Method and device for correcting the posture of a floating wind turbine tower

Abstract

The application discloses a kind of floating wind turbine tower cylinder posture correction method and device, the method includes: when judging tower cylinder posture exception based on target tower cylinder's multi-source feature data, determine the pitch angle increment according to multi-source feature data and current control parameter, and determine the yaw angle increment according to multi-source feature data;According to the pitch angle increment and the yaw angle increment, the posture of target tower cylinder is corrected, and the posture correction is carried out under the constraint of safety boundary;When judging that tower cylinder posture is still abnormal after correction, optimize current control parameter based on new multi-source feature data;Based on the control parameter after optimization and new multi-source feature data, determine new pitch angle increment and carry out again posture correction, until tower cylinder posture returns to normal.The application can timely and accurately correct the posture of floating wind turbine tower cylinder, meet the stable operation demand of floating wind turbine in complex sea state, and avoid safety hazard.

Description

Technical Field

[0001] This invention relates to the field of wind power generation technology, and in particular to a method and apparatus for correcting the attitude of a floating wind turbine tower. Background Technology

[0002] Floating wind turbines (also known as floating wind generators or floating wind turbines) are wind power generation devices supported by floating platforms (such as semi-submersible, Spar, or tension leg types) and fixed to the seabed by mooring systems. They allow installation in deep-sea areas with water depths exceeding 50 meters, enabling the development of richer and higher-quality wind energy resources, and represent the future trend of offshore wind power development.

[0003] Floating wind turbines operate in complex marine environments, influenced by waves, currents, and wind. Their tower attitude changes are more complex and dynamic. Existing monitoring technologies are largely applicable to stationary wind turbines and have limited effectiveness in floating wind turbines. Even when some technologies can detect abnormal tower attitudes, they cannot directly and precisely correct the attitude of floating wind turbines, leaving safety hazards unresolved and failing to meet the requirements for stable operation of floating wind turbines in complex sea conditions.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This specification provides a method and apparatus for correcting the attitude of a floating wind turbine tower, in order to solve the problem that existing monitoring technologies cannot be applied to floating wind turbines and cannot directly and accurately correct the attitude of floating wind turbines.

[0006] Firstly, embodiments of this specification provide a method for correcting the attitude of a floating wind turbine tower, including:

[0007] When judging the tower attitude anomaly based on the multi-source feature data of the target tower, the pitch angle increment is determined according to the multi-source feature data and the current control parameters, and the yaw angle increment is determined according to the multi-source feature data.

[0008] The target tower is attitude-corrected based on the pitch angle increment and yaw angle increment, and the attitude correction is performed under safety boundary constraints.

[0009] If the tower attitude is still abnormal after correction, the current control parameters are optimized based on new multi-source feature data.

[0010] Based on the optimized control parameters and new multi-source feature data, a new pitch angle increment is determined and the attitude is corrected again until the tower attitude returns to normal.

[0011] In some embodiments, the multi-source feature data includes at least one of the following: mean overturning angle, variance of settlement displacement, strain energy density, rate of change of tilt angle, sway spectrum energy in the 0.1-1Hz frequency band, correlation coefficient between wind speed and tilt angle, bending moment distribution gradient, wave and tower motion transfer function values.

[0012] Accordingly, the method further includes:

[0013] Determine whether the mean overturning angle, variance of settlement displacement, and strain energy density are greater than the corresponding preset safety threshold, and / or whether the tilt angle change rate, sway spectrum energy in the 0.1-1Hz frequency band, and correlation coefficient between wind speed and tilt angle are greater than the corresponding preset safety threshold, and / or whether the bending moment distribution gradient and wave and tower motion transfer function values ​​are greater than the corresponding preset safety threshold.

[0014] If so, determine that the target tower's tower attitude is abnormal.

[0015] In some embodiments, the current control parameters include the current proportional coefficient, the current integral coefficient, and the current derivative coefficient; correspondingly, determining the pitch angle increment based on multi-source feature data and the current control parameters includes:

[0016] The overturning angle deviation is determined based on the difference between the mean overturning angle and the preset overturning angle safety threshold.

[0017] Based on the overturning angle deviation, the current proportional coefficient, the current integral coefficient, and the current differential coefficient, determine the pitch angle increment according to the following formula:

[0018]

[0019] Where Δβ is the pitch angle increment; θ err For overturning angle deviation; k p k is the current scaling factor. i k is the current integral coefficient. d θ is the current differential coefficient; t is time; ∫θ err dt represents the deviation of the overturning angle θ err Integrate over time t; Overturning angle deviation θ err The derivative with respect to time t;

[0020] The step of determining the yaw angle increment based on multi-source feature data includes:

[0021] The yaw angle increment is determined based on the bending moment distribution gradient and the correlation coefficient between wind speed and tilt angle.

[0022] In some embodiments, the attitude correction of the target tower based on the pitch angle increment and yaw angle increment includes:

[0023] Based on the pitch angle increment and yaw angle increment, the correction torque is determined according to the following formula:

[0024] M 修正 =K1·C L (β+Δβ)+K2·(C D (β+Δβ)+K3·Δγ

[0025] Among them, M 修正 To correct the torque; K1 is the coupling coefficient of the first structure of the tower; K2 is the coupling coefficient of the second structure of the tower; C L C is the lift / drag coefficient of the first blade. D Δβ is the lift / drag coefficient of the second blade; β is the current actual pitch angle; Δβ is the pitch angle increment; Δγ is the yaw angle increment;

[0026] The attitude of the target tower is corrected based on the corrected torque.

[0027] In some embodiments, the safety boundary constraints are determined based on the tower buckling critical load calculated in real time;

[0028] Accordingly, the attitude correction is performed under safety boundary constraints, including:

[0029] When the measured tower buckling load is greater than the product of the tower buckling critical load and the first preset value, the pitch rate and yaw rate during the attitude correction process are limited.

[0030] In some embodiments, the method further includes:

[0031] By deploying multiple types of sensors at different locations on the target tower, attitude data of the target tower under different working conditions is re-acquired.

[0032] The re-acquired attitude data is preprocessed to extract new multi-source feature data from the preprocessed attitude data.

[0033] After correction, the tower attitude is determined based on the new multi-source feature data to determine whether it is still abnormal.

[0034] In some embodiments, optimizing the current control parameters based on new multi-source feature data includes:

[0035] Based on the new tilt rate of change, the new sloshing spectrum energy in the 0.1-1Hz frequency band, the first weighting factor, and the second weighting factor, the current scaling factor is optimized according to the following formula to obtain the scaling factor increment:

[0036]

[0037] Where ΔK p This represents the increment of the proportional coefficient; For the new rate of change of tilt angle; PSD 0.1-1Hz ' is the new 0.1-1Hz frequency band oscillation spectrum energy; λ1 is the first weighting factor; λ2 is the second weighting factor;

[0038] Based on the new mean overturning angle, calculate the new overturning angle deviation, and combine it with the third weighting factor to optimize the current integral coefficients according to the following formula, thus obtaining the integral coefficient increment:

[0039]

[0040] Where ΔK i θ is the increment of the integral coefficient; μ is the third weighting factor; θ err '(t) represents the new overturning angle deviation; t is any time point from 0 to t; The new overturning angle deviation θ at time point t, from 0 to t. err Integrate the absolute value of '(t);

[0041] Based on the new bending moment distribution gradient and the fourth weighting factor, the differential coefficients are optimized according to the following formula to obtain the differential coefficient increments:

[0042]

[0043] Where ΔK d η is the increment of the differential coefficient; η is the fourth weighting factor; represents the new bending moment distribution gradient; max is the maximum value.

[0044] In some embodiments, the first weighting factor, the second weighting factor, the third weighting factor, and the fourth weighting factor are determined in the following manner:

[0045] New multi-source feature data is used as state parameters input into a deep reinforcement learning model. The adjustment amounts of the first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are used as action parameters and the action parameters are output. The first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are periodically optimized and updated with the goal of maximizing the reward function that includes correction time, overshoot, and energy loss.

[0046] In some embodiments, the method further includes:

[0047] The new yaw angle increment is determined based on the new bending moment distribution gradient and the new wind speed-tilt angle correlation coefficient.

[0048] Accordingly, the further attitude correction includes:

[0049] The target tower is then re-attitude corrected based on the new pitch angle increment and the new yaw angle increment.

[0050] Once the tower's attitude has returned to normal after correction, attitude correction is stopped.

[0051] Secondly, embodiments of this specification also provide a device for correcting the attitude of a floating wind turbine tower, comprising:

[0052] The pitch angle increment and yaw angle increment determination module is used to determine the pitch angle increment based on the multi-source feature data of the target tower and the current control parameters when judging the tower attitude abnormality based on the multi-source feature data of the target tower; and to determine the yaw angle increment based on the multi-source feature data.

[0053] The first attitude correction module is used to correct the attitude of the target tower according to the pitch angle increment and yaw angle increment, and the attitude correction is performed under safety boundary constraints.

[0054] The control parameter optimization module is used to optimize the current control parameters based on new multi-source feature data when the tower attitude is still abnormal after correction.

[0055] The second attitude correction module is used to determine the new pitch angle increment based on the optimized control parameters and new multi-source feature data, and then perform attitude correction again until the tower attitude returns to normal.

[0056] Thirdly, embodiments of this specification also provide an electronic device, including: a memory and a processor, wherein the processor and the memory are communicatively connected to each other, the memory stores computer instructions, and the processor executes the computer instructions to implement the steps of the above-described method for correcting the attitude of a floating wind turbine tower.

[0057] Fourthly, embodiments of this specification also provide a computer-readable storage medium storing computer program instructions that, when executed by a processor, implement the steps of the above-described method for correcting the attitude of a floating wind turbine tower.

[0058] Fifthly, embodiments of this specification also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the above-described method for correcting the attitude of a floating wind turbine tower.

[0059] This specification provides a method and apparatus for correcting the attitude of a floating wind turbine tower. First, when an abnormal tower attitude is determined based on multi-source feature data of the target tower, the pitch angle increment and yaw angle increment are determined based on the multi-source feature data and current control parameters. Then, the tower attitude is corrected based on the pitch angle and yaw angle increments, under safety boundary constraints. If the tower attitude remains abnormal after correction, the current control parameters are optimized based on new multi-source feature data. Finally, based on the optimized control parameters and new multi-source feature data, a new pitch angle increment is determined, and attitude correction is performed again until the tower attitude returns to normal. This approach overcomes the shortcomings of traditional tower monitoring technologies, which have poor application capabilities on floating wind turbines and cannot correct tower attitude. The method of this invention enables floating wind turbines to better adapt to complex marine environments, reducing equipment failures and damage caused by abnormal tower attitudes and extending the unit's service life. Simultaneously, real-time monitoring and timely correction reduce the difficulty and cost of unit maintenance and improve maintenance efficiency. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0061] Figure 1 This is a flowchart illustrating a method for correcting the attitude of a floating wind turbine tower, as provided in the embodiments of this specification.

[0062] Figure 2 This is a schematic diagram of the sensor arrangement of the target tower provided in the embodiments of this specification;

[0063] Figure 3 This is a diagram of the attitude monitoring and real-time control system architecture provided in the embodiments of this specification;

[0064] Figure 4 This is a flowchart of attitude monitoring and real-time control provided in the embodiments of this specification;

[0065] Figure 5 This is a schematic diagram of the structural composition of a floating wind turbine tower attitude correction device provided in the embodiments of this specification;

[0066] Figure 6 This is a schematic diagram of the structural composition of the electronic device provided in the embodiments of this specification. Detailed Implementation

[0067] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0068] As mentioned earlier, the current tower attitude monitoring technology for floating wind turbines is not mature enough. Existing tower monitoring technologies are only applicable to stationary offshore units and lack attitude correction capabilities. Specific problems are as follows:

[0069] 1. Insufficient monitoring accuracy and adaptability: Traditional tower monitoring technologies are mostly applicable to stationary wind turbines, with poor application capabilities to floating wind turbines. Floating wind turbines, situated in complex marine environments, are affected by the coupling effects of waves, currents, and wind, resulting in more complex and dynamic tower attitude changes. Existing monitoring methods struggle to accurately and in real-time capture these changes. In other words, current technologies are only designed for monitoring the towers of stationary wind turbines and cannot handle the dynamic coupling effects of wave-current-wind loads.

[0070] 2. Lack of effective attitude correction function: Even if some technologies can detect abnormal tower attitude, they cannot directly and accurately control the attitude of floating wind turbines, resulting in safety hazards still existing and failing to meet the requirements for stable operation of floating wind turbines in complex sea conditions.

[0071] To address the aforementioned problems, embodiments of this specification provide a method and apparatus for correcting the attitude of a floating wind turbine tower. The method includes: First, when the tower attitude is determined to be abnormal based on multi-source feature data of the target tower, determining the pitch angle increment based on the multi-source feature data and current control parameters, and determining the yaw angle increment based on the multi-source feature data. Then, performing attitude correction on the target tower based on the pitch angle increment and yaw angle increment, wherein the attitude correction is performed under safety boundary constraints. Next, if the tower attitude is still determined to be abnormal after correction, optimizing the current control parameters based on new multi-source feature data. Finally, based on the optimized control parameters and new multi-source feature data, determining a new pitch angle increment and performing attitude correction again, until the tower attitude returns to normal.

[0072] The above solution overcomes the shortcomings of traditional tower monitoring technologies, such as poor applicability to floating wind turbines and the inability to correct tower attitude. This invention enables floating wind turbines to better adapt to complex marine environments, reducing equipment failures and damage caused by abnormal tower attitude, and extending the unit's service life. Simultaneously, real-time monitoring and timely correction reduce maintenance difficulty and costs, and improve maintenance efficiency.

[0073] It should be noted that the information and data related to users involved in the embodiments of this specification are all information and data authorized by the user or fully authorized by the relevant parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with relevant laws, regulations, and standards, do not violate public order and good morals, and provide corresponding operation entry points for users or relevant parties to choose to authorize or refuse.

[0074] It should also be noted that in the embodiments of this specification, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that this application has used or necessarily used such a solution.

[0075] See Figure 1 As shown in the embodiments of this specification, a method for correcting the attitude of a floating wind turbine tower is provided. In specific implementation, this method may include the following:

[0076] S101: When judging the tower attitude abnormality based on the multi-source feature data of the target tower, determine the pitch angle increment according to the multi-source feature data and the current control parameters, and determine the yaw angle increment according to the multi-source feature data.

[0077] In some embodiments, prior to S101 above, the following may be included in a specific implementation:

[0078] Sensor data acquisition steps: Collect attitude data of the target tower under different working conditions by deploying multiple types of sensors at different locations on the target tower;

[0079] Data preprocessing steps: Perform data preprocessing on the collected attitude data;

[0080] Feature extraction steps: Extract multi-source feature data from the pose data after data preprocessing;

[0081] Anomaly assessment steps: Determine whether the target tower's attitude is abnormal based on the extracted multi-source feature data.

[0082] Specifically, the target tower can be a floating wind turbine tower or a floating wind generator tower. The various types of sensors deployed at different locations on the target tower can include: sensors deployed at the top, middle, and bottom of the target tower and on the floating foundation platform, such as a dual-axis tilt sensor and a triaxial accelerometer deployed at the top of the target tower; strain sensors and tilt sensors deployed at the middle and bottom of the target tower; and environmental sensors deployed on the floating foundation platform of the target tower, such as wave radar and current meters. The attitude data can include: tilt data of the target tower collected by the tilt sensor; acceleration data of the target tower collected by the triaxial accelerometer; strain data of the target tower collected by the strain sensor; and environmental data of the target tower collected by the environmental sensor. Data preprocessing can include, but is not limited to: wavelet denoising, temperature compensation, and coordinate transformation processing. Preprocessed pose data can be transmitted to edge computing nodes for storage via the OPC UA protocol. Specifically, it can be stored in a time-series database (such as InfluxDB), retaining 30 days of rolling data. Edge computing nodes are devices or apparatuses that perform data processing and analysis at the network edge. Storing preprocessed pose data improves the efficiency of subsequent multi-source feature data extraction.

[0083] Because floating wind turbines are located in complex marine environments and are affected by factors such as waves, currents, and wind, their tower attitude changes are more complex and dynamic. Therefore, various types of sensors can be installed or deployed at different locations (such as different key parts) of the floating wind turbine tower to more accurately collect attitude data of the floating wind turbine tower under different operating conditions, realize real-time and accurate monitoring of the tower attitude, and thus provide a reliable basis for timely and accurate tower attitude correction.

[0084] See Figure 2 As shown, Figure 2 This diagram illustrates the sensor layout for the target tower. Tilt sensors (specifically, dual-axis tilt sensors) and triaxial accelerometers can be placed, deployed, or installed at the top of the target tower. Strain sensors and tilt sensors can be placed in the middle and bottom of the target tower. Environmental sensors, such as wave radar and current meters, can be placed on the floating foundation platform and correlated with environmental disturbance data. An anemometer can be installed at the top of the tower. Subsequently, anti-interference design can be implemented, such as preprocessing sensor signals through an EMI shielding layer and digital filtering modules to suppress electromagnetic interference from marine salt spray.

[0085] Specifically, the above sensor data acquisition steps adopt a dual-mode strategy of "high-frequency sampling + event triggering": in the normal mode, data is collected at a frequency of 10Hz; when the wave height is greater than 2m or the wind speed is greater than 15m / s, it automatically switches to 50Hz high-frequency sampling.

[0086] The above data preprocessing steps may include:

[0087] Wavelet denoising: The original signal is decomposed using the Morlet wavelet basis, preserving the tower's natural frequency. This method is suitable for tilt sensors and triaxial accelerometers.

[0088]

[0089] Where x(t) is the denoised signal; c k These are wavelet coefficients; a k b is the scaling factor (correlation frequency); k This is the time shift factor.

[0090] It should be noted that strain sensors do not require noise reduction, while environmental sensors require moving average filtering.

[0091] Tilt sensor temperature compensation (sensor temperature compensation refers to the process of correcting sensor measurement errors caused by changes in ambient temperature using specific techniques and algorithms):

[0092] θ 校准 =θ 原始 -δ·(T-T0)

[0093] Where, θ 校准 The calibrated tilt angle; θ 原始 δ is the original tilt angle; δ is the temperature coefficient (0.002° / ℃); T is the real-time temperature; T0 is the reference temperature (25℃).

[0094] It should be noted that the strain sensor uses an automatic temperature compensation bridge, while the triaxial accelerometer and environmental sensor do not require compensation.

[0095] The above feature extraction steps may include:

[0096] Extract the mean overturning angle θ from the tilt angle data of the target tower collected by the tilt sensor. avg σ z 2 Inclination rate of change Its physical meaning is the overall tilt and settlement stability of the tower; the sway spectrum energy (PSD) in the 0.1-1Hz frequency band is extracted from the acceleration data of the target tower collected by the triaxial accelerometer. 0.1-1Hz Resonance peak frequency f res Effective value of acceleration a rms Its physical meaning is the dynamic vibration characteristics and resonance risk of the tower; the bending moment distribution gradient is extracted from the strain data of the target tower collected by the strain sensor. Strain energy density U eIts physical meaning is the stress concentration and buckling risk of the tower structure; the wave-to-tower motion transfer function value (i.e., wave-tower motion transfer function value) H is extracted from the environmental data of the target tower collected by environmental sensors. wave (f) Correlation coefficient between wind speed and tilt angle (i.e., wind speed-tilt angle correlation coefficient) ρ v,θ Its physical meaning is the driving mechanism of attitude by environmental load.

[0097] The above-mentioned posture anomaly detection steps may include:

[0098] Multi-level criteria, such as static over-limit criteria, dynamic trend criteria, and environmental coupling risk criteria, can be constructed based on multi-source feature data. An abnormal tower attitude is determined as long as any one of these criteria is triggered or satisfied. The static over-limit criteria are as follows:

[0099] |θ avg |>5° (the absolute value of the mean overturning angle is greater than 5°), σ z 2 >0.2m 2 (Settlement displacement variance greater than 0.2 square meters) and U e >U 临界应变能 (Strain energy density is greater than critical strain energy);

[0100] The dynamic trend criterion is:

[0101] (tilt change rate greater than 0.5 degrees per second), PSD 0.1-1Hz >0.2g 2 / Hz (the wobble spectrum energy in the 0.1-1Hz frequency band is greater than 0.2 squared acceleration per Hz) and ρ v,θ >0.8 (correlation coefficient between wind speed and tilt angle is greater than 0.8);

[0102] The environmental coupling risk criterion is:

[0103] |H wave (f)|>1.5 and (The absolute value of the wave and tower motion transfer function is greater than 1.5 and the bending moment distribution gradient is greater than the bending moment threshold).

[0104] The priority of the static over-limit criterion can be set to be higher than that of the dynamic trend criterion, and the priority of the dynamic trend criterion can be set to be higher than that of the environmental coupling risk criterion. Accordingly, the higher the priority of the criterion, the shorter the time for attitude correction. For example, if the static over-limit criterion is triggered, the attitude correction step needs to be executed within time t1. If the dynamic trend criterion is triggered, the attitude correction step needs to be executed within time t2. If the environmental coupling risk criterion is triggered, the attitude correction step needs to be executed within time t3. Here, t1 < t2 < t3, such as t1 is 1 second, t2 is 2 seconds, t3 is 5 seconds, etc.

[0105] In some embodiments, the multi-source feature data in S101 above may include at least one of the following: mean overturning angle θ avg σ z 2 Strain energy density U e Inclination rate of change PSD of oscillation spectrum energy in the 0.1-1Hz frequency band 0.1-1Hz Correlation coefficient ρ between wind speed and tilt angle v,θ Bending moment distribution gradient Wave and tower motion transfer function value H wave (f);

[0106] Accordingly, the determination of whether the target tower's attitude is abnormal before step S101 may include:

[0107] Determine the mean overturning angle θ avg σ z 2 and strain energy density U e Whether it exceeds the corresponding preset safety threshold, and / or the rate of change of tilt angle. PSD of oscillation spectrum energy in the 0.1-1Hz frequency band 0.1-1Hz And the correlation coefficient ρ between wind speed and tilt angle v,θ Whether it exceeds the corresponding preset safety threshold, and / or bending moment distribution gradient And the wave and tower motion transfer function value H wave (f) Whether it is greater than the corresponding preset security threshold;

[0108] If so, determine that the target tower's tower attitude is abnormal.

[0109] Specifically, one can determine the absolute value of the mean overturning angle |θ avg |Whether it is greater than 5°, settlement displacement variance σ z 2 Is it greater than 0.2m? 2 And whether the strain energy density is greater than U 临界应变能 If so, then the tower's attitude is determined to be abnormal. This includes 5° and 0.2m. 2 U 临界应变能 As the mean overturning angle θ avg σ z 2 and strain energy density U e The corresponding preset safety threshold. And / or, can be used to determine the rate of change of tilt angle. Does the sway spectrum energy PSD exceed 0.5° / s in the 0.1-1Hz frequency band? 0.1-1HzIs it greater than 0.2g? 2 / Hz and the correlation coefficient between wind speed and tilt angle ρ v,θ If the value is greater than 0.8, then the tower's attitude is considered abnormal. This includes values ​​such as 0.5° / s and 0.2g. 2 / Hz and 0.8 are respectively used as the tilt angle change rate. PSD of oscillation spectrum energy in the 0.1-1Hz frequency band 0.1-1Hz And the correlation coefficient ρ between wind speed and tilt angle v,θ The corresponding preset safety threshold. And / or, can determine the bending moment distribution gradient. Whether it exceeds the bending moment threshold and the tower's motion transfer function value H wave If the absolute value of (f) is greater than 1.5, then the tower attitude of the target tower is determined to be abnormal. M can be... th 1.5 as the bending moment distribution gradient And the wave and tower motion transfer function value H wave (f) corresponds to the preset security threshold.

[0110] If none of the above criteria are met, the target tower's attitude is determined to be normal, and no attitude correction is currently required. Subsequently, if any of the above criteria is triggered, the target tower's attitude can be determined to be abnormal, and the target tower's attitude correction steps can be executed.

[0111] In some embodiments, the sloshing spectrum energy PSD in the 0.1-1Hz frequency band described above 0.1-1Hz It can be determined using the following formula:

[0112]

[0113] Among them, PSD 0.1-1Hz (f) represents the oscillation spectrum energy in the 0.1-1Hz frequency band; M is the number of signal segments; X m (f) is the FFT spectrum (the Fast Fourier Transform (FFT) spectrum can include both amplitude and phase spectra).

[0114] The wave and tower motion transfer function values ​​mentioned above can be determined using the following formula:

[0115] The transfer function between wave height Hwave and tower top acceleration a can be established using the following formula:

[0116]

[0117] Where Hwave(f) is the wave and tower motion transfer function value; s a,h (f) represents the cross-power spectrum; s h (f) is the wave self-power spectrum.

[0118] Based on the above embodiments, this invention utilizes high-precision multi-type sensors deployed at different parts of the target tower to accurately collect attitude data under various operating conditions, thereby accurately extracting multi-source feature data. Based on the extracted multi-source feature data, it determines whether any of the multi-level criteria is triggered, thus enabling timely and accurate tower attitude monitoring, timely detection of tower attitude anomalies, and laying the foundation for subsequent accurate and efficient attitude correction.

[0119] In some embodiments, the current control parameters in S101 may include the current proportional coefficient, the current integral coefficient, and the current derivative coefficient; correspondingly, the determination of the pitch angle increment based on multi-source feature data and the current control parameters in S101 may, in specific implementation, include:

[0120] The overturning angle deviation is determined based on the difference between the mean overturning angle and the preset overturning angle safety threshold.

[0121] Based on the overturning angle deviation, the current proportional coefficient, the current integral coefficient, and the current differential coefficient, determine the pitch angle increment according to the following formula:

[0122]

[0123] Where Δβ is the pitch angle increment; θ err For overturning angle deviation; k p k is the current scaling factor. i k is the current integral coefficient. d θ is the current differential coefficient; t is time; ∫θ err dt represents the deviation of the overturning angle θ err Integrate over time t; Overturning angle deviation θ err The derivative with respect to time t;

[0124] In the above-mentioned S101, determining the yaw angle increment based on multi-source feature data can, in specific implementation, include:

[0125] The yaw angle increment is determined based on the bending moment distribution gradient and the correlation coefficient between wind speed and tilt angle.

[0126] Specifically, the aforementioned floating wind turbine generator can be connected to the wind turbine main control system, sending monitoring results (such as abnormal tower attitude) to the system. The main control system then controls and adjusts the turbine's pitch and yaw (e.g., calculating pitch angle increments and yaw angle increments) to quickly correct the tower attitude. The main control system includes a PID controller, which involves control parameters such as proportional, integral, and derivative coefficients. These coefficients can be adjusted in real-time based on changes in multi-source characteristic data; this will be explained further later and will not be elaborated upon here.

[0127] The steps for calculating the pitch angle increment can be as follows:

[0128] The overturning angle deviation can be determined using the following formula:

[0129] θ err =θ avg -θ safe

[0130] Where, θ err For overturning angle deviation; θ avg The mean overturning angle; θ safe This is a preset overturning angle safety threshold.

[0131] Then, based on the overturning angle deviation θ err Current proportionality coefficient k p Current integral coefficient k i and the current differential coefficient k d The pitch angle increment Δβ is determined using the following formula:

[0132]

[0133] The steps for calculating the yaw angle increment are as follows:

[0134] It can be based on the bending moment gradient Correlation coefficient ρ between wind speed and tilt angle v,θ Dynamic decision-making to determine the yaw angle increment Δγ:

[0135]

[0136] The function f can be a lookup table, fuzzy logic, or neural network mapping; its specific form needs to be calibrated in the actual system. For example, the yaw angle increment Δγ can be determined using the following formula:

[0137]

[0138] Where K is the gain coefficient, a core coefficient that needs to be calibrated through simulation and field testing, converting the physical quantity of bending moment gradient into yaw angle; sign(ρ v,θ ) is a sign function, which can take the sign (positive or negative) of the correlation coefficient between wind speed and tilt angle, and is used to determine the yaw direction; The sign and magnitude of the bending moment gradient directly determine the direction and amplitude of Δγ. The term is an exponentially decaying term, a smoothing function used to prevent [further issues]. When Δγ is at its maximum, it becomes too large, which enhances the smoothness and stability of the system. saturate(|ρ) is the attenuation coefficient. v,θ |) is a saturation function, and |ρ v,θ The value of | is limited to the range [0,1].

[0139] By determining the pitch angle increment and yaw angle increment, basic data can be provided for the target tower attitude correction.

[0140] S102: The attitude of the target tower is corrected based on the pitch angle increment and yaw angle increment, and the attitude correction is performed under safety boundary constraints.

[0141] In some embodiments, the attitude correction of the target tower based on the pitch angle increment and yaw angle increment in S102 above may, in specific implementation, include:

[0142] Based on the pitch angle increment and yaw angle increment, the correction torque is determined according to the following formula:

[0143] M 修正 =K1·C L (β+Δβ)+K2·(C D (β+Δβ))+K3·Δγ

[0144] Among them, M 修正 To correct the torque; K1 is the coupling coefficient of the first structure of the tower; K2 is the coupling coefficient of the second structure of the tower; C L C is the lift / drag coefficient of the first blade. D Δβ is the lift / drag coefficient of the second blade; β is the current actual pitch angle; Δβ is the pitch angle increment; Δγ is the yaw angle increment;

[0145] The attitude of the target tower is corrected based on the corrected torque.

[0146] Specifically, the wind turbine's main control system can first obtain the current actual pitch angle β. This value represents the pitch angle set by the wind turbine to pursue optimal aerodynamic efficiency (or other conventional goals), and is a known, real-time changing control variable. Then, it combines this with the pitch angle increment Δβ, which represents the additional adjustment added (or reduced) to the current actual pitch angle, to calculate β+Δβ. This value simultaneously considers power generation efficiency and structural safety. Finally, it combines this with the previously calculated yaw angle increment Δγ, and the known C... L C D K1 and K2 (calibrated through finite element simulation) can be used to determine the correction torque (i.e., the aerodynamic load correction torque) according to the following formula:

[0147] M 修正 =K1·C L (β+Δβ)+K2·(C D (β+Δβ)+K3·Δγ.

[0148] The target tower's attitude can be accurately and quickly corrected by adjusting the torque. The adjusting torque can counteract the harmful bending moment caused by external loads on the target tower, ensuring the structural safety of the target tower under complex sea conditions.

[0149] In some embodiments, the safety boundary constraints in S102 above are determined based on the tower buckling critical load calculated in real time;

[0150] Accordingly, the attitude correction in S102 above is performed under safety boundary constraints, and in specific implementation, it may include:

[0151] When the measured tower buckling load is greater than the product of the tower buckling critical load and the first preset value, the pitch rate and yaw rate during the attitude correction process are limited.

[0152] Specifically, the aforementioned critical buckling load for the tower refers to the minimum load value at which the target tower, as a slender structure, is about to undergo instability and failure (buckling) under pressure. This can be represented by M. 临界 M indicates 临界 This is the ultimate bearing capacity at which the target tower structure will not experience instability or failure. The aforementioned first preset value can be set to 0.6 based on experimental or practical needs, within M... 实测 >0.6M 临界At this point, it means the target tower structure is already in a high-risk state, but has not yet failed immediately. Safety constraints can be applied during attitude correction, such as limiting the pitch and yaw rates, to prevent structural damage. Specifically, the applied safety constraints can also include forcing the derivative coefficient to 0 (i.e., disabling the derivative term in PID control) and limiting the pitch rate to ≤2° / s. Since the derivative term will respond drastically to rapid changes in the signal, under high structural stress, this drastic response may introduce impact loads. Setting it to zero makes the control law smoother. By limiting the pitch and yaw rates, impact torques on the tower and other structures can be avoided from rapid pitch maneuvers.

[0153] The critical buckling load of the tower can be determined according to the following formula:

[0154]

[0155] Among them, M 临界 Δ is the critical buckling load of the tower; E is the elastic modulus; I is the moment of inertia of the section; L is the target tower height; D is the diameter; Δ 沉降 This represents the settlement displacement.

[0156] In some embodiments, prior to the attitude correction in step S102, the following may also be included:

[0157] Real-time calculation of the critical buckling load of the tower;

[0158] When the measured tower buckling load exceeds the product of the tower buckling critical load and the second preset value, the floating wind turbine generator operation is stopped. The first preset value is less than the second preset value.

[0159] Specifically, the second preset value can be set to 0.7 according to experimental or actual needs, and the second preset value is greater than the first preset value. In M 实测 >0.7M 临界 In such cases, an emergency shutdown can be triggered to perform protective actions such as propeller retraction and yaw braking to prevent tower structure failure.

[0160] S103: If the tower attitude is still abnormal after correction, optimize the current control parameters based on new multi-source feature data.

[0161] In some embodiments, prior to S103 above, the following may be included in a specific implementation:

[0162] By deploying multiple types of sensors at different locations on the target tower, attitude data of the target tower under different working conditions is re-acquired.

[0163] The re-acquired attitude data is preprocessed to extract new multi-source feature data from the preprocessed attitude data.

[0164] After correction, the tower attitude is determined based on the new multi-source feature data to determine whether it is still abnormal.

[0165] Specifically, after the tower attitude is corrected, it is necessary to determine whether the attitude correction is effective. That is, it is necessary to continue to monitor whether the tower attitude is still abnormal. If it is still abnormal, the attitude correction is ineffective and the control parameters need to be further optimized to ensure the correction effect. If the tower attitude returns to normal, the attitude correction is effective and there is no need to further optimize the control parameters.

[0166] Specifically, multiple sensors deployed at the top, middle, and bottom of the target tower and on the floating foundation platform can be used to re-collect attitude data of the target tower under different operating conditions. This re-collected attitude data is then processed with wavelet denoising, temperature compensation, and coordinate transformation to extract new multi-source feature data from the pre-processed attitude data. Finally, the new mean overturning angle θ' can be determined. avg The new settlement displacement variance σ z 2 'and the new strain energy density U e 'Whether it exceeds the corresponding preset safety threshold, and / or the new rate of change of tilt angle.' New 0.1-1Hz frequency band oscillation spectrum energy PSD 0.1-1Hz 'and the new correlation coefficient ρ between wind speed and tilt angle v,θ 'Whether it exceeds the corresponding preset safety threshold, and / or the new bending moment distribution gradient.' And the new wave and tower motion transfer function value H wave (f)' If the value is greater than the corresponding preset safety threshold, then the tower attitude of the target tower is still abnormal.

[0167] In some embodiments, optimizing the current control parameters based on new multi-source feature data in S103 above includes, in specific implementation:

[0168] Based on the new tilt rate of change, the new sloshing spectrum energy in the 0.1-1Hz frequency band, the first weighting factor, and the second weighting factor, the current scaling factor is optimized according to the following formula to obtain the scaling factor increment:

[0169]

[0170] Where ΔK p This represents the increment of the proportional coefficient; For the new rate of change of tilt angle; PSD 0.1-1Hz ' is the new 0.1-1Hz frequency band oscillation spectrum energy; λ1 is the first weighting factor; λ2 is the second weighting factor;

[0171] Based on the new mean overturning angle, calculate the new overturning angle deviation, and combine it with the third weighting factor to optimize the current integral coefficients according to the following formula, thus obtaining the integral coefficient increment:

[0172]

[0173] Where ΔK i θ is the increment of the integral coefficient; μ is the third weighting factor; θ err '(t) represents the new overturning angle deviation; t is any time point from 0 to t; The new overturning angle deviation θ at time point t, from 0 to t. err Integrate the absolute value of '(t);

[0174] Based on the new bending moment distribution gradient and the fourth weighting factor, the differential coefficients are optimized according to the following formula to obtain the differential coefficient increments:

[0175]

[0176] Where ΔK d η is the increment of the differential coefficient; η is the fourth weighting factor; represents the new bending moment distribution gradient; max is the maximum value.

[0177] Specifically, the first weighting factor λ1 and the second weighting factor λ2 mentioned above can also be called the first working condition weighting factor and the second working condition weighting factor, and the third weighting factor μ mentioned above can also be called the cumulative error sensitivity coefficient. The new overturning angle deviation θ mentioned above err '(t)=θ avg '(t)-θ safe (t), the fourth weighting factor η mentioned above can also be called the stress gradient influence factor. It can be determined based on the absolute value of the new tilt angle change rate. New 0.1-1Hz frequency band oscillation spectrum energy PSD 0.1-1Hz 'The first weighting factor λ1 and the second weighting factor λ2 are used to calculate the scaling factor increment ΔK. p The new overturning angle deviation can be determined based on the absolute value of |θ|. err Using '(t)| and the third weighting factor μ, the integral coefficient increment ΔK is calculated. i The maximum value of the new bending moment distribution gradient can be used as a reference. And the fourth weighting factor η, calculate the increment of the differential coefficient ΔK d .

[0178] After calculating the above proportionality coefficient increment ΔK p Integral coefficient increment ΔK i , Differential coefficient increment ΔK dSubsequently, to avoid these theoretically excessive values ​​causing drastic changes in the characteristics of the PID controller, which could lead to sudden changes in control commands, causing violent actions of the wind turbine actuators (pitch and yaw motors), generating mechanical shocks, or even directly causing system instability, a saturation function can be used to process the aforementioned proportional coefficient increment ΔK. p Integral coefficient increment ΔK i , Differential coefficient increment ΔK d This causes the proportionality coefficient to increase by ΔK. p Integral coefficient increment ΔK i , Differential coefficient increment ΔK d Limit it to a reasonable, predefined safety range such as [ΔK] min ,ΔK max [Within]. The expression for the saturation function mentioned above is as follows:

[0179]

[0180] The above-mentioned proportional coefficient increment ΔK is processed using a saturation function. p Integral coefficient increment ΔK i , Differential coefficient increment ΔK d The specific process is as follows:

[0181] Increment the proportionality coefficient ΔK p Integral coefficient increment ΔK i , Differential coefficient increment ΔK d Input the saturation function (sat(x,x) respectively) max The process is carried out in the proportional coefficient increment ΔK. p For example, if ΔK p >ΔK p max The saturation function will then limit the output value to ΔΔK. p max If ΔK p <-ΔK p max The saturation function will then limit the output value to -ΔK. p max Otherwise, the saturation function outputs ΔK as is. p .

[0182] Then, the constrained and safe increment sat(ΔK) p ,ΔK p max (and the current benchmark value of the proportional coefficient) The summation yields the new proportional coefficient increment that will ultimately be applied to the controller.

[0183]

[0184] Finally, the calculated increment of the new proportional coefficient can be... The proportional gain is updated in the PID controller to calculate the pitch angle increment Δβ for the next control cycle.

[0185] Similarly, the increment of the new integral coefficient can be obtained.

[0186]

[0187] The increment of the new differential coefficient can be obtained.

[0188]

[0189] The new proportional coefficient can be incremented. New integral coefficient increment New differential coefficient increment The optimized control parameters obtained in this round can be combined with new multi-source feature data to determine new pitch angle increments.

[0190] The optimized control parameters allow for real-time adjustment of the pitch control law, enabling rapid, stable, and low-energy correction of the tower attitude, thereby enhancing the system's adaptability and safety in complex sea conditions.

[0191] In some embodiments, the first weighting factor, the second weighting factor, the third weighting factor, and the fourth weighting factor described above can be determined in the following manner:

[0192] New multi-source feature data is used as state parameters input into a deep reinforcement learning model. The adjustment amounts of the first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are used as action parameters and the action parameters are output. The first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are periodically optimized and updated with the goal of maximizing the reward function that includes correction time, overshoot, and energy loss.

[0193] Specifically, a deep reinforcement learning model (DRL model) can be used, incorporating new multi-source feature data as state parameters, such as s. t =[θ avg ,σ z 2 PSD 0.1-1Hz ,ρ v,θ H wave (f)], which is then input into the DRL model. The DRL model outputs the adjustment amount of the weight factor as the action parameter, such as a t= [Δλ1, Δλ2, Δμ, Δη]. The DRL model can determine how to adjust the above weights by analyzing the current input state parameters. The model's reward function expression is as follows:

[0194]

[0195] in, This represents the average time required for each attitude correction over the past period; the shorter the time, the higher the reward. E represents the average overshoot during the correction process over the past period; the smaller the overshoot, the higher the reward. 损耗 In the past cycle, the energy loss caused by corrective actions (such as yaw power consumption) is considered. The lower the energy loss, the higher the reward. w1, w2, and w3 are manually set weights used to weigh the importance of the three objectives.

[0196] The specific learning process of the DRL model is as follows:

[0197] 1. The model observes the current state s. t (For example: the energy of tower swaying is very high).

[0198] 2. The model outputs an action a. t (For example: it is recommended to increase the value of λ2, because λ2 is related to PSD).

[0199] 3. This action is applied, and the new weights are used in the basic control layer for 24 hours.

[0200] 4. After 24 hours, calculate the reward R based on the average correction time, overshoot, and energy consumption during this period. t .

[0201] 5. The model is based on the obtained reward R. t This is used to update the parameters of its internal neural network, thereby learning "in state s_t, output action a". t A strategy that can yield high rewards.

[0202] 6. Through repeated iterations, the model continuously optimizes itself and eventually learns an optimal strategy that can automatically adjust the weight factors under different sea conditions.

[0203] By periodically optimizing and updating the first, second, third, and fourth weighting factors through the above methods, the wind turbine main control system achieves self-learning and self-evolution, enabling it to adapt to long-term, complex, and ever-changing marine environments and find the globally optimal control parameter adjustment strategy, thereby comprehensively improving the speed, stability, and economy of the correction process.

[0204] S104: Based on the optimized control parameters and new multi-source feature data, determine the new pitch angle increment and perform attitude correction again until the tower attitude returns to normal.

[0205] In some embodiments, before performing the attitude correction again in S104 above, the following may also be included in the specific implementation:

[0206] The new yaw angle increment is determined based on the new bending moment distribution gradient and the new wind speed-tilt angle correlation coefficient.

[0207] Accordingly, the aforementioned S104, which involves performing another attitude correction, can, in practice, include:

[0208] The target tower is then re-attitude corrected based on the new pitch angle increment and the new yaw angle increment.

[0209] Once the tower's attitude has returned to normal after correction, attitude correction is stopped.

[0210] Specifically, the calculation methods for the new yaw angle increment (or new yaw angle increment) and the new pitch angle increment, the calculation method for the new pitch angle increment (or new pitch angle increment), and the correction formula for re-correcting the attitude of the target tower based on the new pitch angle increment and the new yaw angle increment are all similar to the corresponding formulas mentioned above.

[0211] The aforementioned new yaw angle increment (or new yaw angle increment) can be determined according to the following formula:

[0212]

[0213] The aforementioned new pitch angle increment (or new pitch angle increment) can be determined based on the optimized control parameters and new multi-source characteristic data, according to the following formula:

[0214]

[0215] The meanings of the specific parameters can be found in the previous text, and will not be repeated here.

[0216] The above-mentioned attitude correction of the target tower is performed again based on the new pitch angle increment and the new yaw angle increment. The attitude correction is as follows:

[0217] M 修正 =K1·C L (β+Δβ')+K2·(C D (β+Δβ'))+K3·Δγ'

[0218] Among them, M 修正 This is the second correction torque (or secondary correction torque).

[0219] After a second or subsequent attitude correction, it can be determined whether the second attitude correction is effective. This requires continued monitoring of the tower's attitude to see if it remains abnormal. If it is still abnormal, the second attitude correction is ineffective, and further optimization of control parameters is needed to ensure the correction effect. If the tower's attitude returns to normal, the second attitude correction is effective, and no further optimization of control parameters is necessary; at this point, attitude correction should be stopped. The attitude correction process after a second attitude correction can be found in the second attitude correction process; this manual will not elaborate on it further.

[0220] Based on the above embodiments, the present invention can achieve the following beneficial effects:

[0221] 1. Improve monitoring accuracy and reliability: By developing a dual-mode data acquisition strategy specifically for floating wind turbine towers and installing high-precision multi-type sensors on key parts of the target tower, the attitude data of the target tower under different operating conditions can be acquired more accurately, enabling real-time and precise monitoring of the target tower's attitude and providing a reliable basis for subsequent attitude correction.

[0222] 2. Achieve rapid and effective attitude correction: By closely integrating monitoring results with the wind turbine main control system, when the tower attitude is detected to be out of limit or the swaying shows an upward trend (abnormal attitude), the wind turbine main control system can promptly control the wind turbine pitch and yaw to quickly correct the target tower attitude. Through the control closed-loop mechanism, the control parameters are continuously optimized to ensure that the target tower attitude is always kept within a safe range, effectively reducing the operating risk of floating wind turbine units and improving the stability and reliability of the units.

[0223] 3. Enhanced adaptability and maintainability of the unit: The method of this invention enables floating wind turbine units to better adapt to complex marine environments, reduce equipment failures and damage caused by abnormal tower attitude, and extend the service life of the unit. Simultaneously, through real-time monitoring and timely correction, the difficulty and cost of unit maintenance can be reduced, and maintenance efficiency improved.

[0224] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. For details, please refer to the foregoing descriptions of the relevant processing embodiments; they will not be repeated here.

[0225] The foregoing description of this method is for illustrative purposes only and describes specific embodiments. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps described in the claims may be performed in a different order than those shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0226] The above method will be described below with reference to specific embodiments.

[0227] See Figure 3 As shown, Figure 3 This is a diagram of the attitude monitoring and real-time control system architecture. The above method can be applied to a floating wind turbine tower attitude correction system, which specifically includes a perception layer (for tower and environmental monitoring), a transmission layer (for data synchronization and processing), a decision execution layer, and an execution layer (these two layers are the wind turbine main control system). The perception layer can deploy various types of sensors, such as a dual-axis tilt sensor and a triaxial accelerometer at the top of the tower, strain sensors and tilt sensors at the middle and bottom of the tower, and wave radar and current meters on the floating foundation platform of the tower. In the transmission layer, data can be transmitted to an edge computing gateway via 5G / fiber optics. The edge computing gateway can perform data preprocessing, noise reduction filtering, and timestamp synchronization on the attitude data collected by the various sensors in the perception layer, and finally store the data in a local server database. The decision execution layer can include a data fusion module and a control decision module. The data fusion module can include spatiotemporal alignment, feature extraction, wavelet transformation, LSTM trend prediction, etc., while the control decision module can include safety threshold judgment, sway trend analysis, pitch command generation, and cooperative control strategies. The decision execution layer can retrieve preprocessed attitude data from the transport layer's database, extract multi-source feature data from this attitude data, and then compare the multi-source feature data with corresponding preset safety thresholds to determine whether the tower attitude is abnormal. If abnormal, it generates pitch and yaw commands and sends them to the execution layer. The execution layer can receive the commands output by the decision execution layer and control the pitch and yaw systems to achieve rapid and accurate correction of the tower attitude.

[0228] See Figure 4 As shown, Figure 4 This is a flowchart for attitude monitoring and real-time control. The specific steps are as follows:

[0229] Sensor sampling: A dual-mode strategy of "high-frequency sampling + event triggering" is adopted: data is collected at a frequency of 10Hz in normal mode; when the wave height exceeds 2m or the wind speed exceeds 15m / s, it switches to 50Hz high-frequency sampling.

[0230] Data preprocessing: The collected raw data is processed by wavelet denoising and temperature compensation to improve the accuracy of the data. Then, the processed data is uploaded to the edge computing node in real time via the OPC UA protocol and stored in the time series database.

[0231] Feature extraction and judgment: Calculate the mean overturning angle θ avg σ z 2 Inclination rate of change Based on the characteristic parameters, it can be determined whether the tower attitude exceeds the limit. The overturning angle safety threshold can be set to ±5° and the sway energy threshold to 0.2g / Hz.

[0232] If the tower attitude exceeds the limit or the swaying shows an upward trend, the attitude correction control logic is triggered, which can execute pitch / yaw. For example, the wind turbine main control system controls the wind turbine pitch and yaw to correct the attitude based on the tower attitude deviation. By adjusting the pitch angle and yaw angle, the bending moment of the tower is adjusted so that the tower attitude is restored to a safe range. During the correction process, the tower attitude change is monitored in real time, and the control parameters are optimized to ensure the correction effect.

[0233] After attitude correction, attitude data is collected again to determine whether the attitude correction is effective. If it is, control is maintained; otherwise, adaptive optimization is performed based on PID parameters (i.e., optimization of control parameters), and attitude correction is performed again based on the optimized control parameters. The above process is repeated until the tower attitude returns to normal.

[0234] It should be noted that during attitude correction, the tower buckling critical load can be calculated in real time. When the real-time load exceeds 70% of the critical load, the emergency shutdown protection mechanism is triggered. During emergency shutdown, operations such as adjusting the pitch angle to 90° and locking the yaw brake are performed to ensure the safety of the unit.

[0235] In actual operation, the system can effectively monitor the attitude changes of the floating wind turbine tower and make timely corrections when the tower attitude is abnormal, ensuring the safe and stable operation of the unit. In a sea state with a wave height of 3m, the system successfully corrected the tower tilt angle from 6° to 3° in less than 30s, significantly reducing the operational risk of the unit.

[0236] Although this specification provides the following examples or appendices Figure 5The method or apparatus structure shown may include more or fewer combined operation steps or module units based on conventional or non-inventive methods. In steps or structures where there is no logically necessary causal relationship, the execution order of these steps or the module structure of the apparatus is not limited to the execution order or module structure shown in the embodiments or drawings of this specification. When the method or module structure is applied in actual devices, servers, or terminal products, it can be executed sequentially or in parallel according to the method or module structure shown in the embodiments or drawings (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed processing or server cluster implementation environment). Based on the above-described method for correcting the attitude of a floating wind turbine tower, this specification also proposes an embodiment of a device for correcting the attitude of a floating wind turbine tower. Figure 5 As shown, the device may specifically include the following modules:

[0237] The pitch angle increment and yaw angle increment determination module 501 can be used to determine the pitch angle increment and the yaw angle increment based on the multi-source feature data of the target tower when judging the tower attitude abnormality based on the multi-source feature data of the target tower.

[0238] The first attitude correction module 502 can be used to correct the attitude of the target tower according to the pitch angle increment and yaw angle increment, and the attitude correction is performed under safety boundary constraints.

[0239] The control parameter optimization module 503 can be used to optimize the current control parameters based on new multi-source feature data when the tower attitude is still abnormal after correction.

[0240] The second attitude correction module 504 can be used to determine a new pitch angle increment and perform attitude correction again based on the optimized control parameters and new multi-source feature data until the tower attitude returns to normal.

[0241] In some embodiments, the multi-source feature data in the above-mentioned pitch angle increment and yaw angle increment determination module 501 may include at least one of the following: mean overturning angle, settlement displacement variance, strain energy density, tilt angle change rate, sway spectrum energy in the 0.1-1Hz frequency band, wind speed and tilt angle correlation coefficient, bending moment distribution gradient, wave and tower motion transfer function value; correspondingly, the above-mentioned pitch angle increment and yaw angle increment determination module 501 may also be used to determine whether the mean overturning angle, settlement displacement variance, and strain energy density are greater than the corresponding preset safety threshold, and / or whether the tilt angle change rate, sway spectrum energy in the 0.1-1Hz frequency band, and wind speed and tilt angle correlation coefficient are greater than the corresponding preset safety threshold, and / or whether the bending moment distribution gradient and wave and tower motion transfer function value are greater than the corresponding preset safety threshold; if so, it is determined that the tower attitude of the target tower is abnormal.

[0242] In some embodiments, the current control parameters in the above-mentioned pitch angle increment and yaw angle increment determination module 501 may include the current proportional coefficient, the current integral coefficient, and the current derivative coefficient; correspondingly, the above-mentioned pitch angle increment and yaw angle increment determination module 501 can specifically be used to determine the overturning angle deviation based on the difference between the average overturning angle and the preset overturning angle safety threshold; and to determine the pitch angle increment according to the following formula based on the overturning angle deviation, the current proportional coefficient, the current integral coefficient, and the current derivative coefficient:

[0243]

[0244] Where Δβ is the pitch angle increment; θ err For overturning angle deviation; k p k is the current scaling factor. i k is the current integral coefficient. d θ is the current differential coefficient; t is time; ∫θ err dt represents the deviation of the overturning angle θ err Integrate over time t; Overturning angle deviation θ err The derivative with respect to time t;

[0245] The yaw angle increment is determined based on the bending moment distribution gradient and the correlation coefficient between wind speed and tilt angle.

[0246] In some embodiments, the first attitude correction module 502 described above can be specifically used to determine the correction torque according to the following formula based on the pitch angle increment and yaw angle increment:

[0247] M 修正 =K1·C L (β+Δβ)+K2·(C D (β+Δβ)+K3·Δγ

[0248] Among them, M修正 To correct the torque; K1 is the coupling coefficient of the first structure of the tower; K2 is the coupling coefficient of the second structure of the tower; C L C is the lift / drag coefficient of the first blade. D Δβ is the lift / drag coefficient of the second blade; β is the current actual pitch angle; Δβ is the pitch angle increment; Δγ is the yaw angle increment;

[0249] The attitude of the target tower is corrected based on the corrected torque.

[0250] In some embodiments, the safety boundary constraints in the first attitude correction module 502 are determined based on the tower buckling critical load calculated in real time. Specifically, the first attitude correction module 502 can also be used to limit the pitch rate and yaw rate during the attitude correction process when the measured tower buckling load is greater than the product of the tower buckling critical load and the first preset value.

[0251] In some embodiments, the first attitude correction module 502 described above can also be used to re-acquire attitude data of the target tower under different working conditions by using multiple types of sensors deployed at different locations on the target tower; perform data preprocessing on the re-acquired attitude data to extract new multi-source feature data from the preprocessed attitude data; and determine whether the tower attitude is still abnormal based on the new multi-source feature data after correction.

[0252] In some embodiments, the control parameter optimization module 503 can be specifically used to optimize the current proportional coefficient according to the following formula based on the new tilt angle change rate, the new sway spectrum energy in the 0.1-1Hz frequency band, the first weighting factor, and the second weighting factor, to obtain the proportional coefficient increment:

[0253]

[0254] Where ΔK p This represents the increment of the proportional coefficient; For the new rate of change of tilt angle; PSD 0.1-1Hz ' is the new 0.1-1Hz frequency band oscillation spectrum energy; λ1 is the first weighting factor; λ2 is the second weighting factor;

[0255] Based on the new mean overturning angle, calculate the new overturning angle deviation, and combine it with the third weighting factor to optimize the current integral coefficients according to the following formula, thus obtaining the integral coefficient increment:

[0256]

[0257] Where ΔK i θ is the increment of the integral coefficient; μ is the third weighting factor; θ err '(t) represents the new overturning angle deviation; t is any time point from 0 to t; The new overturning angle deviation θ at time point t, from 0 to t. err Integrate the absolute value of '(t);

[0258] Based on the new bending moment distribution gradient and the fourth weighting factor, the differential coefficients are optimized according to the following formula to obtain the differential coefficient increments:

[0259]

[0260] Where ΔK d η is the increment of the differential coefficient; η is the fourth weighting factor; represents the new bending moment distribution gradient; max is the maximum value.

[0261] In some embodiments, the control parameter optimization module 503 can also be used to input new multi-source feature data as state parameters into a deep reinforcement learning model, use the adjustment amounts of the first weight factor, the second weight factor, the third weight factor, and the fourth weight factor as action parameters and output the action parameters, and periodically optimize and update the first weight factor, the second weight factor, the third weight factor, and the fourth weight factor with the goal of maximizing the reward function that includes correction time, overshoot, and energy loss.

[0262] In some embodiments, the second attitude correction module 504 may be further used to determine a new yaw angle increment based on the new bending moment distribution gradient and the new wind speed and tilt angle correlation coefficient; correspondingly, the second attitude correction module 504 may be used to perform attitude correction on the target tower again based on the new pitch angle increment and the new yaw angle increment; when the tower attitude is determined to have returned to normal after correction, attitude correction is stopped.

[0263] As can be seen from the above, the floating wind turbine tower attitude correction device provided in the embodiments of this specification can achieve the purpose of monitoring the operating attitude of the floating wind turbine tower, correcting abnormal tower attitude, and reducing the operating risk of the floating wind turbine.

[0264] This specification also provides an electronic device based on the above-described method for correcting the attitude of a floating wind turbine tower, including a processor and a memory for storing processor-executable programs / instructions. Specifically, the processor can execute the following steps according to the program / instructions: when the tower attitude is determined to be abnormal based on multi-source feature data of the target tower, the pitch angle increment is determined based on the multi-source feature data and the current control parameters, and the yaw angle increment is determined based on the multi-source feature data; the attitude is corrected based on the pitch angle increment and the yaw angle increment, the attitude correction being performed under safety boundary constraints; if the tower attitude is still abnormal after correction, the current control parameters are optimized based on new multi-source feature data; based on the optimized control parameters and the new multi-source feature data, a new pitch angle increment is determined and the attitude is corrected again until the tower attitude returns to normal.

[0265] To execute the above instructions more accurately, please refer to... Figure 6 As shown in the embodiments of this specification, another specific electronic device is also provided, wherein the electronic device includes a network communication port 601, a processor 602 and a memory 603, and the above structures are connected by internal cables so that the various structures can perform specific data interaction.

[0266] Specifically, the processor 602 can be used to determine the pitch angle increment and yaw angle increment based on the multi-source feature data of the target tower when the tower attitude is determined to be abnormal based on the multi-source feature data of the target tower; to perform attitude correction on the target tower based on the pitch angle increment and yaw angle increment, wherein the attitude correction is performed under safety boundary constraints; if the tower attitude is still abnormal after correction, to optimize the current control parameters based on the new multi-source feature data; and to determine a new pitch angle increment and perform attitude correction again based on the optimized control parameters and the new multi-source feature data, until the tower attitude returns to normal.

[0267] The memory 603 can be used to store the corresponding instruction program.

[0268] In this embodiment, the network communication port 601 can be a virtual port bound to different communication protocols, thereby enabling the sending or receiving of different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0269] In this embodiment, the processor 602 can be implemented in any suitable manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc. This specification is not limiting.

[0270] In this embodiment, the memory 603 may include multiple layers. In a digital system, anything that can store binary data can be a memory. In an integrated circuit, a circuit with storage function but no physical form is also called a memory, such as RAM, FIFO, etc. In a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0271] This specification also provides a computer storage medium based on the above-described method for correcting the attitude of a floating wind turbine tower. The computer storage medium stores a computer program / instruction that, when executed, performs the following: when the tower attitude is determined to be abnormal based on multi-source feature data of the target tower, the pitch angle increment is determined based on the multi-source feature data and the current control parameters, and the yaw angle increment is determined based on the multi-source feature data; the attitude is corrected based on the pitch angle increment and the yaw angle increment, the attitude correction being performed under safety boundary constraints; if the tower attitude is still abnormal after correction, the current control parameters are optimized based on new multi-source feature data; based on the optimized control parameters and the new multi-source feature data, a new pitch angle increment is determined and the attitude is corrected again until the tower attitude returns to normal.

[0272] In this embodiment, the storage medium includes, but is not limited to, Random Access Memory (RAM), Read-Only Memory (ROM), cache, hard disk drive (HDD), or memory card. The memory can be used to store computer program instructions. The network communication unit can be an interface configured according to standards specified in the communication protocol for network connection communication.

[0273] In this embodiment, the specific functions and effects implemented by the program instructions stored in the computer storage medium can be explained in comparison with other implementation methods, and will not be repeated here.

[0274] This specification also provides a computer program product based on the above-described method for correcting the attitude of a floating wind turbine tower. The product includes a non-transitory computer-readable storage medium storing computer programs / instructions. These computer programs / instructions are operable to cause the computer to perform the following steps: when the tower attitude is determined to be abnormal based on multi-source feature data of the target tower, the pitch angle increment is determined based on the multi-source feature data and current control parameters, and the yaw angle increment is determined based on the multi-source feature data; the attitude is corrected based on the pitch angle increment and yaw angle increment, the attitude correction being performed under safety boundary constraints; if the tower attitude is still abnormal after correction, the current control parameters are optimized based on new multi-source feature data; based on the optimized control parameters and new multi-source feature data, a new pitch angle increment is determined and the attitude is corrected again until the tower attitude returns to normal.

[0275] While this specification provides the steps of operation for the methods described in the embodiments or flowcharts, more or fewer steps may be included based on conventional or non-inventive means. The order of steps listed in the embodiments is merely one possible order of execution among many steps and does not represent the only possible order. In actual device or client product execution, the methods shown in the embodiments or drawings may be executed sequentially or in parallel (e.g., in a parallel processor or multi-threaded processing environment, or even a distributed data processing environment). The terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, product, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, product, or apparatus. Without further limitations, the presence of other identical or equivalent elements in a process, method, product, or apparatus that includes said elements is not excluded. The terms "first," "second," etc., are used to denote names and do not indicate any particular order.

[0276] Those skilled in the art will also know that, besides implementing the controller using purely computer-readable program code, the same functions can be achieved by logically programming the method steps, making the controller function as logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers (PLCs), and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the devices within it used to implement various functions can also be considered structures within that hardware component. Alternatively, the devices used to implement various functions can be considered as both software modules implementing the method and structures within a hardware component.

[0277] This specification can be described in the general context of computer-executable instructions that are executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, classes, etc., that perform a specific task or implement a specific abstract data type. This specification can also be practiced in distributed computing environments, where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0278] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this specification can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions of this specification can essentially be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, mobile terminal, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments of this specification.

[0279] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on its differences from other embodiments. This specification can be used in numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable electronic devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices, etc.

[0280] Although this specification has been described by way of examples, those skilled in the art will recognize that many variations of this specification are possible without departing from its spirit, and it is intended that the appended claims cover such variations without departing from the spirit of this specification.

Claims

1. A method for correcting the attitude of a floating wind turbine tower, characterized in that, include: When judging the tower attitude anomaly based on the multi-source feature data of the target tower, the pitch angle increment is determined according to the multi-source feature data and the current control parameters, and the yaw angle increment is determined according to the multi-source feature data. The target tower is attitude-corrected based on the pitch angle increment and yaw angle increment, and the attitude correction is performed under safety boundary constraints. If the tower attitude is still abnormal after correction, the current control parameters are optimized based on new multi-source feature data. Based on the optimized control parameters and new multi-source feature data, a new pitch angle increment is determined and attitude correction is performed again until the tower attitude returns to normal. The optimization of the current control parameters based on the new multi-source feature data includes: Based on the new tilt rate of change, the new sloshing spectrum energy in the 0.1-1Hz frequency band, the first weighting factor, and the second weighting factor, the current scaling factor is optimized according to the following formula to obtain the scaling factor increment: ； in, This represents the increment of the proportional coefficient; For the new rate of change of tilt angle; For the new 0.1-1Hz frequency band of oscillation spectrum energy; It is the first weighting factor; As the second weighting factor; Based on the new mean overturning angle, calculate the new overturning angle deviation, and combine it with the third weighting factor to optimize the current integral coefficients according to the following formula, thus obtaining the integral coefficient increment: in, This represents the increment of the integral coefficient; It is the third weighting factor; For the new overturning angle deviation; For any point in time from 0 to t; For the time point from 0 to t, New overturning angle deviation Integrate by the absolute value; Based on the new bending moment distribution gradient and the fourth weighting factor, the differential coefficients are optimized according to the following formula to obtain the differential coefficient increments: in, This represents the increment of the differential coefficient; It is the fourth weighting factor; This represents the new bending moment distribution gradient; max is the maximum value. The first weighting factor, the second weighting factor, the third weighting factor, and the fourth weighting factor are determined in the following manner: New multi-source feature data is used as state parameters input into a deep reinforcement learning model. The adjustment amounts of the first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are used as action parameters and the action parameters are output. The first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are periodically optimized and updated with the goal of maximizing the reward function that includes correction time, overshoot, and energy loss.

2. The method according to claim 1, characterized in that, The multi-source feature data includes at least one of the following: mean overturning angle, variance of settlement displacement, strain energy density, rate of change of tilt angle, sway spectrum energy in the 0.1-1Hz frequency band, correlation coefficient between wind speed and tilt angle, bending moment distribution gradient, wave and tower motion transfer function values. Accordingly, the method further includes: Determine whether the mean overturning angle, variance of settlement displacement, and strain energy density are greater than the corresponding preset safety threshold, and / or whether the tilt angle change rate, sway spectrum energy in the 0.1-1Hz frequency band, and correlation coefficient between wind speed and tilt angle are greater than the corresponding preset safety threshold, and / or whether the bending moment distribution gradient and wave and tower motion transfer function values ​​are greater than the corresponding preset safety threshold. If so, determine that the target tower's tower attitude is abnormal.

3. The method according to claim 2, characterized in that, The current control parameters include the current proportional coefficient, the current integral coefficient, and the current derivative coefficient; correspondingly, determining the pitch angle increment based on multi-source feature data and the current control parameters includes: The overturning angle deviation is determined based on the difference between the mean overturning angle and the preset overturning angle safety threshold. Based on the overturning angle deviation, the current proportional coefficient, the current integral coefficient, and the current differential coefficient, determine the pitch angle increment according to the following formula: in, This is the pitch angle increment; For overturning angle deviation; k p k is the current scaling factor. i k is the current integral coefficient. d The current differential coefficient; t is time; To account for overturning angle deviation Integrate over time t; Overturning angle deviation The derivative with respect to time t; The step of determining the yaw angle increment based on multi-source feature data includes: The yaw angle increment is determined based on the bending moment distribution gradient and the correlation coefficient between wind speed and tilt angle.

4. The method according to claim 1, characterized in that, The attitude correction of the target tower based on the pitch angle increment and yaw angle increment includes: Based on the pitch angle increment and yaw angle increment, the correction torque is determined according to the following formula: Among them, M 修正 To correct the torque; K1 is the coupling coefficient of the first structure of the tower; K2 is the coupling coefficient of the second structure of the tower; C L C is the lift / drag coefficient of the first blade. D The lift / drag coefficient for the second blade; This represents the current actual pitch angle; This is the pitch angle increment; This is the yaw angle increment; The attitude of the target tower is corrected based on the corrected torque.

5. The method according to claim 1, characterized in that, The safety boundary constraints are determined based on the tower buckling critical load calculated in real time; Accordingly, the attitude correction is performed under safety boundary constraints, including: When the measured tower buckling load is greater than the product of the tower buckling critical load and the first preset value, the pitch rate and yaw rate during the attitude correction process are limited.

6. The method according to claim 1, characterized in that, The method further includes: By deploying multiple types of sensors at different locations on the target tower, attitude data of the target tower under different working conditions is re-acquired. The re-acquired attitude data is preprocessed to extract new multi-source feature data from the preprocessed attitude data. After correction, the tower attitude is determined based on the new multi-source feature data to determine whether it is still abnormal.

7. The method according to claim 1, characterized in that, The method further includes: The new yaw angle increment is determined based on the new bending moment distribution gradient and the new wind speed-tilt angle correlation coefficient. Accordingly, the further attitude correction includes: The target tower is then re-attitude corrected based on the new pitch angle increment and the new yaw angle increment. Once the tower's attitude has returned to normal after correction, attitude correction is stopped.

8. A device for correcting the attitude of a floating wind turbine tower, characterized in that, include: The pitch angle increment and yaw angle increment determination module is used to determine the pitch angle increment based on the multi-source feature data of the target tower and the current control parameters when judging the tower attitude abnormality based on the multi-source feature data of the target tower; and to determine the yaw angle increment based on the multi-source feature data. The first attitude correction module is used to correct the attitude of the target tower according to the pitch angle increment and yaw angle increment, and the attitude correction is performed under safety boundary constraints. The control parameter optimization module is used to optimize the current control parameters based on new multi-source feature data when the tower attitude is still abnormal after correction. The second attitude correction module is used to determine the new pitch angle increment based on the optimized control parameters and new multi-source feature data, and then perform attitude correction again until the tower attitude returns to normal. The optimization of the current control parameters based on the new multi-source feature data includes: Based on the new tilt rate of change, the new sloshing spectrum energy in the 0.1-1Hz frequency band, the first weighting factor, and the second weighting factor, the current scaling factor is optimized according to the following formula to obtain the scaling factor increment: ； in, This represents the increment of the proportional coefficient; For the new rate of change of tilt angle; For the new 0.1-1Hz frequency band of oscillation spectrum energy; It is the first weighting factor; As the second weighting factor; Based on the new mean overturning angle, calculate the new overturning angle deviation, and combine it with the third weighting factor to optimize the current integral coefficients according to the following formula, thus obtaining the integral coefficient increment: in, This represents the increment of the integral coefficient; It is the third weighting factor; For the new overturning angle deviation; For any point in time from 0 to t; For the time point from 0 to t, New overturning angle deviation Integrate by the absolute value; Based on the new bending moment distribution gradient and the fourth weighting factor, the differential coefficients are optimized according to the following formula to obtain the differential coefficient increments: in, This represents the increment of the differential coefficient; It is the fourth weighting factor; This represents the new bending moment distribution gradient; max is the maximum value. The first weighting factor, the second weighting factor, the third weighting factor, and the fourth weighting factor are determined in the following manner: New multi-source feature data is used as state parameters input into a deep reinforcement learning model. The adjustment amounts of the first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are used as action parameters and the action parameters are output. The first weight factor, the second weight factor, the third weight factor, and the fourth weight factor are periodically optimized and updated with the goal of maximizing the reward function that includes correction time, overshoot, and energy loss.

9. An electronic device, characterized in that, include: A memory and a processor, the processor and the memory being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to implement the steps of the method according to any one of claims 1 to 7.

10. A computer storage medium, characterized in that, The computer storage medium stores computer program instructions, which, when executed, implement the steps of the method according to any one of claims 1 to 7.

11. A computer program product, characterized in that, It includes a computer program that, when executed by a processor, implements the steps of the method according to any one of claims 1 to 7.

Images