Fan blade root load flattening phenomenon judgment and correction method

Abstract

This invention discloses a method for judging and correcting the flattening phenomenon of wind turbine blade root load, comprising the following steps: acquiring real-time blade load data under grid-connected operation of the wind turbine; judging whether the blade root oscillation load data in the real-time blade load data has a flattening phenomenon; judging whether it needs to be merged into a single flattening segment based on the interval between each flattening segment, and calculating the cumulative duration of the flattening segment; judging whether the cumulative duration is not greater than a preset cumulative duration threshold; extracting features of the normal segment located in the previous cycle of each flattening segment from the blade root oscillation load data; generating replacement loads for the flattening segment based on the features and performing segment-by-sine replacement on each flattening segment, while smoothly connecting the start and end points of the flattening segment to complete the correction of the flattening phenomenon of wind turbine blade root load. This invention corrects flattening through feature extraction and sinusoidal load generation, meeting the load testing accuracy requirements, avoiding long-term wind turbine shutdowns, and reducing power generation losses.

Description

Technical Field

[0001] This invention relates to the technical field of wind turbine load testing, and in particular to a method for judging and correcting the phenomenon of flattening load at the wind turbine blade root. Background Technology

[0002] Wind turbines are large and complex mechanical components. To monitor and evaluate the structural health of wind turbines under wind loads and operating loads, and to ensure the safety, reliability, and performance of the system, mechanical load testing of the turbine blades is necessary. This identifies potential structural problems, fatigue cracks, or strain concentration areas, thereby preventing equipment failure or damage. Furthermore, analysis based on measured data can improve system reliability, reduce downtime, and lower maintenance costs.

[0003] Blade root load testing typically employs sensors such as resistance strain gauges or fiber optic strain gauges. These sensors are installed at the roots of three blades, and the load signals are connected to a type test data acquisition system. After sensor installation, signal calibration is required to determine the conversion relationship between electrical signals and load values. For blade oscillation load calibration, in light wind conditions below 5 m / s, the blades are opened to 0 degrees and allowed to freely rotate at low speed for at least 3 revolutions. The peak and trough values ​​are obtained from the waveform. Combined with the distance between the strain gauge and the blade root, the blade weight, and the blade's center of mass position, the slope and offset values ​​can be calculated. For flapping load calibration in light wind conditions, the blades are opened to 90 degrees and allowed to freely rotate at low speed for at least 3 revolutions. The peak and trough values ​​are obtained from the waveform. Combined with the distance between the strain gauge and the blade root, the blade weight, and the blade's center of mass position, the slope and offset values ​​can be calculated. Based on the above-calibrated correspondence, the acquired electrical signal values ​​can be converted into load values. Then, by comparing theoretical and simulated values, it can be confirmed whether the wind turbine is operating safely under different wind conditions.

[0004] However, it is well known in this field that the blade root load testing cycle is as long as 3-6 months. Affected by ambient temperature and internal interference, the sensor signal is prone to partial flattening, leading to inaccurate load calculations and invalid data. This, in turn, delays the testing cycle, increases equipment maintenance and labor costs, and a single hub maintenance requires 4-6 hours, resulting in a loss of 5-15 MWh of power generation. Therefore, there is an urgent need for a method that can accurately correct flattened signals without downtime, ensuring that the test data is accurate and usable. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and propose a method for judging and correcting the phenomenon of flattening load at the blade root of a wind turbine. It aims to solve the following technical problems: 1) Flattening of the blade root oscillation signal leads to inaccurate load calculation, affecting the judgment of unit safety; 2) Sensor maintenance requires shutdown and entry into the hub, which is difficult to operate and results in a loss of power generation; 3) Invalid data leads to a longer testing cycle and increased equipment and labor costs; 4) Existing correction methods lack quantitative standards, and the correction effect cannot be verified.

[0006] The objective of this invention is achieved through the following technical solution: a method for judging and correcting the phenomenon of flattening load at the blade root of a wind turbine, comprising the following steps:

[0007] S1. Obtain real-time blade load data of the wind turbine under grid-connected operation status;

[0008] S2. Set the sampling frequency. After collecting real-time blade load data for a preset period, calculate the absolute value of the first-order difference of the blade root oscillation load data in the real-time blade load data. When several or more consecutive absolute values ​​of the first-order difference are less than the preset threshold, it is determined that the blade root oscillation load data has flattened. Record the start index and end index of each flattened segment in the blade root oscillation load data and proceed to step S3; otherwise, it is determined that no flattening has occurred and the operation is exited.

[0009] S3. Calculate the duration of each leveling segment based on the start index and end index. Determine whether the segments need to be merged into one leveling segment based on the interval duration between them. If they need to be merged into one leveling segment, calculate the cumulative duration of the merged leveling segment. If they do not need to be merged into one leveling segment, calculate the cumulative duration of each leveling segment separately.

[0010] S4. Determine whether the cumulative duration obtained in step S3 is not greater than the preset cumulative duration threshold; if the cumulative duration is not greater than the preset cumulative duration threshold, proceed to step S5; if the cumulative duration is greater than the preset cumulative duration threshold, exit the operation.

[0011] S5. Extract the features of the normal segment located in the previous cycle of each flattening segment from the blade root oscillation load data. The features include the impeller rotation cycle, blade root oscillation amplitude, and initial phase.

[0012] S6. Based on the features obtained in step S5, generate the correction load for the flattening segment based on the sinusoidal characteristics and correct each flattening segment segment by segment. At the same time, smoothly connect the start and end points of the flattening segment to complete the correction of the flattening phenomenon of the wind turbine blade root load.

[0013] Furthermore, step S2 includes:

[0014] Set the sampling frequency. After collecting blade load data for a preset period, calculate the absolute value of the first-order difference of the blade root oscillation load data in that blade load data. At the same time, acquire blade root oscillation load data without flattening phenomenon, calculate the standard deviation of the absolute value of the first-order difference of the blade root oscillation load data without flattening phenomenon, and set a threshold of less than 3 times the standard deviation. When several or more consecutive absolute values ​​of the first-order difference are less than the threshold, it is determined that the blade root oscillation load data has flattened phenomenon. Record the start index and end index of each flattening segment in the blade root oscillation load data, and proceed to step S3; otherwise, it is determined that no flattening phenomenon has occurred, and exit the operation.

[0015] Furthermore, step S3 includes:

[0016] For each flattening segment in the blade root oscillation load data, the number of data points in the flattening segment is calculated by subtracting the start index from the end index, thus obtaining the duration of each flattening segment. Further, the interval duration between two flattening segments is calculated. If the interval duration between two flattening segments is not greater than a preset interval duration threshold, it is determined that the two flattening segments are caused by the same flattening phenomenon, and the two flattening segments are merged into one flattening segment. The start index of the former flattening segment and the end index of the latter flattening segment are taken to calculate the cumulative duration of the merged flattening segment. If it is not necessary to merge them into one flattening segment, the cumulative duration of each flattening segment is calculated separately.

[0017] Furthermore, the threshold for the interval duration is 0.1s.

[0018] Furthermore, in step S4, the cumulative duration threshold is 10 seconds.

[0019] Furthermore, step S5 includes:

[0020] The impeller rotation cycle is calculated by analyzing the azimuth angle. If the azimuth angle completes a 360° change, it is considered one cycle. The impeller rotation cycle is determined by finding the interval in the azimuth angle data that increases from 0° to 360°. ,in This represents the number of data points in the azimuth data ranging from 0° to 360°. The period of the sampling frequency;

[0021] The amplitude A of the blade root oscillation is estimated by calculating half the difference between the maximum and minimum values ​​in the normal segment of the blade root oscillation load data. mean , The load data for the normal section in the previous cycle before the leveling section;

[0022] The initial phase is deduced by back-calculating the normal segment data from the previous data point of the flattened segment, ensuring that the starting waveform of the repaired flattened segment is completely continuous with the original signal, and avoiding abrupt transitions.

[0023] Furthermore, step S6 includes:

[0024] S6.1 For each flattening segment, calculate the load on the flattening segment:

[0025]

[0026] In the formula To flatten the first section Corrected load for each data point; S is the starting index of the flattening segment; To flatten the relative time within the segment; initial phase Load from the previous data point of the flattening section By reverse reasoning ; This is the first data point in the flattened segment;

[0027] S6.2 Perform the pre-connection, for the first data point of the flattening segment. Correction value for the first data point in the flattening segment ,use Correct the first data point of the flattening segment This allows it to transition to the normal segment of the previous cycle;

[0028] S6.3. Perform post-connection, for the last data point of the flattening segment. Correction value for the last data point in the flattening segment ,use Correct the last data point of the flattening segment , This represents the load value of the first data point in the normal section after the flattening section ends.

[0029] S6.4. For the middle part of the flattened section, all of it shall be done using... Correction;

[0030] S6.5 Repeat steps S6.1 to S6.4 until all flattened sections are corrected.

[0031] A system for judging and correcting the load flattening phenomenon at the wind turbine blade root, used to implement the above-mentioned method for judging and correcting the load flattening phenomenon at the wind turbine blade root, includes:

[0032] The load data acquisition module is used to acquire real-time blade load data under the grid-connected operation status of the wind turbine and input it into the flattening phenomenon judgment module;

[0033] The flattening phenomenon judgment module is used to determine whether the blade root oscillation load data in the real-time blade load data has experienced flattening. After collecting real-time blade load data for a preset period, it calculates the absolute value of the first-order difference of the blade root oscillation load data in the real-time blade load data. When several or more consecutive absolute values ​​of the first-order difference are less than a preset threshold, it is determined that the blade root oscillation load data has experienced flattening. The start index and end index of each flattening segment in the blade root oscillation load data are recorded and input into the flattening segment merging module. Otherwise, it is determined that no flattening phenomenon has occurred and the system exits.

[0034] The segment merging module calculates the duration of each segment based on the start and end indices, and determines whether to merge them into one segment based on the interval duration between each segment. If they need to be merged into one segment, the module calculates the cumulative duration of the merged segment. If they do not need to be merged into one segment, the module calculates the cumulative duration of each segment separately.

[0035] The cumulative duration judgment module determines whether the cumulative duration calculated by the flattening segment merging module is not greater than the preset cumulative duration threshold; if the cumulative duration is not greater than the preset cumulative duration threshold, it proceeds to the feature extraction module; if the cumulative duration is greater than the preset cumulative duration threshold, it exits the system.

[0036] The feature extraction module is used to extract features of the normal segment located in the previous cycle of each flattening segment from the blade root oscillation load data. The features include the impeller rotation cycle, blade root oscillation amplitude, and initial phase.

[0037] The flattening correction module generates the correction load for the flattening segment based on the features obtained by the feature extraction module and the sinusoidal characteristics, and corrects each flattening segment segment by segment. At the same time, it smoothly connects the start and end points of the flattening segment to complete the correction of the flattening phenomenon of the wind turbine blade root load.

[0038] A non-transitory computer-readable medium storing instructions that, when executed by a processor, perform the steps of the method for judging and correcting the flattening phenomenon of wind turbine blade root load as described above.

[0039] A computing device includes a processor and a memory for storing processor-executable programs. When the processor executes the program stored in the memory, it implements the above-mentioned method for judging and correcting the flattening phenomenon of wind turbine blade root load.

[0040] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0041] 1. High correction accuracy: Through quantitative feature extraction and sinusoidal load generation, the correction results in small deviations in blade root oscillation amplitude and impeller rotation period, and small transition jumps, meeting the accuracy requirements of load testing.

[0042] 2. No need to shut down for maintenance: When the flattening phenomenon occurs, there is no need to shut down the wind turbine and then enter its hub for maintenance sensors. A single correction of the flattening phenomenon can avoid long-term shutdown, reduce power generation loss, and save operation and maintenance costs per wind turbine per year.

[0043] 3. Shortened testing cycle: Effectively avoids retesting due to invalid data, shortens the testing cycle, and reduces equipment rental and labor costs;

[0044] 4. High reproducibility: The parameters, formulas and steps for judging and correcting the flattening phenomenon in this invention are all quantified, and those skilled in the art can directly reproduce it. Attached Figure Description

[0045] Figure 1 This is a side view of the wind turbine structure.

[0046] Figure 2 This is a front view of the wind turbine blade.

[0047] Figure 3 A schematic diagram of the blade root oscillation load waveform before correction of the flattening phenomenon.

[0048] Figure 4 A schematic diagram of the blade root oscillation load waveform after correction for the flattening phenomenon.

[0049] Figure 5 A flowchart for judging and correcting the phenomenon of flattening load at the root of wind turbine blades. Detailed Implementation

[0050] The present invention will be further described below with reference to specific embodiments.

[0051] Example 1

[0052] See Figures 1 to 2 As shown, the wind turbine consists of blades 1, nacelle 2 and tower 3. Blades 1 rotate under the action of wind force, and the load at the root of the three blades 1 will fluctuate with the wind speed.

[0053] Four resistance strain gauges 5 are installed at the root of each of the three blades 1, and then connected to the hub data acquisition device 6 for data acquisition. The data is then transmitted through the fiber optic cable 4, nacelle 2 and tower 3 to the wind turbine data acquisition system 7 for storage, and finally summarized to the industrial control computer 8 for analysis and processing.

[0054] After the strain gauge 5 is installed, the three blades 1 are calibrated at 0° and 90° idle rotation in light winds below 5m / s. After calibration, the calculation relationship between the electrical signal and the load can be obtained. The calculation relationship is written into the industrial control computer 8 for all load data processing.

[0055] See Figure 5As shown, the method for judging and correcting the flattening phenomenon of wind turbine blade root load provided in this embodiment includes the following steps:

[0056] S1. First, it is necessary to determine whether the acquisition devices, such as the resistance strain gauge 5, the hub acquisition device 6, and the wind turbine data acquisition system 7, are normal. If the acquisition devices are all normal, the industrial control computer 8 will obtain the real-time blade load data of the wind turbine in grid-connected operation from the wind turbine data acquisition system 7. If the acquisition devices are not normal, the operation will be terminated.

[0057] S2. Set the sampling frequency to 50Hz, i.e., 50 data points per second. After collecting blade load data for a period of 10 minutes, under the action of wind load, the synthesized blade wobbling load data will form a load signal similar to a sine wave. See [link / reference]. Figure 3 As shown; the azimuth angle of the blade tumbling load data is Calculate the blade root oscillation load data. The continuous first-order difference absolute value; simultaneously, acquire blade root oscillation load data without flattening phenomenon, calculate the standard deviation of the continuous first-order difference absolute value of the blade root oscillation load data without flattening phenomenon, and set a threshold of less than 3 times the standard deviation. In this embodiment, the threshold is set to 0.001; determine ( If the absolute values ​​of five or more consecutive first-order differences are less than 0.001, it is determined that the blade root oscillation load data has flattened. The start index S and end index E of each flattened segment in the blade root oscillation load data are recorded, and the process proceeds to step S3. If less than five consecutive absolute values ​​of first-order differences are less than 0.001, or if five or more consecutive absolute values ​​of first-order differences are greater than 0.001, it is determined that no flattening has occurred, and the operation is exited. See also Figure 3 Detected and The starting index S and ending index E of the flattening segment are respectively. The flattening time is 2.24 seconds, and flattening repair is required.

[0058] S3. For each flattening segment in the blade root oscillation load data, the number of data points in the flattening segment is calculated by subtracting the starting index S and the ending index E, and then the duration of each flattening segment is obtained. Further, the interval duration between two flattening segments is calculated. If the interval duration between two flattening segments is not greater than 0.1s, that is, the interval data points between two flattening segments are not greater than 5, then it is determined that the two flattening segments are caused by the same flattening phenomenon. The two flattening segments are merged into one flattening segment. The starting index of the previous flattening segment and the ending index of the next flattening segment are taken to calculate the cumulative duration of the merged flattening segment. If it is not necessary to merge them into one flattening segment, the cumulative duration of each flattening segment is calculated separately.

[0059] S4. Determine if the cumulative duration obtained in step S3 is not greater than 10s, i.e., 500 data points; if the cumulative duration is not greater than 10s, proceed to step S5; if the cumulative duration is greater than 10s, exit the operation.

[0060] S5. Extract the features of the normal segment located in the previous cycle of each flattening segment from the blade root oscillation load data. The features include the impeller rotation cycle, blade root oscillation amplitude, and initial phase.

[0061] 1) Determine the cycle From the normal data of the first complete impeller cycle before the flattening section, find the number of data points in the azimuth angle interval from 0° to 360°. ,cycle The sampling frequency is 50Hz.

[0062] 2) Calculate the amplitude with the mean : , In the formula, This is the load data for the normal section of the previous cycle before the leveling section.

[0063] 3) Lock the phase connection point: Record the starting index of the flattened segment. Load value of the first data point With corresponding azimuth angle Ensure the initial load of the repair section is consistent with... Connection.

[0064] S6. Based on the features obtained in step S5, generate the corrected load for the flattening segment based on the sinusoidal characteristics and correct each flattening segment segment by segment. At the same time, smoothly connect the start and end points of the flattening segment to complete the correction of the flattening phenomenon of the wind turbine blade root load. This includes the following steps:

[0065] S6.1 For each flattening segment, calculate the load on the flattening segment:

[0066]

[0067] In the formula To flatten the first section Corrected load for each data point; S is the starting index of the flattening segment; To flatten the relative time within the segment; initial phase Load from the previous data point of the flattening section By reverse reasoning ; This is the first data point in the flattened segment;

[0068] S6.2 Perform the pre-connection, for the first data point of the flattening segment. Correction value for the first data point in the flattening segment ,use Correct the first data point of the flattening segment This allows it to transition to the normal segment of the previous cycle;

[0069] S6.3. Perform post-connection, for the last data point of the flattening segment. Correction value for the last data point in the flattening segment ,use Correct the last data point of the flattening segment , This represents the load value of the first data point in the normal section after the flattening section ends.

[0070] S6.4. For the middle part of the flattened section, all of it shall be done using... Correction;

[0071] S6.5. Repeat steps S6.1 to S6.4 until all flattened sections are corrected. See the corrected blade root oscillation load waveform. Figure 4 As shown.

[0072] Example 2

[0073] This embodiment provides a system for judging and correcting the flattening phenomenon of wind turbine blade root load, used to implement the method for judging and correcting the flattening phenomenon of wind turbine blade root load described in Embodiment 1, including:

[0074] The load data acquisition module is used to acquire blade load data under the grid-connected operation state of the wind turbine and input it into the flattening phenomenon judgment module;

[0075] The flattening phenomenon judgment module is used to determine whether the blade root oscillation load data in the real-time blade load data has experienced flattening. After collecting real-time blade load data for a preset period, it calculates the absolute value of the first-order difference of the blade root oscillation load data in the real-time blade load data. When several or more consecutive absolute values ​​of the first-order difference are less than a preset threshold, it is determined that the blade root oscillation load data has experienced flattening. The start index and end index of each flattening segment in the blade root oscillation load data are recorded and input into the flattening segment merging module. Otherwise, it is determined that no flattening phenomenon has occurred and the system exits.

[0076] The segment merging module calculates the duration of each segment based on the start and end indices, and determines whether to merge them into one segment based on the interval duration between each segment. If they need to be merged into one segment, the module calculates the cumulative duration of the merged segment. If they do not need to be merged into one segment, the module calculates the cumulative duration of each segment separately.

[0077] The cumulative duration judgment module determines whether the cumulative duration calculated by the flattening segment merging module is not greater than the preset cumulative duration threshold; if the cumulative duration is not greater than the preset cumulative duration threshold, it proceeds to the feature extraction module; if the cumulative duration is greater than the preset cumulative duration threshold, it exits the system.

[0078] The feature extraction module is used to extract features of the normal segment located in the previous cycle of each flattening segment from the blade root oscillation load data. The features include the impeller rotation cycle, blade root oscillation amplitude, and initial phase.

[0079] The flattening correction module generates the correction load for the flattening segment based on the features obtained by the feature extraction module and the sinusoidal characteristics, and corrects each flattening segment segment by segment. At the same time, it smoothly connects the start and end points of the flattening segment to complete the correction of the flattening phenomenon of the wind turbine blade root load.

[0080] Example 3

[0081] This embodiment discloses a non-transitory computer-readable medium storing instructions, which, when executed by a processor, perform the steps of the wind turbine blade root load flattening phenomenon judgment and correction method according to Embodiment 1.

[0082] In this embodiment, the non-transitory computer-readable medium can be a disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), USB flash drive, portable hard drive, etc.

[0083] Example 4

[0084] This embodiment discloses a computing device, including a processor and a memory for storing processor-executable programs. When the processor executes the program stored in the memory, it implements the method for judging and correcting the flattening phenomenon of wind turbine blade root load as described in Embodiment 1.

[0085] The computing device described in this embodiment may be a desktop computer, laptop computer, smartphone, PDA handheld terminal, tablet computer, programmable logic controller (PLC), or other terminal device with processor function.

[0086] The above-described embodiments are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Therefore, any changes made in accordance with the shape and principle of the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for judging and correcting the phenomenon of flattening load at the root of a wind turbine blade, characterized in that, Includes the following steps: S1. Obtain real-time blade load data of the wind turbine under grid-connected operation status; S2. Set the sampling frequency. After collecting real-time blade load data for a preset period, calculate the absolute value of the first-order difference of the blade root oscillation load data in the real-time blade load data. When several or more consecutive absolute values ​​of the first-order difference are less than the preset threshold, it is determined that the blade root oscillation load data has flattened. Record the start index and end index of each flattened segment in the blade root oscillation load data and proceed to step S3; otherwise, it is determined that no flattening has occurred and the operation is exited. S3. Calculate the duration of each leveling segment based on the start index and end index, and determine whether it needs to be merged into one leveling segment based on the interval duration between each leveling segment; if it needs to be merged into one leveling segment, calculate the cumulative duration of the merged leveling segment. If there is no need to merge them into one leveling segment, then calculate the cumulative duration of each leveling segment separately; S4. Determine whether the cumulative duration obtained in step S3 is not greater than the preset cumulative duration threshold; if the cumulative duration is not greater than the preset cumulative duration threshold, proceed to step S5; If the cumulative duration exceeds the preset cumulative duration threshold, the operation will be terminated. S5. Extract the features of the normal segment located in the previous cycle of each flattening segment from the blade root oscillation load data. The features include the impeller rotation cycle, blade root oscillation amplitude, and initial phase. S6. Based on the features obtained in step S5, generate the correction load for the flattening segment based on the sinusoidal characteristics and correct each flattening segment segment by segment. At the same time, smoothly connect the start and end points of the flattening segment to complete the correction of the flattening phenomenon of the wind turbine blade root load.

2. The method for judging and correcting the flattening phenomenon of wind turbine blade root load according to claim 1, characterized in that, Step S2 includes: Set the sampling frequency. After collecting blade load data for a preset period, calculate the absolute value of the first-order difference of the blade root oscillation load data in that blade load data. At the same time, acquire blade root oscillation load data without flattening phenomenon, calculate the standard deviation of the absolute value of the first-order difference of the blade root oscillation load data without flattening phenomenon, and set a threshold of less than 3 times the standard deviation. When several or more consecutive absolute values ​​of the first-order difference are less than the threshold, it is determined that the blade root oscillation load data has flattened phenomenon. Record the start index and end index of each flattening segment in the blade root oscillation load data, and proceed to step S3; otherwise, it is determined that no flattening phenomenon has occurred, and exit the operation.

3. The method for judging and correcting the flattening phenomenon of wind turbine blade root load according to claim 1, characterized in that, Step S3 includes: For each flattening segment in the blade root oscillation load data, the number of data points in the flattening segment is calculated by subtracting the start index from the end index, thus obtaining the duration of each flattening segment. Further, the interval duration between two flattening segments is calculated. If the interval duration between two flattening segments is not greater than a preset interval duration threshold, it is determined that the two flattening segments are caused by the same flattening phenomenon, and the two flattening segments are merged into one flattening segment. The start index of the former flattening segment and the end index of the latter flattening segment are taken to calculate the cumulative duration of the merged flattening segment. If it is not necessary to merge them into one flattening segment, the cumulative duration of each flattening segment is calculated separately.

4. The method for judging and correcting the flattening phenomenon of wind turbine blade root load according to claim 3, characterized in that, The interval duration threshold is 0.1s.

5. The method for judging and correcting the flattening phenomenon of wind turbine blade root load according to claim 1, characterized in that, In step S4, the cumulative duration threshold is 10 seconds.

6. The method for judging and correcting the flattening phenomenon of wind turbine blade root load according to claim 1, characterized in that, Step S5 includes: The impeller rotation cycle is calculated by analyzing the azimuth angle. If the azimuth angle completes a 360° change, it is considered one cycle. The impeller rotation cycle is determined by finding the interval in the azimuth angle data that increases from 0° to 360°. ,in This represents the number of data points in the azimuth data ranging from 0° to 360°. The period of the sampling frequency; The amplitude A of the blade root oscillation is estimated by calculating half the difference between the maximum and minimum values ​​in the normal segment of the blade root oscillation load data. mean , The load data for the normal section in the previous cycle before the leveling section; The initial phase is deduced by back-calculating the normal segment data from the previous data point of the flattened segment, ensuring that the starting waveform of the repaired flattened segment is completely continuous with the original signal, and avoiding abrupt transitions.

7. The method for judging and correcting the flattening phenomenon of wind turbine blade root load according to claim 6, characterized in that, Step S6 includes: S6.1 For each flattening segment, calculate the load on the flattening segment: ； In the formula To flatten the first section Corrected load for each data point; S is the starting index of the flattening segment; To flatten the relative time within the segment; initial phase Load from the previous data point of the flattening section By reverse reasoning ; This is the first data point in the flattened segment; S6.2 Perform the pre-connection, for the first data point of the flattening segment. Correction value for the first data point in the flattening segment ,use Correct the first data point of the flattening segment This allows it to transition to the normal segment of the previous cycle; S6.

3. Perform post-connection, for the last data point of the flattening segment. Correction value for the last data point in the flattening segment ,use Correct the last data point of the flattening segment , This represents the load value of the first data point in the normal section after the flattening section ends. S6.

4. For the middle part of the flattened section, all of it shall be done using... Correction; S6.5 Repeat steps S6.1 to S6.4 until all flattened sections are corrected.

8. A system for judging and correcting the phenomenon of flattening load at the blade root of a wind turbine, characterized in that, The method for judging and correcting the flattening phenomenon of wind turbine blade root load as described in any one of claims 1-7 includes: The load data acquisition module is used to acquire real-time blade load data under the grid-connected operation status of the wind turbine and input it into the flattening phenomenon judgment module; The flattening phenomenon judgment module is used to determine whether the blade root oscillation load data in the real-time blade load data has experienced flattening. After collecting real-time blade load data for a preset period, it calculates the absolute value of the first-order difference of the blade root oscillation load data in the real-time blade load data. When several or more consecutive absolute values ​​of the first-order difference are less than a preset threshold, it is determined that the blade root oscillation load data has experienced flattening. The start index and end index of each flattening segment in the blade root oscillation load data are recorded and input into the flattening segment merging module. Otherwise, it is determined that no flattening phenomenon has occurred and the system exits. The segment merging module calculates the duration of each segment based on the start and end indices, and determines whether to merge them into one segment based on the interval duration between each segment. If they need to be merged into one segment, the module calculates the cumulative duration of the merged segment. If they do not need to be merged into one segment, the module calculates the cumulative duration of each segment separately. The cumulative duration judgment module determines whether the cumulative duration calculated by the flattening segment merging module is not greater than the preset cumulative duration threshold; if the cumulative duration is not greater than the preset cumulative duration threshold, it proceeds to the feature extraction module; if the cumulative duration is greater than the preset cumulative duration threshold, it exits the system. The feature extraction module is used to extract features of the normal segment located in the previous cycle of each flattening segment from the blade root oscillation load data. The features include the impeller rotation cycle, blade root oscillation amplitude, and initial phase. The flattening correction module generates the correction load for the flattening segment based on the features obtained by the feature extraction module and the sinusoidal characteristics, and corrects each flattening segment segment by segment. At the same time, it smoothly connects the start and end points of the flattening segment to complete the correction of the flattening phenomenon of the wind turbine blade root load.

9. A non-transitory computer-readable medium storing instructions, characterized in that, When the instruction is executed by the processor, the steps of the method for judging and correcting the flattening phenomenon of wind turbine blade root load according to any one of claims 1-7 are performed.

10. A computing device, comprising a processor and a memory for storing a processor-executable program, characterized in that, When the processor executes the program stored in the memory, it implements the method for judging and correcting the flattening phenomenon of wind turbine blade root load as described in any one of claims 1-7.

Images