A control method and device of a wind power generator, a terminal and a storage medium

Abstract

The application discloses a kind of control method, device, terminal and storage medium of wind driven generator, specifically relates to wind driven generator operation control technical field, for solving the problem that existing control method is caused by the difference of wind and wave load direction dynamic response in marine environment, support structure bears intense asymmetric load, by real-time synchronous acquisition wind direction and wave direction data, calculate wind direction change rate and analyze support structure vibration signal to identify wave excitation dominant frequency component, when detecting wave excitation energy mutation and wind direction change rate is over limit, analyze wind wave direction difference, simultaneously determine wave spectrum energy concentration band by analyzing tower top motion trajectory, in combination with direction difference and wave energy concentration band, evaluate load asymmetric risk level, finally, according to risk level, yaw speed and variable pitch rate are adaptively adjusted, the precise identification and early warning of wind wave load direction mismatching condition are realized, and the operation reliability of wind driven generator under extreme weather is improved.

Description

Technical Field

[0001] This invention relates to the field of wind turbine operation control technology, and in particular to a control method, device, terminal and storage medium for a wind turbine. Background Technology

[0002] During operation, offshore wind turbines typically use a yaw system to adjust the rotor's orientation relative to the wind in order to capture maximum wind energy and ensure structural safety. Existing control methods primarily rely on real-time wind speed and direction signals to control the yaw system to track wind direction changes and execute protective shutdowns under extreme wind conditions. This control strategy assumes that environmental loads are in the same direction and does not adequately consider the dynamic response differences between wind and wave loads in the marine environment.

[0003] However, during extreme weather events such as typhoons, wind direction may change rapidly and significantly, while wave loads lag significantly in direction due to the inertia of water. This instantaneous mismatch between wind and wave load directions causes the wind turbine support structure to bear severe asymmetrical loads, significantly increasing the risk of fatigue damage to the foundation structure and even causing early failure of key components, making it difficult to achieve optimal control of structural safety while ensuring power generation efficiency. Summary of the Invention

[0004] This invention addresses the technical problems existing in the prior art by providing a control method, device, terminal, and storage medium for a wind turbine generator.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] This invention provides the following technical solution:

[0007] A control method for a wind turbine includes:

[0008] S1. Real-time acquisition of wind direction and wave direction data of the environment where the wind turbine is located;

[0009] S2. Obtain the current wind direction change rate, and at the same time determine the dominant frequency components of wave excitation by performing spectrum analysis on the vibration signal of the real-time monitored wind turbine support structure.

[0010] S3. When the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds the preset change threshold, analyze the directional difference between wind direction data and wave direction data.

[0011] S4. Analyze the motion trajectory of the top of the wind turbine tower and determine the concentrated frequency band of wave spectrum energy based on the trajectory motion characteristics;

[0012] S5. Assess the risk of load asymmetry on the wind turbine support structure by combining directional differences and wave spectrum energy concentration frequency bands;

[0013] S6. Adjust the operating parameters of the wind turbine according to the risk of load asymmetry. The operating parameters include the yaw speed of the yaw system and the pitch rate of the pitch system.

[0014] Furthermore, real-time wind direction and wave direction data of the environment where the wind turbine is located are acquired, including:

[0015] Wind direction data is obtained by an ultrasonic anemometer installed on the top of the cabin.

[0016] Wave direction data is acquired by a wave radar installed on the support structure;

[0017] Align and synchronize wind direction data and wave direction data using a unified timestamp.

[0018] Furthermore, the current wind direction change rate is obtained, and spectral analysis is performed on the vibration signals of the real-time monitored wind turbine support structure to determine the dominant frequency components of wave excitation, including:

[0019] The rate of change of wind direction per unit time is calculated based on real-time acquired wind direction data;

[0020] Vibration signals are monitored in real time by an acceleration sensor installed on the wind turbine support structure;

[0021] The time-frequency spectrum is obtained by performing continuous wavelet transform on the vibration signal;

[0022] Based on the characteristic frequency range of wave excitation, the frequency component with the highest energy within the characteristic frequency range of wave excitation is identified from the time-spectrum diagram as the dominant frequency component of wave excitation.

[0023] Calculate the energy integral value of the dominant frequency component of wave excitation within a preset bandwidth;

[0024] The cumulative sum control chart algorithm is applied to the energy integral value sequence to detect energy abrupt change points;

[0025] When an energy mutation point is detected, it is determined that the energy of the dominant frequency component of wave excitation has undergone a mutation.

[0026] Furthermore, when the energy of the dominant frequency component of wave excitation undergoes a sudden change and the rate of change of wind direction exceeds a preset threshold, the directional difference between wind direction data and wave direction data is analyzed, including:

[0027] Under the condition that the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds a preset threshold, calculate the absolute angle difference between wind direction data and wave direction data at the same time stamp.

[0028] The absolute angle difference is processed by moving average filtering to obtain a smoothed direction difference value;

[0029] The smoothed directional difference value is compared with the preset directional difference threshold.

[0030] Furthermore, the motion trajectory of the wind turbine tower top is analyzed, and the wave spectrum energy concentration frequency band is determined based on the trajectory motion characteristics, including:

[0031] The three-dimensional motion trajectory of the tower top is obtained by a positioning system installed on the top of the wind turbine tower;

[0032] The three-dimensional motion trajectory is bandpass filtered to retain the motion trajectory components in the frequency band dominated by the wave load.

[0033] Extract the horizontal motion component of the filtered motion trajectory;

[0034] The characteristic parameters of the motion ellipse are obtained by fitting the horizontal plane motion components to an ellipse.

[0035] Power spectral density analysis was performed on the time series of characteristic parameters of the moving ellipse.

[0036] The frequency band with the highest energy in the power spectral density analysis results is identified as the concentrated energy frequency band of the wave spectrum.

[0037] Furthermore, the risk of load asymmetry on the wind turbine support structure is assessed by combining directional differences and wave spectrum energy concentration frequency bands, including:

[0038] The risk level of directional difference is determined by comparing the smoothed directional difference value with the preset directional difference threshold.

[0039] The wave energy risk level is determined by comparing the energy amplitude of the concentrated frequency band of the wave spectrum with the preset energy threshold.

[0040] Establish a mapping table between directional difference risk levels and wave energy risk levels;

[0041] The corresponding load asymmetry risk level is determined based on the mapping table.

[0042] Furthermore, the operating parameters of the wind turbine are adjusted based on the risk of load asymmetry, including:

[0043] The yaw speed adjustment value of the yaw system is determined based on the correspondence between the load asymmetry risk level and the yaw speed adjustment value.

[0044] The pitch rate adjustment value of the pitch system is determined based on the correspondence between the load asymmetry risk level and the pitch rate adjustment value.

[0045] The determined yaw speed adjustment value and pitch rate adjustment value are sent to the yaw system and pitch system respectively for execution.

[0046] On the other hand, the present invention provides a control device for a wind turbine generator, comprising:

[0047] The data acquisition module is used to acquire wind direction and wave direction data of the environment where the wind turbine is located in real time.

[0048] The signal analysis module is used to obtain the current wind direction change rate and to determine the dominant frequency components of wave excitation by performing spectrum analysis on the vibration signal of the real-time monitored wind turbine support structure.

[0049] The difference analysis module is used to analyze the directional difference between wind direction data and wave direction data when the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds a preset change threshold.

[0050] The trajectory analysis module is used to analyze the motion trajectory of the top of the wind turbine tower and determine the concentrated frequency band of wave spectrum energy based on the trajectory motion characteristics.

[0051] The risk assessment module is used to assess the risk of load asymmetry on the wind turbine support structure by combining directional differences and wave spectrum energy concentration frequency bands.

[0052] The parameter adjustment module is used to adjust the operating parameters of the wind turbine according to the risk of load asymmetry. The operating parameters include the yaw speed of the yaw system and the pitch rate of the pitch system.

[0053] On the other hand, the present invention provides a terminal, which includes a processor, a memory, and a program or instructions stored in the memory and executable on the processor. When the program or instructions are executed by the processor, a method for controlling a wind turbine is implemented.

[0054] On the other hand, the present invention provides a storage medium on which a program or instruction is stored, and when the program or instruction is executed by a processor, a method for controlling a wind turbine is implemented.

[0055] The beneficial effects of this invention are:

[0056] 1. By real-time synchronous monitoring of wind and wave direction data and introducing a joint criterion of wind direction change rate and wave excitation frequency components, high-risk conditions of instantaneous mismatch in wind and wave load directions can be effectively identified. The analysis of the tower top motion trajectory is incorporated into the evaluation system. By extracting the characteristic parameters of the motion ellipse and identifying the concentrated frequency band of wave spectrum energy, accurate perception of the dynamic response of the support structure is achieved. It can provide early warning of asymmetric loads caused by differences in wind and wave directions, providing key decision-making basis for subsequent control strategy adjustments.

[0057] 2. Based on the assessment results of load asymmetry risk, the yaw speed and pitch rate are dynamically adjusted to achieve adaptive matching between operating parameters and real-time sea conditions. By establishing a mapping relationship between risk level and control parameters, power generation efficiency under normal operating conditions is guaranteed, while dynamic structural loads can be alleviated by smoothly reducing the action speed under risky operating conditions. The graded control strategy effectively balances the contradiction between power generation efficiency and structural safety, significantly improves the operational reliability of wind turbines under extreme weather conditions such as typhoons, and extends the service life of key components. Attached Figure Description

[0058] Figure 1 This is a flowchart of a control method for a wind turbine generator according to the present invention;

[0059] Figure 2 This is a schematic diagram of the structure of a control device for a wind turbine generator according to the present invention. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] Example 1: Figure 1 The present invention provides a control method for a wind turbine generator, comprising:

[0062] S1. Real-time acquisition of wind direction and wave direction data of the environment where the wind turbine is located;

[0063] S2. Obtain the current wind direction change rate, and at the same time determine the dominant frequency components of wave excitation by performing spectrum analysis on the vibration signal of the real-time monitored wind turbine support structure.

[0064] S3. When the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds the preset change threshold, analyze the directional difference between wind direction data and wave direction data.

[0065] S4. Analyze the motion trajectory of the top of the wind turbine tower and determine the concentrated frequency band of wave spectrum energy based on the trajectory motion characteristics;

[0066] S5. Assess the risk of load asymmetry on the wind turbine support structure by combining directional differences and wave spectrum energy concentration frequency bands;

[0067] S6. Adjust the operating parameters of the wind turbine according to the risk of load asymmetry. The operating parameters include the yaw speed of the yaw system and the pitch rate of the pitch system.

[0068] S1. Real-time acquisition of wind direction and wave direction data of the environment where the wind turbine is located. Specific implementation includes:

[0069] Real-time wind direction data of the wind turbine's environment is acquired via an ultrasonic anemometer mounted on the top of the nacelle. This anemometer uses a four-channel ultrasonic probe array to measure the time difference of sound wave propagation along different paths. The precise wind direction angle is calculated by taking these time differences. The measurement frequency is four times per second, and the results are output as digital signals through the wind turbine's data acquisition system. Real-time wave direction data of the wind turbine's environment is acquired via a wave radar mounted on the support structure. This wave radar uses X-band microwaves to detect the microscale wave spectrum distribution on the sea surface. The wave propagation direction is inverted by analyzing the spatial distribution characteristics of the wave spectrum. The sampling interval is twice per second, and the measurement data is transmitted to the data acquisition system via Ethernet. When aligning and synchronizing wind direction data and wave direction data with a unified timestamp, the Global Positioning System clock is used as the time reference, and a millisecond-level timestamp is added to each data point. Since the two types of data are collected at different frequencies, the high-frequency wind direction data is interpolated into the time series of the wave direction data using a linear interpolation method, so that the two types of data have exactly the same time coordinates. The aligned data is stored in a circular buffer in the form of a structure. Each data packet contains three fields: timestamp, wind direction angle value, and wave direction angle value. The storage format is binary to save storage space.

[0070] The ultrasonic anemometer is installed at the highest point on the top of the nacelle to avoid the tower shadow effect, at a height of approximately 3 meters from the nacelle roof. The mounting base is made of stainless steel to resist corrosion from the marine environment, and the probe is installed parallel to the central axis of the nacelle. The wave radar is installed on the transition platform of the supporting structure, at a height of approximately 15 meters above sea level. The radar antenna elevation angle is set downwards, for example, at 5 degrees, to optimize sea surface detection. The radome is made of fiberglass to prevent salt spray corrosion. During timestamp synchronization, the data acquisition system uses a network time protocol to synchronize with the GPS clock, with clock synchronization accuracy controlled within ±1 millisecond. For linear interpolation calculations, the two wind direction data points closest to the wave data time point are selected for interpolation. The interpolation formula is: wind direction angle equals the previous wind direction angle plus the difference between the next and previous wind direction angles multiplied by a time scaling factor. Data storage adopts a first-in-first-out queue management method, with a buffer size of approximately 10,000 data packets. When the buffer is full, the oldest data is automatically overwritten.

[0071] The wind direction data measurement principle is based on the speed difference of ultrasonic waves propagating with and against the wind. By measuring the propagation time difference on four mutually perpendicular paths, the wind direction angle within a 360-degree horizontal range is calculated, with an angle resolution of 0.1 degrees and a measurement accuracy of ±2 degrees. Wave direction data acquisition is based on the Doppler frequency shift principle. The microwave signal emitted by the radar is backscattered after being modulated by sea surface roughness. By analyzing the Doppler spectrum and phase information of the echo signal, the wave propagation direction spectrum is extracted, with a direction resolution of 1 degree and a measurement accuracy of ±5 degrees. During data alignment, the timestamp adopts the international standard time format, including year, month, day, hour, minute, second, and millisecond fields. The time scaling factor for linear interpolation is calculated by dividing the difference between the target time point and the previous data point's timestamp by the difference between the two data point timestamps. The interpolation calculation is completed in real time in the field-programmable gate array of the data acquisition system.

[0072] Before data storage, validity verification is performed to check whether the wind direction angle value and the wave direction angle value are within the range of 0 to 360 degrees. Data outside the range is marked as invalid and discarded. Simultaneously, the continuity of timestamps is checked; data segments with discontinuous timestamps are marked and a data retransmission mechanism is initiated. During wind direction data acquisition, when yaw motion of the aircraft cabin is detected, the measured wind direction angle value is transformed into a coordinate system, converting the relative wind direction to the wind direction in the absolute geographic coordinate system. The conversion formula is: absolute wind direction equals relative wind direction plus the aircraft cabin yaw angle. During wave radar data acquisition, the beam direction is automatically adjusted according to tidal changes, and the optimal detection angle is maintained through an electric pitch mechanism. The pitch angle adjustment range is, for example, 0 to 10 degrees, with an adjustment step of, for example, 0.1 degrees.

[0073] S2. Obtain the current wind direction change rate, and simultaneously determine the dominant frequency components of wave excitation by performing spectral analysis on the vibration signals of the real-time monitored wind turbine support structure. Specific implementation includes:

[0074] When calculating the rate of change of wind direction per unit time based on real-time acquired wind direction data, a sliding time window method is used to process the wind direction data sequence. The time window length is set to 10 seconds for example, and the sliding step is 1 second each time. Within each time window, the standard deviation of the wind direction angle value is calculated as the rate of change of wind direction for that time period. The calculation formula is the square root of the sum of squares of the deviations of each wind direction angle value from the average value within the window, divided by the number of data points. The wind direction data comes from the data collected by the ultrasonic anemometer in step S1 and timestamped. During the calculation, invalid data points are excluded, and only the wind direction angle values ​​marked as valid are used for calculation. The implementation of the sliding time window method uses a circular buffer to store the wind direction data of the most recent 10 seconds. Each time new data is added, the oldest data point is removed to keep the data volume in the buffer to be the sampling points within 10 seconds.

[0075] When monitoring vibration signals in real time using accelerometers installed on the wind turbine support structure, a triaxial IEPE piezoelectric accelerometer is used. The sensor is installed approximately 5 meters below the transition section of the support structure, with its installation direction parallel to the support structure's axis. The sensor sampling frequency is set to 100Hz, the range to ±5g, and the sensitivity to 100mV / g. The vibration signal is transmitted to the data acquisition system via a shielded cable. The acquisition system uses a 24-bit analog-to-digital converter to digitize the signal, simultaneously acquiring acceleration signals in three directions during sampling. The sensor is installed using a stainless steel mounting base, fixed to the support structure surface with bolts. The mounting surface is ground before installation to ensure good contact rigidity.

[0076] When performing continuous wavelet transform on the vibration signal to obtain the time spectrum, the Morlet wavelet is used as the mother wavelet, the wavelet center frequency is set to 0.8Hz, and the scaling parameter is dynamically adjusted according to the frequency range to be analyzed; the transform frequency range is set to 0.05Hz to 2Hz, the frequency resolution is 0.01Hz, and the time resolution is 0.1 seconds; during the calculation, the vibration signal is first detrended to eliminate the linear drift component in the signal, then the wavelet coefficients are calculated, and finally the squared modulus of the wavelet coefficients is used as the energy representation of the time spectrum; the detrending process adopts the first-order difference method, that is, subtracting the value of the previous data point from each data point to eliminate the slowly changing components in the signal.

[0077] Based on the wave excitation characteristic frequency range determined in advance through flume tests or numerical simulations, when identifying the frequency component with the highest energy within the wave excitation characteristic frequency range as the dominant frequency component of wave excitation from the time-spectrum diagram, the wave excitation characteristic frequency range is determined in the following way: the wave frequency distribution under typical wave conditions in the target sea area is measured through flume tests, or the peak frequency range of the wave energy spectrum is calculated through numerical simulation. This frequency range is usually from 0.05Hz to 0.3Hz. In the time-spectrum diagram, the energy value of each frequency point is integrated along the time axis within this frequency range, and the frequency point with the largest integrated energy is selected as the dominant frequency component of wave excitation. The flume test uses irregular waves to simulate actual sea conditions, with wave heights ranging from, for example, 0.5 meters to 5 meters and period ranging from, for example, 5 seconds to 20 seconds. The main distribution frequency band of wave energy is obtained by analyzing the experimental data.

[0078] When calculating the energy integral value of the dominant frequency component of wave excitation within a preset bandwidth, the preset bandwidth is set to a range of ±0.02Hz centered on the dominant frequency. The energy integral is calculated using the trapezoidal numerical integration method. A frequency slice within the bandwidth near the dominant frequency is extracted from the time-spectrum graph, and the energy value within the slice is numerically integrated. The integration result is taken as the energy integral value at that time point. During the calculation, an energy integral value is obtained every 0.1 seconds, forming a time series of energy integral values. The numerical integration uses the trapezoidal method formula, which is to multiply the sum of the energy values ​​of two adjacent frequency points by the frequency interval and then divide by 2, and finally sum the areas of all small trapezoids.

[0079] When applying the cumulative sum control chart algorithm to detect energy mutation points on the energy integral value sequence, the reference value of the cumulative sum control chart algorithm is set as the moving average of the energy integral value sequence, and the moving average window length is, for example, 60 seconds; the allowable deviation is set to 0.5 times the standard deviation of the sequence; the algorithm calculates the cumulative deviation of each data point from the reference value, and when the cumulative deviation exceeds the control limit, it is determined as a mutation point; the control limit is dynamically adjusted according to the statistical characteristics of the sequence, and is usually set to 5 times the allowable deviation; in the process of calculating the moving average, a recursive update algorithm is adopted, specifically, the new moving average is obtained by multiplying the moving average of the previous moment by the adjustment coefficient (n-1) / n, and then adding it to the result of multiplying the newly collected data point by 1 / n, where n represents the preset sliding window length.

[0080] When an energy mutation point is detected, it is determined that the energy of the dominant frequency component of wave excitation has changed. The confirmation of the mutation point requires the fulfillment of a continuous condition, such as three consecutive data points exceeding the control limit to be considered a valid mutation. At the same time, a false alarm suppression mechanism is established, and when the mutation amplitude is less than 10% of the historical maximum mutation amplitude, it is not considered a valid mutation. The mutation detection result and the wind direction change rate detection result are logically ANDed and used together as the conditions for triggering subsequent steps. The historical maximum mutation amplitude is obtained by recording the maximum amplitude of all mutation points detected in the past 24 hours, and the historical maximum value is updated every 24 hours.

[0081] The specific implementation of the cumulative sum control chart algorithm includes the following steps: First, calculate the moving average and moving standard deviation of the energy integral value sequence. Then, calculate the deviation of each data point from the moving average. Next, calculate the cumulative sum sequence. The cumulative sum is equal to the previous cumulative sum plus the current deviation minus half of the allowable deviation. When the absolute value of the cumulative sum exceeds the control limit, a sudden alarm is generated.

[0082] For the calculation of moving standard deviation, a recursive update algorithm is used. The new moving standard deviation is calculated through the following steps: first, multiply the moving standard deviation of the previous time step by the adjustment factor (n-1) / n; then, multiply the square of the difference between the new data point and the current moving average by 1 / n; and finally, sum the above two results and take the square root, where n represents the preset sliding window length.

[0083] In determining the characteristic frequency range of wave excitation, the flume test uses a model with a geometric scale of 1:50 to measure the wave spectrum under different combinations of wave height and period. The frequency range of the main energy concentration is determined through spectrum analysis. The numerical simulation uses a third-generation wave model, inputs wind field data and sea topography data, calculates the wave direction spectrum, and extracts the peak frequency distribution range from the direction spectrum. The finally determined characteristic frequency range of wave excitation is stored as a configuration file and dynamically loaded for use during actual monitoring.

[0084] A signal quality check mechanism is set up during vibration signal acquisition to monitor the output signal of the accelerometer in real time. When the signal amplitude continuously exceeds 90% of the range, the range is automatically switched. When the signal noise level exceeds the threshold, digital filtering is started. The filtering adopts a 4th order Butterworth low-pass filter with a cutoff frequency of 10Hz. At the same time, the sensor power supply voltage is monitored. When the voltage is lower than the normal value, a sensor fault alarm is issued.

[0085] Before processing the energy integral value sequence, data smoothing is performed using a moving average filtering method with a window length of, for example, 5 data points. Outlier removal is performed on the smoothed sequence using the 3σ criterion, which removes data points that deviate from the moving average by more than 3 times the standard deviation. The missing positions after removal are filled using linear interpolation, with the interpolation using the average of the two valid data points before and after.

[0086] The parameters of the mutation detection algorithm are adjusted based on historical data statistical analysis. The moving average window length is adjusted according to the fluctuation characteristics of the energy integral value sequence. A longer window is used when the fluctuation is large, and a shorter window is used when the fluctuation is small. The allowable deviation coefficient is adjusted according to the stability of the sequence. The coefficient is set to 0.5 when the sequence is stable and 1.0 when the sequence fluctuates greatly. The control limit coefficient is determined according to the balance between the false alarm rate and the false negative rate, and is usually set between 3 and 5.

[0087] The mutation confirmation mechanism also includes trend verification, which requires that data points after the mutation point maintain the same trend of change, such as three consecutive data points maintaining an increasing or decreasing trend; at the same time, a mutation duration threshold is set, and mutations with a duration of less than 0.5 seconds are considered transient interferences and not confirmed; a mutation energy threshold is established, and a mutation is only confirmed as a valid mutation when the energy integral value of the mutation point exceeds 20% of the baseline level.

[0088] The baseline level is determined by the long-term moving average of the energy integral value sequence, with a moving average window length of, for example, 3600 seconds, and the baseline value is updated every hour; the mutation amplitude is calculated as the percentage difference between the energy integral value at the mutation point and the baseline value; the historical maximum mutation amplitude is recorded as the maximum value in the past 30 days, and the record is automatically updated every morning.

[0089] S3. When the energy of the dominant frequency component of wave excitation undergoes a sudden change and the rate of change of wind direction exceeds a preset threshold, analyze the directional difference between wind direction data and wave direction data. Specific implementation includes:

[0090] Under the condition that both the energy of the dominant frequency component of wave excitation changes abruptly and the wind direction change rate exceeds a preset threshold are satisfied, the absolute angle difference between wind direction data and wave direction data at the same timestamp is calculated. The determination of the energy change of the dominant frequency component of wave excitation is derived from the confirmation result of the energy change point detected by the cumulative sum control chart algorithm in step S2. The determination of the wind direction change rate exceeding the preset threshold is derived from the comparison result of the wind direction change rate calculated in step S2 and the preset threshold. The absolute angle difference is calculated by subtracting the angle value of the wave direction data from the angle value of the wind direction data and taking the absolute value. If the calculated result is greater than 180 degrees, it is subtracted from 360 degrees to ensure that the absolute angle difference is always kept within the range of 0 to 180 degrees. During the calculation process, the data pairs after the timestamp alignment processing in step S1 are strictly used to ensure that the wind direction data points and wave direction data points used in each calculation have the same timestamp identifier.

[0091] The determination of the preset change threshold is based on statistical analysis of historical wind direction change rate data. Typically, the 95th percentile of historical data is taken as the threshold benchmark, and appropriate adjustments are made according to the specific sea area characteristics. The threshold setting process includes collecting wind direction change rate data for at least one full year, calculating the frequency distribution of the data, and selecting a critical value that can cover most normal operating conditions but effectively identify abnormal changes. For example, the preset change threshold can be set to 30 degrees per second. This value represents the degree of drastic change in wind direction per unit time. When the wind direction change rate exceeds this threshold, it is considered that the wind direction has changed significantly. During the historical data collection process, data under extreme weather conditions are excluded, and only wind direction change rate data under normal operating conditions are used for analysis.

[0092] When applying a moving average filter to the absolute angle difference to obtain the smoothed direction difference value, the length of the moving window is determined based on the data sampling frequency and application requirements, for example, a 30-second window length. The filtering process uses an equal-weighted moving average method, which calculates the arithmetic mean of all absolute angle differences within the window. During the filtering process, a circular buffer is used to store the most recent absolute angle difference data. When new data is added, the oldest data point is removed to keep the number of data points within the window constant. Each time the smoothed value is calculated, the average of all valid data points within the window is calculated, and data points marked as invalid are excluded. The size of the circular buffer matches the length of the moving window; for example, a 30-second window corresponds to 300 data points (assuming a sampling frequency of 10Hz).

[0093] The window length for moving average filtering is selected based on the typical time scale of directional difference changes, usually 1 to 2 times the wave characteristic period. For example, in a sea area with a wave period of 10 seconds, setting the window length to 30 seconds can effectively smooth the impact of wave period fluctuations. Before filtering, outliers are removed from the absolute angle difference data. The 3σ criterion is used to exclude data points that deviate from the moving average by more than 3 times the standard deviation. The missing positions after removal are filled using linear interpolation. Linear interpolation uses the values ​​of two consecutive valid data points and calculates the interpolation value according to the time ratio.

[0094] When comparing the smoothed directional difference value with the preset directional difference threshold, the preset directional difference threshold is determined by analyzing the correlation between directional difference and load asymmetry in historical working condition data. The threshold setting process includes collecting directional difference data and corresponding support structure load data under different sea states, establishing a correlation model between the directional difference value and the load asymmetry coefficient, and determining the maximum allowable directional difference value according to engineering safety requirements. The preset directional difference threshold can be set to, for example, 45 degrees. When the smoothed directional difference value exceeds this threshold, it is considered that there is a significant risk of load asymmetry. The load asymmetry coefficient is calculated through stress measurement data of key parts of the support structure, reflecting the degree of unevenness of the structure's stress.

[0095] During the comparison process, a hysteresis comparison strategy is used to avoid frequent switching near the threshold. An upper threshold and a lower threshold are set, for example, the upper threshold is 45 degrees and the lower threshold is 40 degrees. When the smoothed directional difference value rises from below the lower threshold to above the upper threshold, a state change is triggered, and the state is released only when it falls from above the upper threshold to below the lower threshold. This hysteresis comparison can effectively avoid misjudgment caused by measurement noise. The size of the hysteresis interval is determined according to the noise level of the measurement data, and is usually set to 2 to 3 times the standard deviation of the measurement noise.

[0096] The dynamic adjustment mechanism of the preset directional difference threshold is adaptively updated according to real-time sea conditions. The threshold is appropriately relaxed under severe sea conditions and tightened under stable sea conditions. The adjustment is based on marine environmental parameters such as significant wave height and wave period. Intelligent adjustment is achieved by establishing a correspondence between the threshold and marine environmental parameters. The threshold adjustment range is limited to ±20% of the base value to avoid excessive adjustment affecting the detection sensitivity. The marine environmental parameters are obtained from the marine environmental monitoring system or numerical forecast data and are updated every 6 hours.

[0097] All comparison results are stored in association with timestamps, recording the smoothed directional difference value, preset directional difference threshold, and comparison result at each time point. Simultaneously, a data quality flag is established; when the input data quality is poor, the detection sensitivity is automatically reduced or the comparison judgment is paused. Data quality assessment includes indicators such as signal strength, signal-to-noise ratio, and data continuity. If any indicator falls below a set threshold, the data quality is considered unsatisfactory. The signal strength threshold is determined based on the sensor specifications, for example, set to 5% of full scale; the signal-to-noise ratio threshold is set to 20dB; and data continuity requires that consecutive invalid data points do not exceed 10% of the total data points.

[0098] During the directional difference analysis, the rationality of the calculation results is monitored in real time. When the directional difference value is detected to exceed 90 degrees continuously for more than 1 minute, the data verification process is initiated to re-verify the reliability and consistency of the wind direction data and wave direction data. The verification method includes cross-comparing the readings of multiple sensors, checking the synchronization of data timestamps, and re-aligning the data if necessary. If obvious anomalies are found during the data verification process, the comparison results within that time period are marked as unreliable and excluded from subsequent analysis.

[0099] The smoothed directional difference value sequence is simultaneously subjected to trend analysis, and its rate of change and acceleration are calculated. When a rapid increase in the directional difference value is detected, an early warning is issued. The warning threshold is set, for example, for a rate of change exceeding 5 degrees per second or an acceleration exceeding 2 degrees per square second. The warning signal and the final comparison result together constitute a complete basis for judging the load asymmetry risk. The rate of change is calculated using the central difference method, using the previous and next data points to calculate the instantaneous rate of change. The acceleration is obtained by the difference calculation of the rate of change sequence.

[0100] S4. Analyze the motion trajectory of the wind turbine tower top, and determine the concentrated frequency band of wave spectrum energy based on the trajectory motion characteristics. Specific implementation includes:

[0101] The three-dimensional motion trajectory of the wind turbine tower is obtained by a positioning system installed on the top of the tower. This positioning system adopts real-time dynamic differential GPS technology, with the receiver sampling frequency set to 10Hz, achieving a positioning accuracy of centimeters. The installation position is located at the center of the tower top and is fixed by a rigid bracket to ensure a rigid connection with the tower structure. The data output format includes three coordinate components: longitude, latitude, and elevation, and records the GPS timestamp of each data point. The positioning data is transmitted to the data processing system via Ethernet. During transmission, a CRC check mechanism is used to ensure data integrity, and data packets that fail the check are required to be retransmitted.

[0102] When performing bandpass filtering on the three-dimensional motion trajectory to retain the motion trajectory components in the wave load-dominant frequency band, an 8th-order Chebyshev Type I filter is used, with the passband frequency range set from 0.05Hz to 0.3Hz. This frequency range is determined based on typical wave energy distribution characteristics. Before filtering, the motion trajectory data is detrended by using third-order polynomial fitting to remove trend terms and eliminate baseline drift caused by measurement equipment drift or slow environmental changes. The filtered data retains motion components in three directions, with each component confined within the wave load-dominant frequency band.

[0103] When extracting the horizontal motion component of the filtered motion trajectory, the three-dimensional motion trajectory data after bandpass filtering is projected onto the horizontal plane, and the motion components in the two orthogonal directions of east and north are extracted. The projection calculation adopts UTM coordinate system transformation to convert GPS latitude and longitude coordinates into plane rectangular coordinates, while keeping the elevation component unchanged during the coordinate transformation process. The horizontal motion component is stored in the form of a two-dimensional time series, including two channels of eastward displacement and northward displacement, with a sampling interval of 0.1 seconds.

[0104] When obtaining the characteristic parameters of the motion ellipse by ellipse fitting of the horizontal motion components, the least squares method is used to fit the motion trajectory data of two mutually orthogonal directions in the horizontal plane to the ellipse. Each fitting uses data points within a complete wave cycle, and the data length is dynamically adjusted according to the dominant wave frequency. For example, when the dominant frequency is 0.1Hz, 100 data points are used for fitting. The ellipse fitting calculates the azimuth angle of the major axis and the ratio of the minor axis length to the major axis length as the characteristic parameters of the motion ellipse. The azimuth angle of the major axis represents the angle between the major axis of the ellipse and the due north direction, and its value ranges from 0 to 180 degrees.

[0105] The specific process of ellipse fitting using the least squares method includes: first, constructing the general quadratic curve equation of the ellipse, and then transforming the equation into the standard ellipse parameter form; during the solution process, the ellipse parameters are calculated using the eigenvalue decomposition method to ensure the mathematical rationality of the fitting results; after each fitting is completed, a goodness-of-fit test is performed, and the ratio of the residual sum of squares to the total sum of squares is calculated as the goodness-of-fit index. When this index is lower than the threshold (e.g., 0.8), the fitting calculation is repeated.

[0106] When performing power spectral density analysis on the time series of characteristic parameters of a moving ellipse, the focus is on applying Fast Fourier Transform (FFT) to the time series of the major axis azimuth angle. Before analysis, the major axis azimuth angle time series is preprocessed, including outlier removal and data standardization. The FFT uses the Hanning window function with a window length of 1024 data points and an overlap rate of 50%. The power spectral density is calculated using the periodogram method with a frequency resolution of 0.01 Hz and a frequency analysis range of 0 to 5 Hz.

[0107] When identifying the highest energy frequency band in the power spectral density analysis results as the concentrated energy frequency band of the wave spectrum, first find the global peak in the power spectral density curve, and then expand to both sides with the peak frequency as the center until the power value drops to a certain proportion of the peak value (e.g., 50%). This frequency range is the concentrated energy frequency band. During the identification process, parabolic interpolation is used to improve the peak frequency positioning accuracy, and the frequency estimation accuracy can reach 0.001Hz. At the same time, the identified frequency band width is required to be no less than 0.02Hz to avoid misjudging noise peaks as concentrated energy frequency bands.

[0108] The characteristic parameters of the motion ellipse obtained by fitting the horizontal motion components to an ellipse include: using the least squares method to fit the motion trajectory data of two mutually orthogonal directions in the horizontal plane to an ellipse, and calculating the azimuth angle of the major axis and the ratio of the minor axis length to the major axis length as the characteristic parameters of the motion ellipse.

[0109] Power spectral density analysis of the time series of characteristic parameters of the moving ellipse includes: applying Fast Fourier Transform to the time series of the major axis azimuth angle in the characteristic parameters of the moving ellipse, calculating its power spectral density distribution, and identifying the frequency band corresponding to the peak power spectral density as the concentrated frequency band of wave spectrum energy.

[0110] The bandpass filter parameters are dynamically adjusted according to the actual sea conditions. In calm sea conditions, the passband range is appropriately narrowed (e.g., 0.08Hz to 0.25Hz), and in complex sea conditions, the passband range is appropriately widened (e.g., 0.04Hz to 0.35Hz). The adjustment of the passband frequency range is based on real-time wave observation data, and the adjustment range is limited to ±20% of the basic range. The filter coefficients adopt a real-time update mechanism, and the filter coefficients are recalculated every 10 minutes to ensure that the motion trajectory components of the wave load dominant frequency band can be effectively preserved under different sea conditions.

[0111] A data quality check mechanism is set up during the ellipse fitting process. When the correlation coefficient between the eastward and northward motion components is too low (e.g., below 0.5), the data is considered not to meet the characteristics of elliptical motion, and the fitting result is abandoned. At the same time, the eccentricity of the fitted ellipse is monitored. When the eccentricity is too large (e.g., greater than 0.9), the motion trajectory is considered to be close to linear motion, and the fitting result is also abandoned. After each abandonment of fitting, the fitting process is restarted after a sufficient amount of new data has been accumulated.

[0112] The power spectral density analysis employs a multi-segment averaging method to improve the accuracy of spectral estimation. The long axis azimuth time series is divided into multiple overlapping segments, and the power spectral density of each segment is calculated separately. Finally, the average is taken to obtain the final power spectral density estimate. The number of segments is determined based on the data length to ensure that each segment contains sufficient wave cycle information. For example, each segment should contain at least 10 complete wave cycles.

[0113] The energy concentration frequency band identification results are verified post-hocly by comparing the identified frequency band with the synchronously measured wave radar data to verify the consistency of the frequency band. At the same time, the width of the energy concentration frequency band is checked. If the frequency band is too wide (e.g., exceeding 0.1Hz), the re-analysis process is initiated to ensure the reliability of the identification results. During the verification process, the temporal stability of the frequency band energy is also checked, requiring the frequency band energy to remain relatively stable over three consecutive analysis periods.

[0114] All analysis results are stored in association with timestamps, recording the original motion trajectory data, filtered data, ellipse characteristic parameters, power spectral density distribution, and identified energy concentration frequency bands for each analysis period. At the same time, the parameter settings used in the analysis process are saved, including filter parameters, fitting window length, spectral analysis parameters, etc., to facilitate subsequent traceability and verification. The storage format adopts a structured data format, and each data record contains three parts: timestamp, data value, and data quality flag.

[0115] Real-time monitoring is implemented during data analysis. When an abnormal motion trajectory is detected (e.g., the amplitude exceeds twice the historical maximum value), the abnormal handling process is automatically triggered, including data re-acquisition, parameter recalibration, and adaptive adjustment of algorithm parameters. The monitoring indicators include the consistency of motion amplitude, motion frequency, and motion direction. Any indicator that exceeds the normal range will trigger the corresponding processing mechanism.

[0116] The final determined wave spectrum energy concentration frequency band is accompanied by a confidence index; the confidence level is comprehensively evaluated based on data quality, consistency and stability of analysis results, and is divided into three levels: high, medium and low; analysis results with low confidence levels need to be re-analyzed in subsequent use.

[0117] S5. Assess the risk of load asymmetry on the wind turbine support structure by combining directional differences and wave spectrum energy concentration frequency bands. Specific implementation includes:

[0118] When determining the risk level of directional difference based on the comparison between the smoothed directional difference value and the preset directional difference threshold, the preset directional difference threshold is set to multiple levels, such as 30 degrees, 45 degrees, and 60 degrees. The comparison process adopts a multi-level judgment logic: when the smoothed directional difference value is less than 30 degrees, it is rated as a low-risk level; between 30 and 45 degrees, it is rated as a medium-risk level; between 45 and 60 degrees, it is rated as a high-risk level; and when it exceeds 60 degrees, it is rated as an extremely high-risk level. Each risk level corresponds to a numerical identifier, for example, low risk is 1, medium risk is 2, high risk is 3, and extremely high risk is 4. The preset directional difference threshold is determined based on the statistical analysis of long-term monitoring data, collecting directional difference data for at least one complete year, calculating the cumulative distribution function of the data, and selecting a specific percentile as the level threshold.

[0119] When determining the wave energy risk level based on the comparison between the energy amplitude of the concentrated frequency band of the wave spectrum and the preset energy threshold, the preset energy threshold is determined based on statistical analysis of historical wave energy data. For example, the 80th, 90th, and 95th percentile values ​​of historical data are used as the three level thresholds. The energy amplitude is calculated as the integral area of ​​the power spectral density value within the concentrated frequency band of the wave spectrum. The comparison process also adopts a multi-level judgment logic: when the energy amplitude is below the 80th percentile, it is rated as a low-risk level; when it is between the 80th and 90th percentiles, it is rated as a medium-risk level; when it is between the 90th and 95th percentiles, it is rated as a high-risk level; and when it exceeds the 95th percentile, it is rated as an extremely high-risk level. Each risk level corresponds to a numerical identifier, consistent with the directional difference risk level. The process of establishing the wave energy risk level threshold includes collecting historical wave energy data, calculating the energy amplitude distribution under different sea states, and determining the boundary values ​​of each risk level based on structural design specifications and operational experience.

[0120] When establishing the mapping relationship table between directional difference risk level and wave energy risk level, the mapping relationship is determined based on historical load data and structural response analysis. By analyzing the stress measurement data of the supporting structure under different combinations of directional difference and wave energy in historical operating data, the correspondence between risk level combinations and load asymmetry coefficients is established. The mapping relationship table adopts a two-dimensional matrix form, with rows representing directional difference risk levels (1 to 4), columns representing wave energy risk levels (1 to 4), and each cell storing the corresponding load asymmetry risk level. The load asymmetry risk level is also divided into 4 levels, representing slight, moderate, severe, and extreme asymmetry risks, respectively. The establishment of the mapping relationship table adopts a data-driven method, collecting historical operating data including directional difference values, wave energy values, and corresponding stress measurement data of key parts of the supporting structure.

[0121] When determining the corresponding load asymmetry risk level based on the mapping table, a real-time query mechanism is adopted. The current direction difference risk level and wave energy risk level are used as indexes to directly read the corresponding load asymmetry risk level from the mapping table. During the query process, data validity checks are implemented. When the input risk level exceeds the valid range (less than 1 or greater than 4), a conservative strategy is adopted, and the higher risk level is automatically selected. At the same time, the timestamp of each query and the input and output parameters are recorded to form a complete risk assessment log. During the risk assessment process, a multi-verification mechanism is implemented to compare the assessment results based on the mapping table with the estimation results based on the physical model. When the difference between the two exceeds the allowable range, a manual review process is initiated.

[0122] All threshold parameters and mapping tables are stored in the configuration file, supporting online updates and version management. After each parameter update, the old version of the parameters is retained, and the new parameters need to undergo a verification period before they can be officially used. The verification period is usually set to 7 days, during which the old and new sets of parameters are run in parallel to compare and evaluate the consistency of the results. Thresholds are updated regularly, with an update cycle of, for example, 6 months, to ensure that the thresholds can reflect changes in current environmental conditions. Threshold settings take into account seasonal changes, and different threshold standards can be used in different seasons. For example, the threshold level can be appropriately increased during the winter storm season.

[0123] The final determined load asymmetry risk level output includes a confidence index, which is calculated based on the quality of the input data, the consistency of the assessment results, and the completeness of historical data. Assessment results with low confidence will require reassessment in subsequent use. The confidence level is calculated using a fuzzy logic method, considering the weighted summation of multiple influencing factors. The risk assessment results are stored in association with timestamps, recording complete assessment flow data, including input parameters, intermediate results, output level, and confidence index.

[0124] Data analysis excludes data under extreme weather conditions, focusing instead on the distribution characteristics of data under normal operating conditions. The mapping table is recalculated monthly to ensure it reflects the latest structural response characteristics. Simultaneously, the time-varying characteristics of the assessment results are monitored, and data quality checks are initiated when risk levels fluctuate frequently to eliminate misassessments caused by measurement noise. Data is stored in a structured format for easy traceability analysis and system optimization. An automatic data cleanup mechanism is also implemented to periodically remove expired data and maintain the database's efficient operation.

[0125] S6. Adjust the operating parameters of the wind turbine according to the risk of load asymmetry. Specific implementation includes:

[0126] Based on the correspondence between the load asymmetry risk level and the yaw speed adjustment value, the yaw speed adjustment value of the yaw system is determined through structural dynamics simulation analysis and historical operating data statistics. The load asymmetry risk level is divided into four levels, each corresponding to a different yaw speed adjustment strategy. For risk level 1 (slight asymmetry), the yaw speed adjustment value is 0% of the normal speed, i.e., maintaining the original yaw speed unchanged. For risk level 2 (moderate asymmetry), the yaw speed adjustment value is -30% of the normal speed, i.e., reducing the yaw speed by 30%. For risk level 3 (severe asymmetry), the yaw speed adjustment value is -50% of the normal speed. For risk level 4 (extreme asymmetry), the yaw speed adjustment value is -70% of the normal speed. The normal yaw speed is determined based on the rated parameters and design specifications of the wind turbine, and is usually set in the range of 0.5 degrees / second to 1 degree / second, with the specific value adjusted according to the characteristics of the turbine model.

[0127] Based on the correspondence between the load asymmetry risk level and the pitch rate adjustment value, the pitch rate adjustment value of the pitch system is determined by the correspondence based on the blade load distribution characteristics and structural fatigue analysis. The pitch rate adjustment adopts a graded adjustment strategy: for risk level 1, the pitch rate adjustment value is 0% of the normal rate; for risk level 2, the pitch rate adjustment value is -20% of the normal rate; for risk level 3, the pitch rate adjustment value is -40% of the normal rate; and for risk level 4, the pitch rate adjustment value is -60% of the normal rate. The normal pitch rate is determined according to the blade aerodynamic characteristics and control system requirements, and is usually set in the range of 3 degrees / second to 8 degrees / second. The process of determining the adjustment value comprehensively considers the strain measurement data and theoretical calculation values ​​at the blade root, and finds the optimal adjustment parameters through optimization algorithms.

[0128] When the determined yaw speed adjustment value and pitch rate adjustment value are sent to the yaw system and pitch system respectively for execution, an industrial Ethernet communication protocol, such as Modbus TCP protocol or Profinet protocol, is used. The data packet sent contains the adjustment value, timestamp, command priority and check code. The execution process adopts a gradual adjustment method, and the adjustment rate is limited to a change of no more than 10% of the original value per second to avoid additional load impact caused by abrupt changes. The command sending frequency is consistent with the sampling period of the control system, usually set to an interval of 100 milliseconds to 500 milliseconds.

[0129] In determining the yaw speed adjustment value, a wind direction change rate correction factor is introduced. When the wind direction change rate is detected to exceed the threshold, the adjustment range is appropriately reduced to avoid excessive reduction of yaw speed affecting wind accuracy. The correction factor is dynamically calculated based on real-time wind direction data, with a value range of 0.5 to 1.0. At the same time, the current power output level of the wind turbine is considered. A more conservative adjustment strategy is adopted when the power output is high, and a more aggressive adjustment strategy is adopted when the power output is low.

[0130] The pitch rate adjustment value is determined using a closed-loop control method. The load distribution on the blade is monitored in real time by a fiber optic strain sensor installed at the blade root. When the deviation between the measured load and the theoretical value exceeds the allowable range, an online correction mechanism is activated. The correction process uses a proportional-integral-derivative control algorithm to dynamically adjust the pitch rate adjustment value according to the load deviation, ensuring that the blade load is always controlled within a safe range. The allowable range is determined based on the blade material properties and the design safety factor.

[0131] The command transmission adopts a dual redundancy communication mechanism, with the primary and backup communication channels transmitting adjustment commands simultaneously. When the communication quality of the primary channel deteriorates, it automatically switches to the backup channel. The communication data includes a cyclic redundancy check code, and the receiving end can only execute the command after the verification is passed. The execution result is returned to the main control system in real time through the feedback channel, including the actual yaw speed value, pitch rate value, and execution status flag.

[0132] During the adjustment process, stress changes in key parts of the supporting structure are monitored in real time, and data is collected by stress sensors installed at the connection between the tower and the foundation. When the stress value is detected to exceed the safety threshold, the adjustment process is immediately stopped and restored to the safe parameter settings. The safety threshold is determined according to the structural design specifications and is usually set to 80% of the allowable stress of the material. The stop mechanism adopts a hard-wired safety loop to ensure that the protection action can be triggered in a timely manner under any working condition.

[0133] All adjustment operations are recorded in detailed operation logs, including adjustment time, parameters before adjustment, parameters after adjustment, execution results, and monitoring data. The log data is stored in a structured format and retained for no less than 30 days for subsequent analysis and optimization. At the same time, an adjustment effect evaluation mechanism is established to evaluate the effectiveness of the adjustment strategy by comparing the load measurement data before and after the adjustment, and to optimize the corresponding relationship of adjustment values ​​accordingly.

[0134] The correspondence of adjustment values ​​is updated regularly, with an update cycle of 3 months. The updates are based on operational data analysis and structural health monitoring results. The update process adopts a gradual adjustment strategy. New parameters need to be verified through a 7-day trial run. During the verification period, the execution effects of the old and new sets of parameters are recorded in parallel, and the parameter settings with better performance are selected as the final solution. The trial run verification includes tests under various operating conditions to ensure that the new parameters can work stably under different environmental conditions.

[0135] Before final execution, a manual confirmation step is set up. When the adjustment exceeds the predetermined range, such as yaw speed adjustment exceeding the normal value by 50% or pitch rate adjustment exceeding the normal value by 40%, operator confirmation is required before execution. The confirmation interface displays detailed adjustment information, expected effects, and safety assessment results, and provides options to confirm or cancel execution. The confirmation timeout is set to 30 seconds. If no confirmation is made within the timeout period, the adjustment will be automatically canceled, and the original parameters will be maintained.

[0136] During execution, the system status is continuously monitored, including key parameters such as yaw system motor current and pitch system hydraulic pressure. When an abnormal state is detected, the adjustment is immediately stopped and an alarm signal is issued. The judgment of abnormal state is based on a comprehensive evaluation of multiple parameters, including parameter over-limit, abnormal rate of change, abnormal equipment status, etc. The alarm signal is divided into multiple levels, and corresponding handling measures are taken according to the severity of the abnormality.

[0137] Example 2: Figure 2 A schematic diagram of a control device for a wind turbine generator according to the present invention is provided. The control device for a wind turbine generator includes:

[0138] The data acquisition module is used to acquire wind direction and wave direction data of the environment where the wind turbine is located in real time.

[0139] The signal analysis module is used to obtain the current wind direction change rate and to determine the dominant frequency components of wave excitation by performing spectrum analysis on the vibration signal of the real-time monitored wind turbine support structure.

[0140] The difference analysis module is used to analyze the directional difference between wind direction data and wave direction data when the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds a preset change threshold.

[0141] The trajectory analysis module is used to analyze the motion trajectory of the top of the wind turbine tower and determine the concentrated frequency band of wave spectrum energy based on the trajectory motion characteristics.

[0142] The risk assessment module is used to assess the risk of load asymmetry on the wind turbine support structure by combining directional differences and wave spectrum energy concentration frequency bands.

[0143] The parameter adjustment module is used to adjust the operating parameters of the wind turbine according to the risk of load asymmetry. The operating parameters include the yaw speed of the yaw system and the pitch rate of the pitch system.

[0144] Example 3: A terminal, comprising: a processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein the program or instructions, when executed by the processor, implement a control method for a wind turbine generator.

[0145] Example 4: A storage medium storing a program or instructions, which, when executed by a processor, implements a control method for a wind turbine generator.

[0146] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0147] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0148] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0149] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0150] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0151] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0152] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0153] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0154] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0155] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A control method for a wind turbine generator, characterized in that, include: S1. Real-time acquisition of wind direction and wave direction data of the environment where the wind turbine is located; S2. Obtain the current wind direction change rate, and at the same time determine the dominant frequency components of wave excitation by performing spectrum analysis on the vibration signal of the real-time monitored wind turbine support structure. S3. When the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds the preset change threshold, analyze the directional difference between wind direction data and wave direction data. S4. Analyze the motion trajectory of the top of the wind turbine tower and determine the concentrated frequency band of wave spectrum energy based on the trajectory motion characteristics; S5. Assess the risk of load asymmetry on the wind turbine support structure by combining directional differences and wave spectrum energy concentration frequency bands; S6. Adjust the operating parameters of the wind turbine according to the risk of load asymmetry. The operating parameters include the yaw speed of the yaw system and the pitch rate of the pitch system.

2. The control method for a wind turbine generator according to claim 1, characterized in that, Real-time acquisition of wind direction and wave direction data of the environment where the wind turbine is located, including: Wind direction data is obtained by an ultrasonic anemometer installed on the top of the cabin. Wave direction data is acquired by a wave radar installed on the support structure; Align and synchronize wind direction data and wave direction data using a unified timestamp.

3. The control method for a wind turbine generator according to claim 1, characterized in that, The current wind direction change rate is obtained, and the dominant frequency components of wave excitation are determined by performing spectral analysis on the vibration signals of the real-time monitored wind turbine support structure, including: The rate of change of wind direction per unit time is calculated based on real-time acquired wind direction data; Vibration signals are monitored in real time by an acceleration sensor installed on the wind turbine support structure; The time-frequency spectrum is obtained by performing continuous wavelet transform on the vibration signal; Based on the characteristic frequency range of wave excitation, the frequency component with the highest energy within the characteristic frequency range of wave excitation is identified from the time-spectrum diagram as the dominant frequency component of wave excitation. Calculate the energy integral value of the dominant frequency component of wave excitation within a preset bandwidth; The cumulative sum control chart algorithm is applied to the energy integral value sequence to detect energy abrupt change points; When an energy mutation point is detected, it is determined that the energy of the dominant frequency component of wave excitation has undergone a mutation.

4. The control method for a wind turbine generator according to claim 1, characterized in that, When the energy of the dominant frequency component of wave excitation undergoes a sudden change and the rate of change of wind direction exceeds a preset threshold, the directional difference between wind direction data and wave direction data is analyzed, including: Under the condition that the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds a preset threshold, calculate the absolute angle difference between wind direction data and wave direction data at the same time stamp. The absolute angle difference is processed by moving average filtering to obtain a smoothed direction difference value; The smoothed directional difference value is compared with the preset directional difference threshold.

5. The control method for a wind turbine generator according to claim 1, characterized in that, Analyze the motion trajectory of the wind turbine tower top, and determine the wave spectrum energy concentration frequency band based on the trajectory motion characteristics, including: The three-dimensional motion trajectory of the tower top is obtained by a positioning system installed on the top of the wind turbine tower; The three-dimensional motion trajectory is bandpass filtered to retain the motion trajectory components in the frequency band dominated by the wave load. Extract the horizontal motion component of the filtered motion trajectory; The characteristic parameters of the motion ellipse are obtained by fitting the horizontal plane motion components to an ellipse. Power spectral density analysis was performed on the time series of characteristic parameters of the moving ellipse. The frequency band with the highest energy in the power spectral density analysis results is identified as the concentrated energy frequency band of the wave spectrum.

6. The control method for a wind turbine generator according to claim 1, characterized in that, The risk of load asymmetry on wind turbine support structures is assessed by combining directional differences and wave spectrum energy concentration frequency bands, including: The risk level of directional difference is determined by comparing the smoothed directional difference value with the preset directional difference threshold. The wave energy risk level is determined by comparing the energy amplitude of the concentrated frequency band of the wave spectrum with the preset energy threshold. Establish a mapping table between directional difference risk levels and wave energy risk levels; The corresponding load asymmetry risk level is determined based on the mapping table.

7. The control method for a wind turbine generator according to claim 1, characterized in that, Adjusting the operating parameters of wind turbines based on load asymmetry risk includes: The yaw speed adjustment value of the yaw system is determined based on the correspondence between the load asymmetry risk level and the yaw speed adjustment value. The pitch rate adjustment value of the pitch system is determined based on the correspondence between the load asymmetry risk level and the pitch rate adjustment value. The determined yaw speed adjustment value and pitch rate adjustment value are sent to the yaw system and pitch system respectively for execution.

8. A control device for a wind turbine generator, used to implement the control method for a wind turbine generator according to any one of claims 1-7, characterized in that, include: The data acquisition module is used to acquire wind direction and wave direction data of the environment where the wind turbine is located in real time. The signal analysis module is used to obtain the current wind direction change rate and to determine the dominant frequency components of wave excitation by performing spectrum analysis on the vibration signal of the real-time monitored wind turbine support structure. The difference analysis module is used to analyze the directional difference between wind direction data and wave direction data when the energy of the dominant frequency component of wave excitation changes abruptly and the rate of change of wind direction exceeds a preset change threshold. The trajectory analysis module is used to analyze the motion trajectory of the top of the wind turbine tower and determine the concentrated frequency band of wave spectrum energy based on the trajectory motion characteristics. The risk assessment module is used to assess the risk of load asymmetry on the wind turbine support structure by combining directional differences and wave spectrum energy concentration frequency bands. The parameter adjustment module is used to adjust the operating parameters of the wind turbine according to the risk of load asymmetry. The operating parameters include the yaw speed of the yaw system and the pitch rate of the pitch system.

9. A terminal, characterized in that, The terminal includes: A processor, a memory, and a program or instructions stored in the memory and executable on the processor, wherein when the program or instructions are executed by the processor, they implement a control method for a wind turbine as described in any one of claims 1-7.

10. A storage medium, characterized in that, A program or instruction is stored on a storage medium, and when the program or instruction is executed by a processor, it implements a wind turbine control method as described in any one of claims 1-7.

Images