical score; t is the time index.
[0026] In an optional implementation, the preprocessed test set data is input into the trained autoencoder model to obtain an expression for the abnormality score of the test set data:
[0027] Se t =λ·score t +(1-λ)·Se t-1
[0028] Among them, t is the time index; λ is the weight of the historical score; Se t Score for the abnormality of the test set data at time t; score t is the current score of the test set data; Se t-1 Score anomalies for historical test set data.
[0029] In a second aspect, the present invention further provides a fault detection device for a wind turbine gearbox, comprising:
[0030] A preprocessing module, used to obtain data samples of the gearbox and preprocess the data samples; the data samples include test set data and training set data;
[0031] A first calculation module is used to input the preprocessed training set data into the autoencoder model for model training, and calculate the abnormality score of the training set data;
[0032] A second calculation module is used to smooth the abnormality score based on the abnormality score of the training set data using an exponentially weighted moving average control chart to determine a fault threshold of the gearbox;
[0033] A third calculation module is used to input the preprocessed test set data into the trained autoencoder model to obtain an abnormality score of the test set data;
[0034] A judgment module is used to judge whether the gearbox is faulty based on the abnormality score of the test set data and the fault threshold of the gearbox.
[0035] In an optional implementation, the first calculation module includes:
[0036] An encoding unit, used for encoding the preprocessed training set data based on the input layer of the autoencoder model;
[0037] An input unit, used for inputting the encoded training set data into the hidden layer of the autoencoder model;
[0038] A decoding unit, used for decoding and reconstructing the preprocessed training set data in the hidden layer;
[0039] An output unit is used to output the reconstructed training set data based on the output layer of the autoencoder model.
[0040] In a third aspect, the present invention further provides an electronic device, comprising: a memory, a processor, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus;
[0041] The memory is used to store at least one executable instruction, and the executable instruction enables the processor to execute the operation of the above-mentioned wind turbine gearbox fault detection method.
[0042] In a fourth aspect, the present invention also provides a computer-readable storage medium, wherein at least one executable instruction is stored in the storage medium. When the executable instruction is executed on a fault detection device of an electronic device / wind gearbox, the fault detection device of the electronic device / wind gearbox performs the operation of the fault detection method of the wind gearbox as described above.
[0043] The above description is only an overview of the technical solution of the embodiment of the present invention. In order to more clearly understand the technical means of the embodiment of the present invention, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the embodiment of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0044] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0045] Figure 1 It is a flow chart of a fault detection method for a wind turbine gearbox provided by an embodiment of the present invention;
[0046] Figure 2 is a flow chart of a fault detection method for a wind turbine gearbox provided by another embodiment of the present invention;
[0047] Figure 3 It is a structural schematic diagram of a fault detection device for a wind turbine gearbox provided by an embodiment of the present invention;
[0048] Figure 4 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0049] Exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.
[0050] The following describes a specific embodiment of a fault detection method for a wind turbine gearbox according to the present invention. Figure 1 It is a flow chart of a fault detection method for a wind turbine gearbox provided by an embodiment of the present invention. This specification provides method operation steps such as the embodiment or flow chart, but may include more or fewer operation steps based on conventional or non-creative labor. The order of steps listed in the embodiment is only one way of executing the steps among many orders, and does not represent the only order of execution. When the actual system or server product is executed, it can be executed in the order shown in the embodiment or the drawings or in parallel (for example, in a parallel processor or multi-threaded processing environment). Specifically, Figure 1 As shown, the method may include the following steps:
[0051] Step S100, obtaining data samples of the gearbox and preprocessing the data samples; the data samples include test set data and training set data.
[0052] Step S100 specifically includes: obtaining data samples of the gearbox based on the data acquisition and monitoring control system; screening the data acquisition and monitoring control data samples using the quartering method to remove abnormal values in the data acquisition and monitoring control data samples; and normalizing the data acquisition and monitoring control data samples after removing the abnormal values.
[0053] In this embodiment, the data sample includes the generator speed, operating status, wind speed, wind direction, grid active power and grid reactive power data of the wind turbine group; the training set data is the existing historical data sample, and the test set data is the data of the online test. The SCADA data sample is screened by using the four-point method to remove the abnormal values in the SCADA data sample; wherein, the removal of the abnormal values is mainly carried out by wind speed and grid active power. When the wind speed and grid active power are abnormal, the four-point method is used to remove the abnormal wind speed and abnormal grid active power. At the same time, the other parameters of the SCADA data corresponding to the abnormal wind speed and abnormal grid active power at the same time are eliminated together, thereby screening out the abnormal data in the data sample; to ensure that the whole set of data related to the abnormal wind speed and abnormal grid active power are identified and processed. After screening out the abnormal values in the SCADA data sample, all other parameters of the SCADA data are normalized to eliminate the dimension of the data sample, thereby speeding up the optimization process and improving the accuracy of the training model. The abnormal values in the data sample are screened by the four-point method, so that the result of identifying the abnormal values is more objective.
[0054] Step S200: input the preprocessed training set data into the autoencoder model for model training, and calculate the abnormality score of the training set data.
[0055] like Figure 2 As shown, step S200 specifically includes:
[0056] Step S210, encoding the preprocessed training set data based on the input layer of the autoencoder model;
[0057] Step S220, inputting the encoded training set data into the hidden layer of the autoencoder model;
[0058] Step S230, decoding and reconstructing the preprocessed training set data in the hidden layer;
[0059] Step S240: output the reconstructed training set data based on the output layer of the autoencoder model.
[0060] In this embodiment, the autoencoder model includes an input layer, a hidden layer, and an output layer. The preprocessed training set data is received through the input layer of the autoencoder model, and the input layer compresses and abstractly represents the training set data to form encoded data. The encoded data is then transmitted to the hidden layer through the input layer, and the encoded data is decoded and reconstructed in the hidden layer, that is, the hidden layer will restore the approximate form of the original training set data according to the encoded data, so that the reconstructed training set data is as close as possible to the original input training set data; finally, the reconstructed training set data is output through the output layer. The reconstructed training set data output can be compared with the original input training set data, and the difference between the two is calculated to evaluate the learning effect and performance of the autoencoder model, so that the autoencoder model can learn the important features and patterns in the training set data.
[0061] After training, the autoencoder model can learn the patterns and features of normal training set data. When new data is input, the autoencoder model will try to reconstruct the input data. If the input data is normal, the difference between the input data and the reconstructed data will be relatively small; if the input data is abnormal, the difference between the input data and the reconstructed data may be large. By calculating the reconstruction error, the abnormality score of each data point can be obtained.
[0062] The expression for calculating the anomaly score of the training set data is:
[0063]
[0064] Among them, score i is the abnormal score of the fan failure degree; X i,k is the input training set data, The training set data for the reconstruction output.
[0065] Step S300: Based on the abnormality score of the training set data, an exponentially weighted moving average control chart is used to smooth the abnormality score and determine the fault threshold of the gearbox.
[0066] In this embodiment, after obtaining the abnormal score of the training set data, an exponentially weighted moving average (EWMA) control chart is used to smooth the abnormal score. The EWMA control chart performs a weighted average on the historical information of the data, giving a higher weight to recent data, so that the changing trend of the data can be reflected more promptly. Through the EWMA control chart, the abnormal score can be further smoothed, the impact of noise and short-term fluctuations can be reduced, and the outliers in the data can be made more obvious and stable. A fault threshold is determined based on the smoothed results and the actual situation. By using the EWMA control chart to smooth the abnormal score, the changing trend of the abnormal score can be observed more clearly, helping to timely discover potential abnormalities or faults, and providing a basis for further analysis and decision-making.
[0067] The expression for determining the fault threshold of the gearbox is:
[0068]
[0069] Among them, UCL(t) is the upper threshold limit; LCL(t) is the lower threshold limit; μ is the mean of the training set data score; and ε is the standard deviation of the training set data score; K is a constant; λ is the weight of the historical score; t is the time index. Among them, the value of K can be 1.25 times the maximum anomaly score in the training set data.
[0070] Step S400: input the preprocessed test set data into the trained autoencoder model to obtain an abnormality score of the test set data.
[0071] In this embodiment, the expression for obtaining the abnormality score of the test set data is:
[0072] Se t =λ·score t +(1-λ)·Se t-1
[0073] Among them, t is the time index; λ is the weight of the historical score; Se t Score for the abnormality of the test set data at time t; score t is the current score of the test set data; Se t-1 is the anomaly score of the historical test set data. In order to make the anomaly score of the test set data at time t relatively stable, change smoothly, and not easily affected by the drastic impact of the score at a single time point, the value of λ can be set to 0.3. The value of λ can be limited according to specific usage requirements, and can also be 0.2, 0.5, 0.7, etc., without specific limitation.
[0074] The autoencoder model trained with the training set data learns the characteristics and patterns of normal wind turbine operation data. Then, the test data set is input into the trained autoencoder model, which processes and analyzes the input test data and gives an abnormality score for each test data point. The abnormality score reflects the degree of deviation of the test data point from the normal pattern. The higher the score, the more likely the test data point is abnormal. By continuously using this formula, the abnormality score of the test set data at each time point can be gradually calculated based on the output of the autoencoder model and the abnormality score of the historical test set data. This can identify which data points in the test set data may be abnormal, or how abnormal the data is.
[0075] Step S500: judging whether the gearbox is faulty based on the abnormality score of the test set data and the fault threshold of the gearbox.
[0076] In this embodiment, after the anomaly score of the test set data is obtained, the anomaly score of the test set data is compared with the fault threshold of the gearbox, that is, the anomaly score of the test set data is compared with the EWMA control chart; if the anomaly score of the test set data is within the fault threshold range of the gearbox, it can be judged that the operating state of the gearbox is relatively normal and there is no obvious sign of failure; when the anomaly score of the test set data is less than the minimum fault threshold of the gearbox or greater than the maximum fault threshold of the gearbox, it is determined that the gearbox may be faulty or in an abnormal operating state, and the staff can be reminded in time to check the gearbox to avoid damage to the fan due to gearbox failure.
[0077] In the embodiment of the present invention, by obtaining the test set data and training set data of the gearbox, and preprocessing the test set data and the training set data, the optimization process can be accelerated, and the accuracy of the model can be improved when training the model. The preprocessed training set data is input into the autoencoder model, and the autoencoder model is trained by the training set data. The abnormal score of the training set data can be further calculated by the trained model; then, according to the abnormal score of the training set data, the abnormal score is smoothed using the exponentially weighted moving average control chart to determine the fault threshold of the gearbox; by learning and analyzing the normal training set data, a standard for judging abnormalities is established; the preprocessed test set data is input into the trained autoencoder model, and the abnormal score of the test set data can be obtained. By comparing the abnormal score of the test set data with the fault threshold of the gearbox, when the abnormal score of the test set data is not within the fault threshold range of the gearbox, it can be quickly and accurately judged that the gearbox is faulty, and the staff can be reminded to repair the gearbox in time to avoid the wind turbine shutdown caused by gearbox damage.
[0078] The embodiment of the present invention also provides a fault detection device for a wind turbine gearbox. Figure 3 is a schematic diagram of the structure of an embodiment of a fault detection device for a wind turbine gearbox provided by an embodiment of the present invention, such as Figure 3 As shown, the device comprises:
[0079] The preprocessing module 100 is used to obtain data samples of the gearbox and preprocess the data samples; the data samples include test set data and training set data;
[0080] The first calculation module 200 is used to input the preprocessed training set data into the autoencoder model for model training and calculate the abnormality score of the training set data; the expression is:
[0081]
[0082] Among them, score i is the abnormal score of the fan failure degree; Xi,k is the input training set data, Training set data for reconstruction output;
[0083] The second calculation module 300 is used to determine the fault threshold of the gearbox based on the abnormality score of the training set data and smooth the abnormality score using an exponentially weighted moving average control chart; the expression is:
[0084]
[0085] Among them, UCL(t) is the upper threshold limit; LCL(t) is the lower threshold limit; μ is the mean of the training set data score; and ε is the standard deviation of the training set data score; K is a constant; λ is the weight of the historical score; t is the time index;
[0086] The third calculation module 400 is used to input the preprocessed test set data into the trained autoencoder model to obtain the abnormality score of the test set data; the expression is:
[0087] Se t =λ·score t +(1-λ)·Se t-1
[0088] Among them, t is the time index; λ is the weight of the historical score; Se t Score for the abnormality of the test set data at time t; score t is the current score of the test set data; Se t-1 Score anomalies for historical test set data;
[0089] The judgment module 500 is used to judge whether the gearbox is faulty based on the abnormality score of the test set data and the fault threshold of the gearbox.
[0090] In one embodiment, the pre-processing module 100 includes:
[0091] An acquisition unit, used for acquiring data samples of the gearbox based on a data acquisition and monitoring control system;
[0092] A screening unit, used to screen the data acquisition and monitoring control data samples by using a four-part method to remove abnormal values in the data acquisition and monitoring control data samples;
[0093] The processing unit is used to normalize the data collection and monitoring control data samples after removing the abnormal values.
[0094] In one embodiment, the first computing module 200 includes:
[0095] An encoding unit, used for encoding the preprocessed training set data based on the input layer of the autoencoder model;
[0096] An input unit, used to input the encoded training set data into the hidden layer of the autoencoder model;
[0097] A decoding unit, used for decoding and reconstructing the preprocessed training set data in the hidden layer;
[0098] The output unit is used to output the reconstructed training set data based on the output layer of the autoencoder model.
[0099] The device and method embodiments in the embodiments of this application are based on the same application concept.
[0100] Figure 4 The schematic diagram of the structure of an embodiment of an electronic device provided by an embodiment of the present invention is shown. The specific embodiment of the present invention does not limit the specific implementation of the electronic device.
[0101] like Figure 4 As shown, the electronic device may include: a processor (processor) 502 , a communication interface (Communications Interface) 504 , a memory (memory) 506 , and a communication bus 508 .
[0102] Wherein: the processor 502, the communication interface 504, and the memory 506 communicate with each other through the communication bus 508. The communication interface 504 is used to communicate with other devices such as a client or other server network elements. The processor 502 is used to execute the program 510, which can specifically execute the relevant steps in the above-mentioned embodiment of the fault detection method for the wind turbine gearbox.
[0103] Specifically, the program 510 may include program code including computer-executable instructions.
[0104] The processor 502 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention. The one or more processors included in the electronic device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.
[0105] The memory 506 is used to store the program 510. The memory 506 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
[0106] The program 510 may be specifically called by the processor 502 to enable the electronic device to execute the relevant steps in the above-mentioned embodiment of the fault detection method for a wind turbine gearbox.
[0107] It can be understood by those skilled in the art that Figure 4 The structure shown is only for illustration and does not limit the structure of the above-mentioned device. Figure 4 More or fewer components as shown, or with Figure 4 Different configurations shown.
[0108] An embodiment of the present invention provides a computer-readable storage medium, which stores at least one executable instruction. When the executable instruction is executed on a fault detection device of an electronic device / wind gearbox, the fault detection device of the electronic device / wind gearbox executes the fault detection method of the wind gearbox in any of the above-mentioned method embodiments.
[0109] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system or other device. In addition, the embodiments of the present invention are not directed to any particular programming language.
[0110] In the description provided herein, a large number of specific details are described. However, it is understood that embodiments of the present invention can be practiced without these specific details. Similarly, in order to simplify the present invention and help understand one or more of the various inventive aspects, in the above description of exemplary embodiments of the present invention, the various features of the embodiments of the present invention are sometimes grouped together into a single embodiment, figure, or description thereof. Wherein, the claims that follow the specific embodiment are hereby expressly incorporated into the specific embodiment, wherein each claim itself is a separate embodiment of the present invention.
[0111] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and further may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive.
[0112] It should be noted that the above embodiments illustrate the present invention rather than limit it, and that those skilled in the art may devise alternative embodiments without departing from the scope of the appended claims. In the claims, any reference symbol between brackets shall not be construed as a limitation on the claims. The word "comprising" does not exclude the presence of elements or steps not listed in the claims. The word "one" or "an" preceding an element does not exclude the presence of a plurality of such elements. The present invention may be implemented by means of hardware comprising a number of different elements and by means of a suitably programmed computer. In a unit claim enumerating a number of devices, several of these devices may be embodied by the same hardware item. The use of the words first, second, and third, etc. does not indicate any order. These words may be interpreted as names. The steps in the above embodiments, unless otherwise specified, should not be understood as limitations on the order of execution.
Claims
1. A fault detection method for a fan gearbox, It is characterized in that The method comprises: Acquire data samples of the gearbox and preprocess the data samples; the data samples include test set data and training set data; Inputting the preprocessed training set data into the autoencoder model for model training, and calculating the anomaly score of the training set data; Based on the abnormality score of the training set data, smoothing the abnormality score using an exponentially weighted moving average control chart to determine a fault threshold of the gearbox; Inputting the preprocessed test set data into the trained autoencoder model to obtain an abnormality score of the test set data; Based on the abnormality score of the test set data and the fault threshold of the gearbox, it is determined whether the gearbox is faulty.
2. The fault detection method for a wind turbine gearbox according to claim 1, It is characterized in that The step of obtaining a data sample of the gearbox and preprocessing the data sample comprises: Acquiring data samples of the gearbox based on a data acquisition and monitoring control system; Using the quartile method to screen the data acquisition and monitoring control data samples, and removing abnormal values in the data acquisition and monitoring control data samples; The data acquisition and monitoring control data samples are normalized after the outliers are removed.
3. The fault detection method for a wind turbine gearbox according to claim 1, It is characterized in that The step of inputting the preprocessed training set data into the autoencoder model for model training includes: Encoding the preprocessed training set data based on the input layer of the autoencoder model; Inputting the encoded training set data into the hidden layer of the autoencoder model; Decoding and reconstructing the preprocessed training set data in the hidden layer; The reconstructed training set data is output based on the output layer of the autoencoder model.
4. The wind turbine gearbox fault detection method according to claim 1, It is characterized in that The expression for calculating the abnormal score of the training set data is: Among them, score i is the abnormal score of the fan failure degree; X i,k is the input training set data, The training set data for the reconstruction output.
5. The fault detection method for a wind turbine gearbox according to claim 1, It is characterized in that The abnormal score based on the training set data is smoothed using an exponentially weighted moving average control chart to determine the fault threshold of the gearbox: Among them, UCL(t) is the upper threshold limit; LCL(t) is the lower threshold limit; μ is the mean of the training set data score; and ε is the standard deviation of the training set data score; K is a constant; λ is the weight of the historical score; t is the time index.
6. The method for detecting a fault of a wind turbine gearbox according to claim 1, It is characterized in that The preprocessed test set data is input into the trained autoencoder model to obtain the expression of the abnormal score of the test set data: To t λ score t +(1-λ)·Se t-1 Among them, t is the time index; λ is the weight of the historical score; Se t Score for the abnormality of the test set data at time t; score t is the current score of the test set data; Se t-1 Score anomalies for historical test set data.
7. A fault detection device for a fan gearbox, It is characterized in that include: A preprocessing module, used for acquiring data samples of the gearbox and preprocessing the data samples; The data samples include test set data and training set data; A first calculation module is used to input the preprocessed training set data into the autoencoder model for model training, and calculate the abnormality score of the training set data; A second calculation module is used to determine a fault threshold of the gearbox by smoothing the abnormal score based on the abnormal score of the training set data using an exponentially weighted moving average control chart; A third calculation module is used to input the preprocessed test set data into the trained autoencoder model to obtain an abnormality score of the test set data; A judgment module is used to judge whether the gearbox is faulty based on the abnormality score of the test set data and the fault threshold of the gearbox.
8. The fault detection device for a wind turbine gearbox according to claim 7, It is characterized in that The first computing module comprises: An encoding unit, used for encoding the preprocessed training set data based on the input layer of the autoencoder model; An input unit, used for inputting the encoded training set data into the hidden layer of the autoencoder model; A decoding unit, used for decoding and reconstructing the preprocessed training set data in the hidden layer; An output unit is used to output the reconstructed training set data based on the output layer of the autoencoder model.
9. An electronic device, It is characterized in that include: A memory, a processor, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform the operation of the wind turbine gearbox fault detection method according to any one of claims 1-6.
10. A computer-readable storage medium, It is characterized in that The storage medium stores at least one executable instruction, and when the executable instruction is executed on the fault detection device of the electronic device / wind gearbox, the fault detection device of the electronic device / wind gearbox executes the operation of the fault detection method of the wind gearbox as described in any one of claims 1-6.
