Fault diagnosis method and system for anemometer of wind turbine generator
By dynamically constructing diagnostic communities and integrating environmental data, and employing machine learning models and a dual-threshold judgment mechanism, the accuracy and reliability issues of wind turbine anemometer fault diagnosis have been resolved, enabling precise fault location and operation and maintenance guidance.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- GUANGDONG YUEDIAN ZHUHAI OFFSHORE WIND POWER CO LTD
- Filing Date
- 2026-02-04
- Publication Date
- 2026-05-12
AI Technical Summary
Existing fault diagnosis methods for wind turbine anemometers fail to effectively adapt to real-time wind condition changes, resulting in decreased wind speed correlation, reduced fitting accuracy, and failure to effectively integrate environmental interference with turbine health data. This leads to misjudgment or omission of fault types, increasing maintenance costs and safety hazards.
By dynamically constructing diagnostic communities, calculating wind speed correlation coefficients, integrating environmental and unit health data, using machine learning models for wind speed fitting, and combining a dual-threshold judgment mechanism and sliding window residual features, fault type subdivision and root cause localization are achieved.
It improved the accuracy of wind speed fitting, reduced the misjudgment rate, enabled precise fault type location and root cause analysis, and reduced operation and maintenance costs.
Smart Images
Figure CN122017287A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of fault diagnosis technology, specifically relating to a fault diagnosis method and system for wind turbine anemometers. Background Technology
[0002] Wind speed data is a crucial variable in wind turbines used to evaluate the power characteristics, output performance, and operating status of the entire wind turbine. Currently, wind speed is typically measured by anemometers installed in the wind turbine. The accuracy of the wind speed data measured by the anemometer directly affects the evaluation indicators of the entire wind turbine. Therefore, it is necessary to diagnose faults in the anemometer itself to ensure its stability and reliability.
[0003] Existing anemometer diagnostic methods often employ a fixed number of adjacent wind turbines to form a fault diagnosis cluster, failing to consider the decrease in wind speed correlation between turbines due to real-time wind condition changes, thus reducing fitting accuracy. They rely solely on basic data such as wind speed and active power, without incorporating environmental interference and wind turbine health data, failing to offset the interference of non-wind speed factors on diagnostic results. Furthermore, judging faults based solely on a single instantaneous residual threshold is prone to misjudgment due to instantaneous airflow fluctuations and cannot differentiate fault types or pinpoint root causes. These existing diagnostic methods result in low accuracy and reliability in anemometer fault diagnosis, easily leading to misjudgments or missed diagnoses, increasing maintenance costs, or creating safety hazards. Summary of the Invention
[0004] Based on the aforementioned problems in the existing technology, the purpose of this invention is to provide a fault diagnosis method and system for wind turbine anemometers.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: A fault diagnosis method for a wind turbine anemometer includes: Step S1: Collect the first wind speed data of the wind turbine under test and other wind turbines in the wind farm where it is located, and calculate the wind speed correlation coefficient between the wind turbine under test and other wind turbines. Step S2: Include the wind turbine under test and the adjacent wind turbines whose wind speed correlation coefficient is greater than a preset threshold into the diagnostic community; Step S3: Collect the second wind speed data and active power data of all wind turbines in the diagnostic cluster, and collect environmental data and unit health data. After standardized preprocessing, the data are fused to form a multi-dimensional feature dataset. Step S4: Input the multi-dimensional feature dataset into the pre-trained machine learning model to calculate the wind speed fitting value of the wind turbine under test; Step S5: Calculate the instantaneous residual between the fitted wind speed value and the measured value of the anemometer equipped with the wind turbine under test, and calculate the root mean square of the residual within a preset time window; Step S6: When the instantaneous residual is greater than the first threshold and the root mean square of the residual is greater than the second threshold, it is determined that the anemometer is faulty. Step S7: Extract residual sequence features under fault conditions, determine the fault type based on the preset fault feature library, and match the corresponding fault root cause.
[0006] A further improvement of this invention is that the wind speed correlation coefficient r in step S1 is calculated using the Pearson correlation coefficient, and the calculation formula is as follows: ; In the formula, The wind speed sequence of the wind turbine unit under test. For the other wind turbine wind speed sequences, The mean value of the wind speed sequence of the wind turbine generator under test. The mean of the wind speed sequence of the other wind turbine units.
[0007] A further improvement of the present invention is that step S1 further includes: When the wind direction change of the wind farm exceeds a preset angle and / or the gust intensity exceeds a preset value, the wind speed correlation coefficient is recalculated. Based on the newly calculated wind speed correlation coefficient, the adjacent units that meet the conditions are re-selected to form a new diagnostic community. The adjacent units whose wind speed correlation decreases due to changes in wind direction are removed, and the new adjacent units with high wind speed correlation are included.
[0008] A further improvement of the present invention is that the step S3, which involves standardizing and preprocessing all the above data and then fusing them to form a multi-dimensional feature dataset, includes: The 3σ principle is used to remove outliers from all data; missing data is filled using linear interpolation; when consecutive missing data exceed a preset value, a data acquisition anomaly alarm is triggered. The environmental data in step S3 includes atmospheric temperature data, relative humidity data, and atmospheric pressure data; the unit health data includes nacelle vibration acceleration data, anemometer power supply voltage data, and gearbox oil temperature data.
[0009] A further improvement of the present invention is that the pre-trained machine learning model in step S4 is trained using a random forest regression model, and the model training parameters are optimized by 5-fold cross-validation. The number of decision trees in the random forest regression model is set to 50 to 200.
[0010] A further improvement of the present invention is that step S5 includes: Calculate the instantaneous residual The formula is as follows: ; In the formula, This is the actual measured value from the anemometer. The fitted value for the wind speed; The preset time window is set to a sliding time window of 30–120 minutes, and the root mean square of the residuals within the sliding time window is calculated. The formula is as follows: ; Where N is the number of data points in the sliding time window.
[0011] A further improvement of the present invention is that the first threshold in step S6 is determined based on the 3σ principle of historical normal operation data; the second threshold is the 90% to 95% quantile of the root mean square of the residuals under historical normal conditions.
[0012] A further improvement of the present invention is that the step S7 of extracting residual sequence features under fault conditions includes: The residual sequence under the fault state is decomposed by wavelet transform, and the features of the residual sequence, including the residual mean, are extracted. Residual fluctuation frequency Number of residual mutations ; The fault types include: measurement drift, abnormal fluctuation, partial jamming, and complete jamming.
[0013] A fault diagnosis system for a wind turbine anemometer, comprising: The correlation coefficient calculation module is used to collect the first wind speed data of the wind turbine under test and other wind turbines in the wind farm where it is located, and to calculate the wind speed correlation coefficient between the wind turbine under test and other wind turbines. The diagnostic community establishment module is used to include the wind turbine under test and the adjacent wind turbines whose wind speed correlation coefficient is greater than a preset threshold into the diagnostic community. The dataset acquisition module is used to collect the second wind speed data and active power data of all wind turbines in the diagnostic community, as well as environmental data and unit health data, and after standardized preprocessing, they are fused to form a multi-dimensional feature dataset. The wind speed fitting calculation module is used to input the multi-dimensional feature dataset into a pre-trained machine learning model and calculate the wind speed fitting value of the wind turbine under test. The fault diagnosis module is used to calculate the instantaneous residual between the fitted wind speed value and the measured value of the anemometer equipped with the wind turbine under test, and to calculate the root mean square of the residual within a preset time window. The fault determination module is used to determine that the anemometer is faulty when the instantaneous residual is greater than a first threshold and the root mean square of the residual is greater than a second threshold. The fault type analysis module is used to extract residual sequence features under fault conditions, determine the fault type based on a preset fault feature library, and match the corresponding fault root cause.
[0014] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the fault diagnosis method for the wind turbine anemometer.
[0015] Compared with the prior art, the present invention has at least the following beneficial technical effects: This invention provides a fault diagnosis method and system for wind turbine anemometers. By dynamically constructing a diagnostic cluster, this invention can adapt to real-time wind condition changes, ensuring a high wind speed correlation between the turbines within the cluster and the turbine under test, thus improving wind speed fitting accuracy. By fusing environmental data and turbine health data, it increases data dimensions and effectively counteracts interference from non-wind speed factors on diagnostic results. Employing a dual-threshold judgment mechanism, combined with instantaneous residual and sliding window residual characteristics, it reduces the misjudgment rate caused by instantaneous airflow fluctuations. Furthermore, it enables fault type subdivision and root cause localization, providing precise guidance for operation and maintenance and reducing maintenance costs. Attached Figure Description
[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0017] Figure 1 This is a flowchart illustrating the fault diagnosis method for the wind turbine anemometer according to Embodiment 1 of the present invention. Figure 2 This is a schematic diagram of the module of the wind turbine anemometer fault diagnosis system according to Embodiment 2 of the present invention. Detailed Implementation
[0018] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.
[0019] In the description of this invention, it should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.
[0020] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.
[0021] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.
[0022] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.
[0023] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.
[0024] Example 1 like Figure 1 As shown in the figure, this embodiment provides a fault diagnosis method for a wind turbine anemometer, the fault diagnosis method including: Step S1: Collect the first wind speed data of the wind turbine under test and other wind turbines in the wind farm where it is located, and calculate the wind speed correlation coefficient between the wind turbine under test and other wind turbines.
[0025] In this embodiment, the wind speed correlation coefficient r is calculated using the Pearson correlation coefficient, and the calculation formula is as follows: ; In the formula, This is the wind speed sequence of the wind turbine unit to be tested. For the wind speed sequence of other wind turbine units, The mean value of the wind speed sequence of the wind turbine unit under test. This represents the average wind speed sequence of other wind turbine units.
[0026] As an optional embodiment, step S1 further includes: when the wind direction change of the wind farm exceeds a preset angle or the gust intensity exceeds a preset value, recalculating the wind speed correlation coefficient; based on the newly calculated wind speed correlation coefficient, re-screening eligible neighboring units to form a new diagnostic community, removing neighboring units whose wind speed correlation has decreased due to wind direction change, and including new neighboring units with high wind speed correlation.
[0027] Step S2: Include the wind turbine under test and the adjacent wind turbines whose wind speed correlation coefficient is greater than the preset threshold into the diagnostic community.
[0028] Step S3: Collect the second wind speed data and active power data of all wind turbines in the diagnostic cluster, and collect environmental data and unit health data. After standardizing and preprocessing all the above data, merge them to form a multi-dimensional feature dataset.
[0029] In this embodiment, environmental data includes atmospheric temperature data, relative humidity data, and atmospheric pressure data; unit health data includes nacelle vibration acceleration data, anemometer power supply voltage data, and gearbox oil temperature data.
[0030] As an optional embodiment, step S3, which involves standardizing and preprocessing all the above data before fusing them to form a multi-dimensional feature dataset, includes: Outliers in all data are removed using the 3σ principle; missing data is filled using linear interpolation; and a data acquisition anomaly alarm is triggered when consecutive missing data exceed a preset value.
[0031] Step S4: Input the multi-dimensional feature dataset into the pre-trained machine learning model to calculate the wind speed fitting value of the wind turbine under test.
[0032] As an optional embodiment, the pre-trained machine learning model in step S4 is trained using a random forest regression model. This model is trained on historical data and can predict the wind speed of the wind turbine under test based on the multi-dimensional feature data of the diagnostic community, outputting a fitted wind speed value. Cross-validation is used during model training to optimize parameters and improve fitting accuracy. The number of decision trees in the random forest regression model is set to 50–200.
[0033] Step S5: Calculate the instantaneous residual between the fitted wind speed value and the measured value of the anemometer equipped with the wind turbine under test, and calculate the root mean square of the residual within the preset time window.
[0034] In this embodiment, step S5 includes: calculating the instantaneous residual. The formula is as follows: ; In the formula, This is the actual measured value from the anemometer. The fitted value for the wind speed is denoted as .
[0035] As an optional embodiment, a preset time window is set to a sliding time window of 30–120 minutes, and the root mean square of the residuals within the sliding time window is calculated. The formula is as follows: ; Where N is the number of data points in the sliding time window.
[0036] Step S6: When the instantaneous residual is greater than the first threshold and the root mean square of the residual is greater than the second threshold, the anemometer is determined to be faulty.
[0037] As an optional embodiment, the first threshold in step S6 is determined based on the 3σ principle of historical normal operation data; the second threshold is the 90% to 95% quantile of the root mean square of the residuals under historical normal conditions.
[0038] Step S7: Extract residual sequence features under fault conditions, determine the fault type based on the preset fault feature library, and match the corresponding fault root cause.
[0039] In this embodiment, step S7, which involves extracting residual sequence features under fault conditions, includes: Wavelet transform decomposition is performed on the residual sequence under fault conditions to extract residual sequence features, including the residual mean. Residual fluctuation frequency Number of residual mutations ; As an optional embodiment, the fault types include: measurement drift, abnormal fluctuation, partial jamming, and complete jamming. For example, when the residual mean value is stable within the range of ±1.0 to 2.0 m / s and the fluctuation frequency is below 1 Hz, it is considered a measurement drift fault; when the residual fluctuation frequency is above 5 Hz and the number of abrupt changes is greater than 5 times / minute, it is considered an abnormal fluctuation; when the residual remains stable after abrupt change and the fluctuation frequency drops sharply to below 0.1 Hz, it is considered a partial jamming fault; when the residual is fixed at a certain value (such as 0 or a specific wind speed value) for more than 5 minutes, it is considered a complete jamming fault.
[0040] As an optional embodiment, a fault feature library is constructed to store the feature vectors and root causes corresponding to various fault types, including: measurement drift corresponds to sensor surface contamination and calibration offset; abnormal fluctuation corresponds to poor line contact and electromagnetic interference; partial jamming corresponds to sensor mechanical jamming and bearing wear; complete jamming corresponds to sensor motor failure and power interruption. In this embodiment, the cosine similarity between the current fault feature vector and the feature vector in the fault feature library is calculated, and the fault type and root cause with the highest similarity (≥0.85) is taken as the diagnostic result.
[0041] This embodiment provides a fault diagnosis method for wind turbine anemometers. By dynamically constructing a diagnostic cluster, it can adapt to real-time wind condition changes, ensuring a high wind speed correlation between the units within the cluster and the unit under test, thus improving wind speed fitting accuracy. By fusing environmental data and unit health data, it increases the data dimension and can effectively offset the interference of non-wind speed factors on the diagnostic results. The dual-threshold judgment mechanism, combined with instantaneous residual and sliding window residual characteristics, reduces the misjudgment rate caused by instantaneous airflow fluctuations. It can also achieve fault type subdivision and root cause location, providing precise guidance for operation and maintenance and reducing operation and maintenance costs.
[0042] Example 2 like Figure 2 As shown, this embodiment provides a fault diagnosis system 10 for a wind turbine anemometer. The fault diagnosis system includes: a correlation coefficient calculation module 11, a diagnostic community establishment module 12, a dataset acquisition module 13, a wind speed fitting calculation module 14, a fault judgment module 15, a fault determination module 16, and a fault type analysis module 17.
[0043] The correlation coefficient calculation module 11 is used to collect the first wind speed data of the wind turbine under test and other wind turbines in the wind farm where it is located, and to calculate the wind speed correlation coefficient between the wind turbine under test and other wind turbines.
[0044] The diagnostic community establishment module 12 is used to include the wind turbine under test and adjacent wind turbines whose wind speed correlation coefficient is greater than a preset threshold into the diagnostic community.
[0045] The dataset acquisition module 13 is used to collect the second wind speed data and active power data of all wind turbines in the diagnostic community, as well as environmental data and unit health data. After standardizing and preprocessing all the above data, they are fused to form a multi-dimensional feature dataset.
[0046] The wind speed fitting calculation module 14 is used to input the multi-dimensional feature dataset into the pre-trained machine learning model and calculate the wind speed fitting value of the wind turbine under test.
[0047] The fault diagnosis module 15 is used to calculate the instantaneous residual between the fitted wind speed value and the measured value of the anemometer equipped with the wind turbine under test, and to calculate the root mean square of the residual within a preset time window.
[0048] The fault determination module 16 is used to determine that the anemometer is faulty when the instantaneous residual is greater than the first threshold and the root mean square of the residual is greater than the second threshold.
[0049] The fault type analysis module 17 is used to extract residual sequence features under fault conditions, determine the fault type based on a preset fault feature library, and match the corresponding fault root cause.
[0050] This embodiment provides a fault diagnosis system for wind turbine anemometers. This system is based on the fault diagnosis method for wind turbine anemometers provided in Embodiment 1 above; specific details are not repeated here. This fault diagnosis system dynamically constructs a diagnostic cluster, adapting to real-time wind condition changes and ensuring a high wind speed correlation between the turbines within the cluster and the turbine under test, thus improving wind speed fitting accuracy. By fusing environmental data and turbine health data, it increases data dimensions and effectively counteracts interference from non-wind speed factors on the diagnostic results. It employs a dual-threshold judgment mechanism, combining instantaneous residual and sliding window residual characteristics, reducing the misjudgment rate caused by instantaneous airflow fluctuations. Furthermore, it enables fault type subdivision and root cause localization, providing precise guidance for operation and maintenance and reducing maintenance costs.
[0051] Example 3 This embodiment relates to a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the fault diagnosis method for the wind turbine anemometer of Embodiment 1 described above.
[0052] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.
[0053] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered illustrative and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the scope of the invention. No reference numerals in the claims should be construed as limiting the scope of the claims.
[0054] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that can be understood by those skilled in the art. The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.
Claims
1. A fault diagnosis method for anemometers in wind turbine generators, characterized in that, include: Step S1: Collect the first wind speed data of the wind turbine under test and other wind turbines in the wind farm where it is located, and calculate the wind speed correlation coefficient between the wind turbine under test and other wind turbines. Step S2: Include the wind turbine under test and the adjacent wind turbines whose wind speed correlation coefficient is greater than a preset threshold into the diagnostic community; Step S3: Collect the second wind speed data and active power data of all wind turbines in the diagnostic cluster, and collect environmental data and unit health data. After standardized preprocessing, the data are fused to form a multi-dimensional feature dataset. Step S4: Input the multi-dimensional feature dataset into the pre-trained machine learning model to calculate the wind speed fitting value of the wind turbine under test; Step S5: Calculate the instantaneous residual between the fitted wind speed value and the measured value of the anemometer equipped with the wind turbine under test, and calculate the root mean square of the residual within a preset time window; Step S6: When the instantaneous residual is greater than the first threshold and the root mean square of the residual is greater than the second threshold, it is determined that the anemometer is faulty. Step S7: Extract residual sequence features under fault conditions, determine the fault type based on the preset fault feature library, and match the corresponding fault root cause.
2. The fault diagnosis method for a wind turbine anemometer according to claim 1, characterized in that, The wind speed correlation coefficient r in step S1 is calculated using the Pearson correlation coefficient, and the calculation formula is as follows: ; In the formula, The wind speed sequence of the wind turbine unit under test. For the other wind turbine wind speed sequences, The mean value of the wind speed sequence of the wind turbine generator under test. The mean of the wind speed sequence of the other wind turbine units.
3. The fault diagnosis method for a wind turbine anemometer according to claim 1, characterized in that, Step S1 further includes: When the wind direction change of the wind farm exceeds a preset angle and / or the gust intensity exceeds a preset value, the wind speed correlation coefficient is recalculated; Based on the newly calculated wind speed correlation coefficient, the adjacent units that meet the conditions are re-selected to form a new diagnostic community. The adjacent units whose wind speed correlation decreases due to changes in wind direction are removed, and the new adjacent units with high wind speed correlation are included.
4. The fault diagnosis method for a wind turbine anemometer according to claim 1, characterized in that, The step S3, which involves standardizing and preprocessing all the above data and then fusing them to form a multi-dimensional feature dataset, includes: Outliers in all data are removed using the 3σ principle; missing data are filled using linear interpolation; and a data acquisition anomaly alarm is triggered when consecutive missing data exceed a preset value. The environmental data in step S3 includes atmospheric temperature data, relative humidity data, and atmospheric pressure data; the unit health data includes nacelle vibration acceleration data, anemometer power supply voltage data, and gearbox oil temperature data.
5. A fault diagnosis method for a wind turbine anemometer according to claim 1, characterized in that, The pre-trained machine learning model in step S4 is trained using a random forest regression model, and the model training parameters are optimized through 5-fold cross-validation. The number of decision trees in the random forest regression model is set to 50-200.
6. The fault diagnosis method for a wind turbine anemometer according to claim 1, characterized in that, Step S5 includes: Calculate the instantaneous residual The formula is as follows: ; In the formula, This is the actual measured value from the anemometer. The fitted value for the wind speed; The preset time window is set to a sliding time window of 30–120 minutes, and the root mean square of the residuals within the sliding time window is calculated. The formula is as follows: ; Where N is the number of data points in the sliding time window.
7. The fault diagnosis method for a wind turbine anemometer according to claim 1, characterized in that, The first threshold in step S6 is determined based on the 3σ principle of historical normal operation data; the second threshold is the 90% to 95% quantile of the root mean square of the residuals under historical normal conditions.
8. The fault diagnosis method for a wind turbine anemometer according to claim 1, characterized in that, The step S7, which involves extracting residual sequence features under fault conditions, includes: The residual sequence under the fault state is decomposed by wavelet transform, and the features of the residual sequence, including the residual mean, are extracted. Residual fluctuation frequency Number of residual mutations ; The fault types include: measurement drift, abnormal fluctuation, partial jamming, and complete jamming.
9. A fault diagnosis system for a wind turbine anemometer, characterized in that, include: The correlation coefficient calculation module is used to collect the first wind speed data of the wind turbine under test and other wind turbines in the wind farm where it is located, and to calculate the wind speed correlation coefficient between the wind turbine under test and other wind turbines. The diagnostic community establishment module is used to include the wind turbine under test and the adjacent wind turbines whose wind speed correlation coefficient is greater than a preset threshold into the diagnostic community. The dataset acquisition module is used to collect the second wind speed data and active power data of all wind turbines in the diagnostic community, as well as environmental data and unit health data, and after standardized preprocessing, they are fused to form a multi-dimensional feature dataset. The wind speed fitting calculation module is used to input the multi-dimensional feature dataset into a pre-trained machine learning model and calculate the wind speed fitting value of the wind turbine under test. The fault diagnosis module is used to calculate the instantaneous residual between the fitted wind speed value and the measured value of the anemometer equipped with the wind turbine under test, and to calculate the root mean square of the residual within a preset time window. The fault determination module is used to determine that the anemometer is faulty when the instantaneous residual is greater than a first threshold and the root mean square of the residual is greater than a second threshold. The fault type analysis module is used to extract residual sequence features under fault conditions, determine the fault type based on a preset fault feature library, and match the corresponding fault root cause.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements a fault diagnosis method for a wind turbine anemometer according to any one of claims 1 to 8.