A method and system for real-time diagnosis of wind turbine azimuth sensor in reverse installation

Abstract

This invention discloses a real-time diagnostic method and system for reverse installation of a wind turbine azimuth sensor. The method includes the following steps: S1, acquiring a dataset containing the wind turbine azimuth and power output from the wind turbine azimuth sensor, filtering out valid operating condition data when the wind turbine is in a stable power generation state, and then verifying the validity of the filtered azimuth data to obtain an initial azimuth data sequence; S2, performing angle jump processing on the initial azimuth data sequence obtained in step S1, achieving continuous cyclic data through dynamic compensation of accumulated offset, and then smoothing and optimizing the continuously unfolded angle sequence to obtain a smooth continuous angle sequence; S3, performing least squares linear regression on the smooth continuous angle sequence obtained in step S2 and the corresponding time series index to calculate the slope of the trend line, and determining the installation status of the wind turbine azimuth sensor based on the slope of the trend line. This invention has advantages such as high diagnostic accuracy and strong anti-interference capability.

Description

Technical Field

[0001] This invention mainly relates to the field of wind power generation technology, specifically to a method and system for real-time diagnosis of wind turbine azimuth angle sensors installed in reverse. Background Technology

[0002] The azimuth sensor is a critical component of wind turbine generators, and the angle data it collects directly affects the unit's accuracy in tracking wind direction and its power generation efficiency. In actual operation and maintenance, the azimuth sensor may be installed in reverse due to installation errors, causing abnormal changes in angle data, which in turn can lead to problems such as yaw control failure and power loss.

[0003] Traditional diagnostic methods rely on periodic manual inspections or offline data analysis, which have drawbacks such as high latency, low efficiency, and high cost. Especially in special scenarios such as offshore wind farms, they cannot meet the real-time requirements. There is an urgent need to use algorithms to automatically replace manual inspections to achieve minute-level diagnosis of the installation status.

[0004] The angle data jumps make it difficult to identify anomalies. The azimuth sensor outputs cyclic data from 0° to 360°. In actual operation, there are positive jumps from 360° to 0° when the wind direction changes normally and negative jumps from 0° to 360° when installed in reverse. The existing algorithm does not distinguish between the two types of jumps, which directly leads to errors in the judgment of the angle sequence trend.

[0005] Installation orientation diagnosis faces a contradiction between accuracy and anti-interference. Sensor data is easily affected by wind speed fluctuations, mechanical vibrations, etc. Traditional threshold methods fail to diagnose under low wind speed conditions due to gradual angle changes, and are prone to triggering false diagnoses due to instantaneous jumps under strong wind interference. Summary of the Invention

[0006] To address the technical problems existing in the prior art, this invention provides a method and system for real-time diagnosis of wind turbine azimuth angle sensors with high diagnostic accuracy through reverse installation.

[0007] To solve the above-mentioned technical problems, the technical solution proposed by this invention is as follows:

[0008] A method for real-time diagnostics of a wind turbine azimuth sensor installed in reverse, comprising the following steps:

[0009] S1. Obtain the dataset containing the wind turbine azimuth angle and power output from the wind turbine azimuth angle sensor, filter out the valid operating condition data of the wind turbine in a stable power generation state, and then verify the validity of the filtered azimuth angle data to obtain the initial azimuth angle data sequence.

[0010] S2. Perform angle jump processing on the initial azimuth data sequence obtained in step S1, realize the continuous cyclic data through dynamic compensation of accumulated offset, and then perform smoothing optimization on the continuously unfolded angle sequence to obtain a smooth continuous angle sequence.

[0011] The dynamic compensation for cumulative offset is as follows: taking the first value of the initial azimuth data sequence as the reference, the offset is initialized to zero, and the angle difference between adjacent data points is calculated iteratively; when the angle difference is less than the preset jump threshold, it is determined to be a normal positive jump, and the offset is accumulated by 360°; when the angle difference is greater than the preset jump threshold, it is determined to be an abnormal reverse jump, and the offset is accumulated by 360°; iterative compensation generates a continuously unfolded angle sequence.

[0012] S3. Perform least squares linear regression on the smooth continuous angle sequence obtained in step S2 and the corresponding time series index to calculate the slope of the trend line. Determine the installation status of the wind turbine azimuth sensor based on the slope of the trend line.

[0013] Preferably, the specific process of step S1 is as follows:

[0014] S101. Obtain the dataset containing wind turbine azimuth sensor data, and check whether the dataset contains key data columns for power and azimuth; if the key columns are missing, terminate the process.

[0015] S102. Set a power threshold M based on the rated power of the wind turbine, and filter out valid data whose actual power is greater than the power threshold M, so as to retain valid operating condition data of the wind turbine in power generation state.

[0016] S103. Verify the filtered azimuth data and remove all zero value sequences and abnormal data sequences with outlier ratios exceeding the preset ratio.

[0017] Preferably, in step S103, the proportion of outliers is detected by the interquartile range method; wherein the preset proportion is 20%-40%.

[0018] Preferably, in step S2, a 5-point moving average filter is used to smooth and optimize the continuously unfolded angle sequence.

[0019] Preferably, in step S3, the specific process of determining the installation status of the wind turbine azimuth sensor based on the slope of the trend line is as follows:

[0020] If the slope of the trend line is greater than the first preset value K1, it is determined that "the wind turbine azimuth sensor is installed correctly";

[0021] If the slope of the trend line is less than the second preset value K2, it is determined that "the wind turbine azimuth sensor is installed in reverse".

[0022] Where K1 > K2.

[0023] Preferably, the first preset value and the second preset value are obtained through testing multiple sets of positive and negative samples.

[0024] Preferably, in step S3, result verification is performed simultaneously, including verifying whether the number of data points in the smooth continuous angle sequence meets the minimum sample size requirement, and calculating the confidence interval of the slope of the trend line to determine its trend significance, thereby comprehensively outputting the final diagnostic conclusion.

[0025] The present invention also discloses a computer program product, comprising a computer program that, when executed by a processor, performs the steps of the method described above.

[0026] The present invention further discloses a computer-readable storage medium having a computer program stored thereon, the computer program executing the steps of the method described above when run by a processor.

[0027] The present invention also discloses a real-time diagnostic system for reverse installation of a wind turbine azimuth sensor, comprising a memory and a processor connected to each other, wherein the memory stores a computer program, and the computer program executes the steps of the method described above when run by the processor.

[0028] Compared with the prior art, the advantages of the present invention are as follows:

[0029] The method of this invention effectively filters out invalid and abnormal operating condition data through data acquisition and preprocessing, laying a high-quality data foundation for diagnosis. Furthermore, through angle sequence expansion and smoothing, and an innovative dynamic compensation mechanism for cumulative offset, it accurately resolves the trend breakage problem caused by cyclic jumps from 0° to 360°, clearly distinguishing between normal wind direction changes and reverse installation anomalies. In addition, the moving average filtering effectively suppresses high-frequency interference on site, generating a continuous, stable, and reliable angle sequence. Finally, in trend analysis and installation direction diagnosis, intelligent judgment based on linear regression slope and dynamic threshold, combined with dual verification of sample size and confidence interval, ensures that the diagnostic conclusions have both high accuracy and strong anti-interference ability.

[0030] The method of this invention ultimately shortens the diagnostic cycle from the traditional hours or days to minutes, enabling online real-time monitoring and rapid response to sensor installation status. This not only greatly reduces the reliance on high-cost and high-risk manual labor such as on-site inspections and tower climbing, significantly saving operation and maintenance costs, but also eliminates the risk of turbine yaw control failure, power generation loss, and potential equipment damage and safety accidents caused by sensor data errors. It fundamentally ensures the reliability, stability, and power generation efficiency of wind turbine operation, providing key technical support for the intelligent and lean operation and maintenance of wind farms, especially offshore wind farms with extremely high operation and maintenance difficulties. Attached Figure Description

[0031] Figure 1 This is a flowchart of an embodiment of the real-time diagnostic method of the present invention.

[0032] Figure 2 This is a flowchart of the angle jump processing method of the present invention.

[0033] Figure 3 This is a flowchart of the wind turbine azimuth sensor installation status determination method of the present invention. Detailed Implementation

[0034] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0035] like Figure 1 As shown, the real-time diagnostic method for reverse installation of a wind turbine azimuth sensor provided in this embodiment of the invention includes the following steps:

[0036] S1. Obtain the dataset containing wind turbine azimuth sensor data, check the existence of key columns (such as power and azimuth), filter out valid operating condition data, and perform data cleaning; specifically:

[0037] S101, Key Data Column Integrity Verification: Obtain the dataset containing wind turbine azimuth sensor data, and check whether the dataset contains key data columns for power and azimuth; if the key columns are missing, terminate the process.

[0038] S102. Operating condition data screening and power threshold verification: Based on the rated power of the wind turbine, a power threshold M is set, and valid data with actual power greater than the power threshold M (the wind turbine is in power generation state) are screened to retain valid operating condition data when the wind turbine is in power generation state; data below the power threshold M usually corresponds to the wind turbine shutdown or startup phase, and the azimuth data fluctuates greatly and has no diagnostic value; the above data screening is used to retain rated operating condition data to improve the accuracy of trend judgment.

[0039] S103. Azimuth data validity verification: Verify the azimuth data, including detecting all-zero value sequences and extended verification (such as using the IQR method (interquartile range method) to detect the proportion of outliers; if the proportion of outliers exceeds 30% (adjustable), then mark the data as abnormal) to avoid extreme values ​​interfering with subsequent trend analysis.

[0040] S2. Extract the azimuth data sequence and set the jump threshold; perform angle jump processing on the wind turbine azimuth obtained in step S1, and realize the continuity of cyclic data through cumulative offset dynamic compensation; when the adjacent angle difference is less than the jump threshold, the offset is accumulated by 360°, and when the adjacent angle difference is greater than the jump threshold, the offset is subtracted by 360°. Generate a continuously expanded angle sequence through iterative compensation, and at the same time use moving average filtering to eliminate high-frequency noise to ensure that the angle sequence meets the continuity requirements of trend analysis. As Figure 2 shown, it specifically includes:

[0041] S201. Based on the effective azimuth data sequence obtained in S1, set the jump threshold;

[0042] S202. Cumulative offset dynamic compensation: Initialize the offset to zero and use the first angle data as the reference angle; then traverse and process each angle data one by one to calculate the adjacent angle difference;

[0043] When a positive jump caused by a normal wind direction change in adjacent angles is detected, the sequence trend is continued by accumulating 360° in the offset; when a reverse jump caused by reverse installation is encountered, the sequence direction is corrected by subtracting 360° from the offset to provide an uninterrupted complete data basis for subsequent trend analysis;

[0044] Generate a continuously expanded angle sequence through iterative compensation;

[0045] S203. Smoothing and optimization of the expanded sequence: Apply 5-point moving average filtering (the window size is adjusted according to the sampling frequency) to the expanded angle sequence to eliminate high-frequency jitter and noise of the sensor.

[0046] S3. For the angle sequence obtained in step S2, trend judgment and installation direction diagnosis are based on linear regression slope analysis to realize state recognition. Perform least squares linear regression on the expanded angle sequence and the time axis, and calculate the trend line slope: If the slope > K1, it is determined as "the tilt sensor is installed correctly"; if the slope < K2, it is determined as "the tilt sensor is installed reversely", otherwise "invalid data" is output; where K1 and K2 are both preset values and K1 > K2.

[0047] As Figure 3 shown, it specifically includes:

[0048] S301. Construction of the linear regression model: Based on the continuous angle sequence and the time series index t obtained in step S2, perform least squares linear regression to calculate the trend line slope;

[0049] S302. Dynamic judgment of the slope threshold: Set the slope thresholds K1 and K2 (where K1 and K2 are obtained through more than 200 sets of forward and reverse installation sample tests to ensure a diagnostic accuracy rate > 95% and at the same time tolerate environmental interference fluctuations of ±0.05°);

[0050] If the slope of the trend line > K1, it is determined as "the wind turbine azimuth sensor is installed correctly"; if the slope of the trend line < K2, it is determined as "the wind turbine azimuth sensor is installed reversely"; if the slope of the trend line is between K2 and K1, "invalid data" is output.

[0051] During the process of step S3, it also includes result reliability verification: if the length of the continuous angle sequence < 100 data points, it is directly determined as invalid data to avoid slope calculation deviation caused by too small sample size.

[0052] Slope confidence interval verification: By calculating the slope standard deviation and the 95% confidence interval, if the interval contains 0, it is determined that the trend is not significant.

[0053] The method of the present invention effectively screens out invalid and abnormal working condition data through data acquisition and preprocessing, laying a high-quality data foundation for diagnosis; furthermore, through angle sequence expansion and smoothing processing, and an innovative cumulative offset dynamic compensation mechanism, it accurately solves the problem of trend break caused by the 0° to 360° cyclic jump, clearly distinguishes normal wind direction changes from reverse installation abnormalities, and effectively suppresses on-site high-frequency interference with moving average filtering, generating a continuous, stable and reliable angle sequence; finally, in trend analysis and installation direction diagnosis, based on the intelligent judgment of the linear regression slope and dynamic threshold, and combined with the double verification of sample size and confidence interval, it ensures that the diagnosis conclusion has both high accuracy and strong anti-interference ability.

[0054] The method of the present invention finally shortens the diagnosis cycle from the traditional hour level or day level to the minute level, realizing online real-time monitoring and rapid response to the installation state of the sensor. It not only greatly reduces the high-cost and high-risk manual dependence on on-site patrols and tower climbing inspections, significantly saves the operation and maintenance costs, but also fundamentally eliminates the problems of yaw control failure, power generation loss, potential equipment damage and safety accidents caused by incorrect sensor data, and fundamentally guarantees the reliability, stability and power generation efficiency of the operation of wind turbines, providing key technical support for the intelligent and lean operation and maintenance of wind farms, especially offshore wind farms with extremely high operation and maintenance difficulties.

[0055] The present invention also discloses a computer program product, including a computer program, and the steps of the above-mentioned method are executed when the computer program is run by a processor.

[0056] The present invention further discloses a computer-readable storage medium, on which a computer program is stored, and the steps of the above-mentioned method are executed when the computer program is run by a processor.

[0057] The present invention also discloses a real-time diagnostic system for reverse installation of a wind turbine azimuth sensor, comprising a memory and a processor connected to each other, wherein the memory stores a computer program, and the computer program executes the steps of the method described above when run by the processor.

[0058] The products, media, and systems of the present invention, corresponding to the methods described above, also possess the advantages described above.

[0059] The present invention can implement all or part of the processes in the methods of the above embodiments, or it can be implemented by hardware related to computer program instructions. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, it can implement the steps of the above method embodiments. The computer program includes computer program code, which can be in the form of source code, object code, executable file, or some intermediate form. The computer-readable storage medium includes: any entity or device capable of carrying computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. The memory is used to store computer programs and / or modules. The processor implements various functions by running or executing the computer programs and / or modules stored in the memory, and by calling data stored in the memory. The memory may include high-speed random access memory, as well as non-volatile memory, such as hard disks, RAM, plug-in hard disks, smart media cards (SMC), secure digital (SD) cards, flash cards, at least one disk storage device, flash memory device, or other volatile solid-state storage devices.

[0060] The above are merely preferred embodiments of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should be considered within the scope of protection of the present invention.

Claims

1. A method for real-time diagnosis of a wind turbine azimuth sensor installed in reverse, characterized in that, Including the following steps: S1. Obtain the dataset containing the wind turbine azimuth angle and power output from the wind turbine azimuth angle sensor, filter out the valid operating condition data of the wind turbine in a stable power generation state, and then verify the validity of the filtered azimuth angle data to obtain the initial azimuth angle data sequence. S2. Perform angle jump processing on the initial azimuth data sequence obtained in step S1, realize the continuous cyclic data through dynamic compensation of accumulated offset, and then perform smoothing optimization on the continuously unfolded angle sequence to obtain a smooth continuous angle sequence. The dynamic compensation for cumulative offset is as follows: taking the first value of the initial azimuth data sequence as the reference, the offset is initialized to zero, and the angle difference between adjacent data points is calculated iteratively; when the angle difference is less than the preset jump threshold, it is determined to be a normal positive jump, and the offset is accumulated by 360°; when the angle difference is greater than the preset jump threshold, it is determined to be an abnormal reverse jump, and the offset is accumulated by 360°; iterative compensation generates a continuously unfolded angle sequence. S3. Perform least squares linear regression on the smooth continuous angle sequence obtained in step S2 and the corresponding time series index to calculate the slope of the trend line. Determine the installation status of the wind turbine azimuth sensor based on the slope of the trend line. The specific process of step S1 is as follows: S101. Obtain the dataset containing wind turbine azimuth sensor data, and check whether the dataset contains key data columns for power and azimuth; if the key columns are missing, terminate the process. S102. Set a power threshold M based on the rated power of the wind turbine, and filter out valid data whose actual power is greater than the power threshold M, so as to retain valid operating condition data of the wind turbine in power generation state. S103. Verify the filtered azimuth data and remove all zero value sequences and abnormal data sequences with outlier ratios exceeding the preset ratio. In step S103, the proportion of outliers is detected by the interquartile range method; the preset proportion is 20%-40%.

2. The method for real-time diagnosis of wind turbine azimuth sensor in reverse installation according to claim 1, characterized in that, In step S2, a 5-point moving average filter is used to smooth and optimize the continuously unfolded angle sequence.

3. The method for real-time diagnosis of wind turbine azimuth sensor in reverse installation according to claim 1, characterized in that, In step S3, the specific process of determining the installation status of the wind turbine azimuth sensor based on the slope of the trend line is as follows: If the slope of the trend line is greater than the first preset value K1, it is determined that "the wind turbine azimuth sensor is installed correctly"; If the slope of the trend line is less than the second preset value K2, it is determined that "the wind turbine azimuth sensor is installed in reverse". Where K1 > K2.

4. The method for real-time diagnosis of wind turbine azimuth sensor in reverse installation according to claim 3, characterized in that, The first preset value K1 and the second preset value K2 were obtained through multiple sets of positive and negative sample tests.

5. The method for real-time diagnosis of wind turbine azimuth sensor in reverse installation according to claim 1, 2, or 3, characterized in that, In step S3, the results are verified simultaneously, including verifying whether the number of data points in the smooth continuous angle sequence meets the minimum sample size requirement, and calculating the confidence interval of the slope of the trend line to determine its trend significance, thereby comprehensively outputting the final diagnostic conclusion.

6. A computer program product, comprising a computer program, characterized in that, The computer program is executed by the processor to perform the steps of the method as described in any one of claims 1-5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, The computer program, when run by a processor, performs the steps of the method as described in any one of claims 1-5.

8. A real-time diagnostic system for reverse installation of a wind turbine azimuth sensor, comprising a memory and a processor interconnected, wherein the memory stores a computer program, characterized in that, The computer program, when run by a processor, performs the steps of the method as described in any one of claims 1-5.

Images