Wind turbine abnormality diagnosis device and diagnosis method thereof

Abstract

The present invention discloses a wind turbine abnormality diagnosis device and a diagnosis method thereof, which specifically include the following steps: S1, data collection; S2, data cleaning; S3, data analysis; S4, feature engineering; S5, model establishment; S6, abnormality detection; S7, result interpretation; S8, repeated training; S9, model optimization; The present invention relates to the field of abnormality diagnosis technology. The wind turbine abnormality diagnosis device and the diagnosis method thereof are helpful to optimize the performance of the wind turbine, improve the operating efficiency, and can help prevent potential failures and losses, and can continuously improve and optimize the model to ensure that it has better performance and reliability in practical applications; by remotely connecting several wind turbines to a control center, wind turbines with the same environmental factors can be compared with each other, and some environmental factors can be removed. By mutual detection, update and optimization, the system can find the cause of the failure faster and more accurately.

Description

Technical Field

[0001] The present invention relates to the technical field of abnormality diagnosis, and in particular to a wind turbine abnormality diagnosis device and a diagnosis method thereof. Background Art

[0002] Output power is the main indicator for evaluating the power generation performance of wind turbines. The existing technical difficulty of output power evaluation is relatively small, because there are fewer variables to be analyzed and the statistical methods are relatively simple. However, the difficulty of the actual output power evaluation problem lies in the fact that when the output power is abnormal, it may be caused by many root causes. At this time, experienced engineers need to spend a long time on data analysis to identify the root cause of the output power abnormality, and then solve the problem in a targeted manner based on the locked root cause.

[0003] The existing wind turbine generator output power abnormality diagnosis method still has the following defects:

[0004] Difficult to diagnose accurately: The output power of a wind turbine is affected not only by the equipment itself, but also by external factors such as wind speed, temperature and humidity of the working environment. Therefore, a single diagnostic method may not be able to accurately identify abnormalities.

[0005] Lack of comprehensiveness: Existing diagnostic methods may only diagnose specific abnormalities, lack comprehensive analysis, and are unable to fully identify various possible causes of abnormalities. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a wind turbine abnormality diagnosis device and a diagnosis method thereof, which solves the problem that the existing wind turbine output power abnormality diagnosis method has a single, incomplete and lacks comprehensiveness in diagnosis mode.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: A method for diagnosing abnormality of a wind turbine generator, specifically comprising the following steps:

[0008] S1. Data collection: Collect the output power and related environmental parameters of wind turbines;

[0009] S2, data cleaning: cleaning and preprocessing the collected data;

[0010] S3. Data analysis: Analyze the data using statistical methods;

[0011] S4. Feature engineering: extract and engineer features from the data to find out the characteristic variables related to power output anomalies;

[0012] S5. Model establishment: Establish relevant models, conduct modeling and analysis on power output and related characteristic variables, and find out the causes of abnormalities;

[0013] S6. Anomaly Detection:

[0014] Model training: Use the established model to train the data;

[0015] Anomaly detection: Anomaly detection is performed through the difference between model prediction and actual power output;

[0016] S7. Result interpretation: Use the established model to detect abnormalities in the power output of the wind turbine, find out the time point and cause of the abnormality, and take corresponding measures to repair it according to the analysis results;

[0017] S8, repeat training: repeat the abnormality detection to determine whether the cause is correct. If the abnormality still exists, remove the cause and retrain the detection until the abnormality is eliminated;

[0018] S9, Model optimization: Record each test and add the data to the model to optimize model performance.

[0019] Preferably, the data collected in S1 specifically includes:

[0020] Define power output time series data: P = (P1, P2, ..., P n );

[0021] Collect wind speed data: W = (W1, W2, ..., W n );

[0022] Collect temperature data: T = (T1, T2, ..., T n );

[0023] Collect humidity data: H = (H1, H2, ..., H n ).

[0024] Preferably, the cleaning in S2 includes outlier processing: using statistical methods to detect and process outliers;

[0025] Preprocessing includes missing value handling: using interpolation methods to fill in missing values.

[0026] Preferably, the data analysis in S3 includes:

[0027] Correlation analysis: Calculate the correlation coefficient between power output and wind speed, temperature and humidity;

[0028] Regression analysis: Establish a regression model between power output and environmental parameters;

[0029] Time series analysis: Use time series models to analyze the changing trends of power output.

[0030] Preferably, the S4 feature engineering specifically includes:

[0031] Feature extraction: extracting characteristic variables related to power output anomalies;

[0032] Feature engineering: transform, combine or reduce the features to improve the model performance.

[0033] Preferably, the model building in S5 specifically includes:

[0034] Establish a regression model: P = f(W, T, H) + V, where f() is the regression function and V is the error;

[0035] Build a neural network model: Use a neural network structure to build a nonlinear model of power output.

[0036] The present invention also discloses a wind turbine abnormality diagnosis device, comprising N wind turbine ends and a control center, wherein the N wind turbine ends remotely and wirelessly transmit detection data to the control center for processing, and the control center comprises a wireless receiving module, a data cleaning module, a data analysis module, a data engineering module, a model building module, an abnormality detection module, a result analysis module, a repeated training module and a model optimization module, wherein the output end of the wireless receiving module is connected to the input end of the data cleaning module, the output end of the data cleaning module is connected to the input end of the data analysis module, the output end of the data analysis module is connected to the input end of the data engineering module, the output end of the data engineering module is connected to the input end of the model building module, the output end of the model building module is connected to the input end of the abnormality detection module, the output end of the abnormality detection module is connected to the input end of the result analysis module, the output end of the result analysis module is respectively connected to the input ends of the repeated training module and the model optimization module, the output end of the repeated training module is connected to the input end of the abnormality detection module, and the output end of the model optimization module is connected to the input end of the model building module.

[0037] Preferably, the wind turbine generator end includes a data acquisition unit, a central processing unit and a wireless transmission module, the output end of the data acquisition unit is connected to the input end of the central processing unit, the output end of the central processing unit is connected to the input end of the wireless transmission module, and the wireless transmission module wirelessly sends data to the wireless receiving module.

[0038] Beneficial Effects

[0039] The present invention provides a wind turbine abnormality diagnosis device and a diagnosis method thereof. Compared with the prior art, it has the following beneficial effects:

[0040] 1. The wind turbine abnormality diagnosis device and its diagnosis method can establish a comprehensive data set by collecting the output power and related environmental parameters of the wind turbine, which is helpful to understand the operation of the wind turbine under different conditions. Cleaning and preprocessing data can help remove erroneous, missing or abnormal data to ensure the accuracy and reliability of subsequent analysis and modeling. Analyzing the data using statistical methods can reveal the relationship and rules between the data, providing a basis for subsequent feature engineering and modeling. Through feature extraction and engineering, characteristic variables related to power output abnormality can be found to help understand the root cause of wind turbine abnormality. Establishing relevant models can help understand the complex relationship between power output and related characteristic variables, so as to find out the cause of abnormality and conduct modeling analysis. Through model training and anomaly detection, abnormal power output of wind turbines can be discovered in time, which is helpful to take repair measures in advance and reduce losses. By explaining the time point and cause of the abnormality, it can provide important reference for the maintenance and management of wind turbines and help engineers take corresponding measures for repair and improvement. In general, these steps help optimize the performance of wind turbines, improve operating efficiency, and can help prevent potential failures and losses.

[0041] 2. The wind turbine abnormality diagnosis device and its diagnosis method, by repeatedly performing abnormality detection and determining the cause, can re-detect when the initial diagnosis does not solve the abnormality, automatically remove the error value and loop detection until the final problem is found, and can help identify and solve abnormal conditions in the model to ensure the accuracy and stability of the model. Recording the results of each test and supplementing the data to the model helps to continuously optimize the performance of the model and improve its predictive ability and adaptability. Through these steps, the model can be continuously improved and optimized to ensure better performance and reliability in practical applications.

[0042] 3. The abnormal diagnosis device and the diagnosis method of wind turbines connect several wind turbines to a control center remotely, so that the detected data can be sent together for analysis and processing. It is also convenient to compare and analyze the environmental data of similar wind turbines. Wind turbines with the same environmental factors can be compared with each other, and some environmental factors can be removed. Through mutual detection, update and optimization, the system can find the cause of the fault faster and more accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is the overall system principle block diagram of the present invention;

[0044] Figure 2 This is a principle block diagram of the wind turbine generator end of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions 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 are within the scope of protection of the present invention.

[0046] The present invention discloses a method for diagnosing abnormality of a wind turbine generator, which specifically comprises the following steps:

[0047] S1. Data collection: Collect the output power and related environmental parameters of the wind turbine. The collected data specifically include:

[0048] Define power output time series data: P = (P1, P2, ..., P n );

[0049] Collect wind speed data: W = (W1, W2, ..., W n );

[0050] Collect temperature data: T = (T1, T2, ..., T n );

[0051] Collect humidity data: H = (H1, H2, ..., H n );

[0052] S2. Data cleaning: Clean and preprocess the collected data, where cleaning includes outlier processing: use statistical methods to detect and process outliers;

[0053] Preprocessing includes missing value handling: using interpolation methods to fill missing values;

[0054] S3. Data analysis: Analyze the data using statistical methods. Data analysis includes:

[0055] Correlation analysis: Calculate the correlation coefficient between power output and wind speed, temperature and humidity;

[0056] Regression analysis: Establish a regression model between power output and environmental parameters;

[0057] Time series analysis: Use time series models to analyze the changing trend of power output;

[0058] S4. Feature Engineering: Extract and engineer features from the data to find out the characteristic variables related to power output anomalies, including:

[0059] Feature extraction: extracting characteristic variables related to power output anomalies;

[0060] Feature engineering: transform, combine or reduce the features to improve the model effect;

[0061] S5. Model establishment: Establish relevant models, conduct modeling and analysis on power output and related characteristic variables, and find out the abnormal causes. Model establishment specifically includes:

[0062] Establish a regression model: P = f(W, T, H) + V, where f() is the regression function and V is the error;

[0063] Establishing a neural network model: Using a neural network structure to establish a nonlinear model of power output;

[0064] S6. Anomaly Detection:

[0065] Model training: Use the established model to train the data;

[0066] Anomaly detection: Anomaly detection is performed through the difference between model prediction and actual power output;

[0067] S7. Result interpretation: Use the established model to detect abnormalities in the power output of the wind turbine, find out the time point and cause of the abnormality, and take corresponding measures to repair it according to the analysis results;

[0068] By collecting the output power and related environmental parameters of wind turbines, a comprehensive data set can be established, which helps to understand the operation of wind turbines under different conditions. Cleaning and preprocessing data can help remove erroneous, missing or abnormal data to ensure the accuracy and reliability of subsequent analysis and modeling. Using statistical methods to analyze data can reveal the relationship and rules between data, providing a basis for subsequent feature engineering and modeling. Through feature extraction and engineering, characteristic variables related to abnormal power output can be found to help understand the root cause of wind turbine abnormalities. Establishing relevant models can help understand the complex relationship between power output and related characteristic variables, so as to find out the cause of abnormalities and conduct modeling analysis. Through model training and anomaly detection, abnormalities in wind turbine power output can be discovered in time, which helps to take repair measures in advance and reduce losses. By explaining the time point and cause of the abnormality, it can provide important reference for the maintenance and management of wind turbines, helping engineers take corresponding measures for repair and improvement. In general, these steps help optimize the performance of wind turbines, improve operating efficiency, and can help prevent potential failures and losses.

[0069] S8, repeat training: repeat the abnormality detection to determine whether the cause is correct. If the abnormality still exists, remove the cause and retrain the detection until the abnormality is eliminated;

[0070] S9, Model optimization: Record each test and add the data to the model to optimize model performance.

[0071] By repeatedly performing anomaly detection and determining the cause, it is possible to detect again when the initial diagnosis does not resolve the anomaly, automatically remove erroneous values ​​and repeat the detection until the final problem is found. It can also help identify and resolve anomalies in the model and ensure the accuracy and stability of the model. Recording the results of each test and supplementing the data into the model will help to continuously optimize the performance of the model and improve its predictive ability and adaptability. Through these steps, the model can be continuously improved and optimized to ensure better performance and reliability in practical applications.

[0072] See also Figure 1-2 The present invention also discloses a wind turbine abnormality diagnosis device, comprising N wind turbine terminals and a control center. The N wind turbine terminals remotely and wirelessly transmit detection data to the control center for processing. The control center comprises a wireless receiving module, a data cleaning module, a data analysis module, a data engineering module, a model building module, an abnormality detection module, a result analysis module, a repeated training module and a model optimization module. The output end of the wireless receiving module is connected to the input end of the data cleaning module, the output end of the data cleaning module is connected to the input end of the data analysis module, the output end of the data analysis module is connected to the input end of the data engineering module, the output end of the data engineering module is connected to the input end of the model building module, the output end of the model building module is connected to the input end of the abnormality detection module, the output end of the abnormality detection module is connected to the input end of the result analysis module, the output end of the result analysis module is respectively connected to the input ends of the repeated training module and the model optimization module, the output end of the repeated training module is connected to the input end of the abnormality detection module, and the output end of the model optimization module is connected to the input end of the model building module.

[0073] The wind turbine generator end includes a data acquisition unit, a central processing unit and a wireless transmission module. The data acquisition unit includes a wind speed sensor, a temperature and humidity sensor, a circuit element for monitoring the output power, and a time module. The output end of the data acquisition unit is connected to the input end of the central processing unit, and the output end of the central processing unit is connected to the input end of the wireless transmission module. The wireless transmission module wirelessly sends the data to the wireless receiving module.

[0074] By remotely connecting several wind turbines to a control center, it is convenient to send the detected data together for analysis and processing. It is also convenient to compare and analyze the environmental data of similar wind turbines. Wind turbines with the same environmental factors can be compared with each other, and some environmental factors can be removed. Through mutual detection, update and optimization, the system can find the cause of the fault faster and more accurately.

[0075] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0076] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0077] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for diagnosing abnormality of a wind turbine, characterized in that: The specific steps include: S1. Data collection: Collect the output power and related environmental parameters of wind turbines; S2, data cleaning: cleaning and preprocessing the collected data; S3. Data analysis: Analyze the data using statistical methods; The data analysis in S3 includes: Correlation analysis: Calculate the correlation coefficient between power output and wind speed, temperature and humidity; Regression analysis: Establish a regression model between power output and environmental parameters; Time series analysis: Use time series models to analyze the changing trend of power output; S4. Feature engineering: extract and engineer features from the data to find out the characteristic variables related to power output anomalies; S5. Model establishment: Establish relevant models, conduct modeling and analysis on power output and related characteristic variables, and find out the causes of abnormalities; S6. Anomaly Detection: Model training: Use the established model to train the data; Anomaly detection: Anomaly detection is performed through the difference between model prediction and actual power output; S7. Result interpretation: Use the established model to detect abnormalities in the power output of the wind turbine, find out the time point and cause of the abnormality, and take corresponding measures to repair it according to the analysis results; S8, repeat training: repeat the abnormality detection to determine whether the cause is correct. If the abnormality still exists, remove the cause and retrain the detection until the abnormality is eliminated; S9, Model optimization: Record each test and add the data to the model to optimize model performance.

2. A wind turbine abnormality diagnosis method according to claim 1, characterized in that: The data collected in S1 specifically include: Define power output time series data: P=(P1, P2, ..., P n ); Collect wind speed data: W=(W1, W2, ..., W n ); Collect temperature data: T=(T1, T2, ..., T n ); Collect humidity data: H=(H1, H2, ..., H n ).

3. A wind turbine abnormality diagnosis method according to claim 1, characterized in that: The cleaning in S2 includes outlier processing: using statistical methods to detect and process outliers; Preprocessing includes missing value handling: using interpolation methods to fill in missing values.

4. A wind turbine abnormality diagnosis method according to claim 1, characterized in that: The S4 feature engineering specifically includes: Feature extraction: extracting characteristic variables related to power output anomalies; Feature engineering: transform, combine or reduce the features to improve the model performance.

5. A wind turbine abnormality diagnosis method according to claim 2, characterized in that: The model establishment in S5 specifically includes: Establish a regression model: P=f(W,T,H)+V, where f() is the regression function and V is the error; Build a neural network model: Use a neural network structure to build a nonlinear model of power output.

6. A wind turbine abnormality diagnosis device, used to implement the wind turbine abnormality diagnosis method according to any one of claims 1 to 5, comprising N wind turbine terminals and a control center, wherein the N wind turbine terminals remotely and wirelessly transmit detection data to the control center for processing, characterized in that: The control center includes a wireless receiving module, a data cleaning module, a data analysis module, a data engineering module, a model building module, an anomaly detection module, a result analysis module, a repeated training module and a model optimization module. The output end of the wireless receiving module is connected to the input end of the data cleaning module, the output end of the data cleaning module is connected to the input end of the data analysis module, the output end of the data analysis module is connected to the input end of the data engineering module, the output end of the data engineering module is connected to the input end of the model building module, the output end of the model building module is connected to the input end of the anomaly detection module, the output end of the anomaly detection module is connected to the input end of the result analysis module, the output end of the result analysis module is respectively connected to the input ends of the repeated training module and the model optimization module, the output end of the repeated training module is connected to the input end of the anomaly detection module, and the output end of the model optimization module is connected to the input end of the model building module.

7. A wind turbine abnormality diagnosis device according to claim 6, characterized in that: The wind turbine generator end includes a data acquisition unit, a central processing unit and a wireless transmission module. The output end of the data acquisition unit is connected to the input end of the central processing unit, the output end of the central processing unit is connected to the input end of the wireless transmission module, and the wireless transmission module wirelessly transmits data to the wireless receiving module.

Images