A wind turbine load rapid evaluation method for wind farm coordinated regulation

Abstract

The application discloses a wind turbine load rapid evaluation method for wind farm collaborative regulation, and relates to the technical field of wind turbine safety regulation. The specific implementation scheme is as follows: standard pulse instructions are applied to the wind turbine to extract vibration characteristic parameters and load oscillation initial amplitude; based on the rate of the candidate regulation instruction and the environmental wind speed, a transient excitation intensity function is set and effective excitation equivalent is calculated; according to the ratio of the effective excitation equivalent and the reference excitation equivalent, the relative impact intensity and the load oscillation amplitude are determined; the envelope function is constructed by combining the vibration characteristic parameters of the wind turbine, and the earliest safety execution time required for attenuation to the safety threshold is solved; the executability of the regulation instruction is determined by comparing with the maximum waiting time set by the rated blade bending moment load, and the load evaluation is completed; the collaborative controller can dynamically and accurately arrange the instruction sequence, and the overall safety of the wind turbine is improved.

Description

Technical Field

[0001] This application relates to the field of wind turbine safety control technology, specifically to a method for rapid load assessment of wind turbines for coordinated control of wind farms. Background Technology

[0002] In the field of modern wind power generation, as the scale of wind farms continues to expand, the coordinated control of wind turbine units has become increasingly important. Effective wind turbine load assessment can not only improve power generation efficiency, but also extend the service life of equipment and reduce maintenance costs. In the coordinated control of wind farms, in order to respond to grid dispatch and execute rapid pitch limiting power, the control command will excite transient load oscillation of the unit, that is, the blade flapping, tower front and rear modes will show periodic fluctuations with gradually decreasing amplitude within a few seconds to tens of seconds after the command ends.

[0003] However, existing technologies mainly rely on static models and empirical formulas for wind turbine load monitoring and evaluation, resulting in a lack of sensitivity and real-time performance to load changes under dynamic operating conditions. This makes it difficult to effectively cope with rapidly changing wind farm environments. In addition, to avoid load exceeding limits during transient oscillations, existing engineering practices use a fixed large safety factor to limit the rate or amplitude of control commands, which leads to an excessive sacrifice of the response speed of wind farm coordinated control and the inability to fully realize the potential for increasing power generation. This deficiency leads to the collaborative controller either blindly waiting for a fixed and conservative time, sacrificing response speed, or risking issuing subsequent commands, which may cause oscillations to superimpose, load accumulation to exceed limits, and trigger blade fatigue or tower resonance. The shortcoming of existing technologies is that they cannot accurately and quickly assess the load status of wind turbines under different control commands, especially when wind field conditions change drastically. The lag and inflexibility of traditional methods significantly reduce their application effectiveness.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] In view of this, the purpose of this application is to provide a rapid load assessment method for wind turbine units for coordinated control of wind farms, which can accurately arrange command sequences, improve the overall safety of wind turbine units, and thus improve the power generation efficiency and economic benefits of wind farms.

[0006] According to a first aspect of this application, a method for rapid load assessment of wind turbine generators for coordinated wind farm control is provided, comprising: A standard pulse-shaped control command is applied to each wind turbine. For any wind turbine, the oscillation decay signal of the wind turbine is collected after the control command ends, so as to extract the vibration characteristic parameters of the wind turbine and the initial amplitude of the load oscillation corresponding to the standard pulse command. The candidate control commands issued by the current wind farm coordinating controller are obtained, the duration of the candidate control command and the command rate at each moment are extracted, and an instantaneous excitation intensity function is set based on the command rate and the ambient wind speed. The excitation intensity function is superimposed over the duration of the command to obtain the effective excitation equivalent of the candidate control command. The relative impact intensity is determined based on the ratio of the effective excitation equivalent of the candidate control command to the pre-stored reference excitation equivalent of the standard pulse command. Based on this relative impact intensity and the initial amplitude of the load oscillation, the amplitude of the load oscillation excited by the candidate control command is determined. Based on the load oscillation amplitude of the candidate control command and combined with the vibration characteristic parameters of the wind turbine, the envelope function of the load oscillation amplitude of each wind turbine under the candidate control is constructed. For any wind turbine, the earliest safe execution time required for the envelope to decay to the preset safety threshold when executing the candidate control command is solved. Based on the rated blade bending moment load of each wind turbine, the maximum waiting time for each wind turbine is set. The earliest safe execution time of each wind turbine is compared with the maximum waiting time. Based on the comparison results, the feasibility judgment result of the candidate control command on the corresponding wind turbine is obtained.

[0007] In some embodiments, the standard pulse shape control command is the standard yaw rate command of the wind turbine. The oscillation attenuation signal of the wind turbine is the time sequence of the blade root flapping moment of each blade of the wind turbine, and the vibration characteristic parameters include the damping ratio and undamped natural frequency of the wind turbine.

[0008] In some embodiments, the specific method for obtaining the initial amplitude of the load oscillation corresponding to the standard pulse command and the vibration characteristic parameters of the wind turbine is as follows: For any wind turbine, based on the blade root flapping moment time sequence of any blade, the blade root flapping moment time sequence of the blade is high-pass filtered to obtain the filtered signal. The maximum value of the absolute peak value is determined in the filtered signal, and the maximum value of the absolute peak value is recorded as the initial amplitude of the load oscillation of the blade corresponding to the standard pulse command. Iterate through all blades of the wind turbine and take the initial amplitude of the average load oscillation corresponding to the standard pulse command for each blade as the initial amplitude of the load oscillation corresponding to the standard pulse command for the wind turbine.

[0009] In some embodiments, the method for obtaining the vibration characteristic parameters of the wind turbine is as follows: For any blade, the maximum absolute peak value and several subsequent peak values ​​in the same direction are extracted from the blade root flapping moment time series to form a decay peak sequence. Based on the amplitude ratio of adjacent peaks in the attenuation peak sequence, the logarithmic reduction, which characterizes the degree of vibration attenuation, is calculated, and the damping ratio of the blade is determined based on a preset damping conversion relationship. Based on the time interval between adjacent peaks in the attenuation peak sequence, the damped natural frequency of the blade is determined, and the damped natural frequency is converted by combining the damping ratio to obtain the undamped natural frequency of the blade.

[0010] In some embodiments, the candidate control command specifically refers to the yaw rate command of each wind turbine in the wind farm; the logic for setting the instantaneous excitation intensity function is as follows: extract the duration of the candidate control command and the command rate at each moment; for any wind turbine, within the duration of the candidate control command, determine the active power of the wind turbine after the corresponding control at different moments, and the average wind speed at the hub height of the wind turbine, wherein the average wind speed specifically refers to the historical average wind speed within a preset detection window; map the historical average wind speed to different moments within the candidate control command as the ambient wind speed at the corresponding moment; based on the ambient wind speed and active power at the corresponding moment, calculate the wind energy utilization coefficient of the wind turbine at the corresponding moment; and record the ratio of the wind energy utilization coefficient to the maximum wind energy utilization coefficient of the wind turbine as the relative aerodynamic efficiency of the wind turbine at the corresponding moment. The ratio of the ambient wind speed to the preset rated wind speed at the corresponding moment of the wind turbine is normalized to obtain the basic wind speed term. The product of the basic wind speed term and the relative aerodynamic efficiency is used as the joint weighting factor of the wind turbine at the corresponding moment. The joint weighting factor is multiplied by the absolute value of the yaw rate of the candidate control command at the corresponding moment to obtain the instantaneous excitation intensity function of the wind turbine.

[0011] In some embodiments, calculating the wind energy utilization coefficient of any wind turbine at the time corresponding to the candidate control command specifically includes: The active power of the wind turbine after the corresponding regulation is obtained, and the theoretical wind energy input value at the corresponding time is determined by combining the ambient wind speed, air density and rotor sweep area at the corresponding time. Based on the correspondence between the active power and the theoretical wind energy input value, the wind energy utilization coefficient of the wind turbine at the corresponding moment is calculated to characterize the aerodynamic energy conversion efficiency of the wind turbine under the current operating conditions. The active power of the wind turbine after corresponding regulation is predicted based on a deep learning model. The input of the deep learning model is the ambient wind speed, wind direction, candidate regulation command and corresponding yaw angle at the corresponding time, and the output is the active power of the candidate regulation command at the corresponding time.

[0012] In some embodiments, when calculating the effective excitation equivalent of a candidate control command, for any wind turbine, the temporal forgetting coefficient is determined based on the undamped natural frequency of the wind turbine, and the exponential forgetting weight corresponding to each discrete moment is determined based on the time interval between each discrete moment during the duration of the candidate control command and the end moment of the candidate control command. The instantaneous excitation intensity function value at each discrete moment is weighted with the corresponding exponential forgetting weight to obtain the weighted excitation value at each discrete moment. The weighted excitation values ​​at all discrete moments are then integrated and accumulated to obtain the effective excitation equivalent of the candidate control command for the wind turbine. The exponential forgetting weight is used to characterize the contribution of instantaneous excitation at different discrete moments during the duration of the candidate control command to the load oscillation response. The closer the discrete moment is to the end of the candidate control command, the greater the corresponding exponential forgetting weight.

[0013] In some embodiments, candidate control commands are replaced with standard pulse commands, and the same weighted integral method as that used for candidate control commands is employed to calculate the reference effective excitation equivalent of the standard pulse commands for the wind turbine. The relative impact intensity corresponding to the candidate control command is determined based on the effective excitation equivalent of the candidate control command and the reference effective excitation equivalent of the standard pulse command. The relative impact intensity is used to characterize the excitation impact level of the candidate control command relative to the standard pulse command. The relative impact intensity is positively correlated with the effective excitation equivalent of the candidate control command and negatively correlated with the reference effective excitation equivalent of the standard pulse command.

[0014] In some embodiments, the logic for determining the earliest safe execution time is as follows: Based on the relative impact intensity, the initial amplitude of the load oscillation corresponding to the standard pulse command, and the preset safety factor, the amplitude of the load oscillation excited by the candidate control command is determined. Based on the measured leaf root bending moment and its variation characteristics at the current moment, determine the equivalent amplitude corresponding to the current oscillation state; The equivalent amplitude is superimposed with the load oscillation amplitude excited by the candidate control command, and compared with a preset safety threshold. If the sum of the results is less than the preset security threshold, then the earliest safe execution time is determined to be 0. If the superposition result is greater than or equal to the preset safety threshold, an amplitude attenuation envelope is constructed based on the damping ratio and undamped natural frequency of the wind turbine, and the earliest safe execution time is determined according to the time it takes for the amplitude attenuation envelope to decay to the preset safety threshold.

[0015] In some embodiments, the maximum waiting time of each wind turbine is proportional to its rated blade bending moment load; For any wind turbine, if its earliest safe execution time is greater than 0 and less than the corresponding maximum waiting time, then the candidate control command is executed after delaying the earliest safe execution time. If the earliest safe execution time exceeds the corresponding maximum waiting time, the candidate control command is determined to be unexecutable on the wind turbine.

[0016] One embodiment of the above application has the following advantages or beneficial effects: This invention, by applying a standard pulse shape control command and acquiring the oscillation attenuation signal of the wind turbine, can accurately extract the vibration characteristic parameters and initial amplitude of the load oscillation of the wind turbine; using this as a benchmark, it provides important basic data for subsequent load assessment, making the assessment process real-time and accurate. By calculating the relative impact intensity and further analyzing the initial amplitude of load oscillations, this scheme can accurately determine the initial amplitude of load oscillations excited by candidate control commands. The resulting envelope function can provide a clear load decay curve for each wind turbine. Without requiring full time-domain simulation, it can quickly predict the decay envelope function of transient load oscillations excited by control commands of arbitrary shapes and directly output the safe waiting time. This allows the collaborative controller to dynamically and accurately arrange the command sequence, improving the overall safety of the wind turbine. Ultimately, it yields the load changes and the earliest safe execution time when executing candidate control commands, thereby improving the power generation efficiency and economic benefits of the wind farm.

[0017] Other effects of the above-mentioned alternative methods will be described below in conjunction with specific embodiments. Attached Figure Description

[0018] The accompanying drawings are provided for a better understanding of this solution and do not constitute a limitation of this application. Wherein: Figure 1 This is a schematic diagram of a method for rapid load assessment of wind turbine generators provided in an embodiment of this application; Figure 2 This is a schematic diagram of active power prediction error provided in an embodiment of this application. Detailed Implementation

[0019] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0020] To facilitate understanding of this application, the embodiments of this application will be briefly described below: In the scenario of coordinated control of wind farms, it is necessary to quickly assess the load status of wind turbines after receiving control commands to ensure their safe operation and performance optimization in variable wind speed environments. For example, when wind farms are controlled, wind turbines need to monitor their oscillation attenuation signals in real time according to the applied standard pulse shape commands in order to extract vibration characteristic parameters and the initial amplitude of load oscillation.

[0021] See Figure 1 This is a flowchart illustrating a method for rapid load assessment of wind turbine units for coordinated wind farm control, provided in an embodiment of this application. Figure 1 The execution subject of the method shown can be a combination of software and / or hardware, specifically, it can be one or more of various types of terminals, hardware systems, cloud computing, etc.

[0022] Figure 1 The method for rapid load assessment of wind turbines for coordinated wind farm control, as shown, includes steps S101 to S105, as detailed below: S101: Apply a standard pulse shape control command to each wind turbine. For any wind turbine, collect the oscillation decay signal of the wind turbine after the control command ends, so as to extract the vibration characteristic parameters of the wind turbine and the initial amplitude of the load oscillation corresponding to the standard pulse command.

[0023] The standard pulse shape control command is the standard yaw rate command of the wind turbine; a trapezoidal velocity curve is selected as the standard pulse; the trapezoidal velocity curve has finite jerk, is the most commonly used in engineering and is easy to reproduce; at the same time, its spectral energy is concentrated in the low frequency band, which can effectively excite the main mode of the unit; The standard yaw rate command is typically set by increasing linearly from 0 to... The rise time is set to 0.5-1 second, maintained at a constant speed for 1 second, and then linearly reduced to 0. The fall time is correspondingly set to 0.5-1 second. The entire instruction duration is between 1.5-3 seconds. It is the median value of typical yaw rate, which can produce measurable oscillation amplitude without triggering protection shutdown, and is suitable for most wind turbine units.

[0024] The oscillation attenuation signal of the wind turbine is the time sequence of the blade root flapping moment of each blade of the wind turbine, and the vibration characteristic parameters include the damping ratio and undamped natural frequency of the wind turbine. Under normal operating conditions and relatively stable wind speeds, such as wind speeds in the range of 8 to 10 m / s and turbulence intensity less than 0.1, execute a standard yaw rate command once. Use the unit's built-in load sensors or a high-speed SCADA (Supervisory and Data Acquisition) system to record the time series of blade root flapping moments at a sampling rate of not less than 50 Hz, and simultaneously record the hub-average wind speed during the command execution period. It should be noted that the wind turbine's built-in load sensors are typically installed at the blade root to monitor changes in blade root moment in real time. At the same time, a high-speed SCADA system is configured to record data at a sampling rate of not less than 50 Hz to ensure sufficient temporal resolution, especially during wind speed changes or command execution. In wind turbine units, high-precision load sensors are typically used to monitor changes in blade root bending moment. For example, the Kistler 9121 is a high-performance dynamic force sensor suitable for measuring dynamic loads and providing a high-frequency response; or the HBM U10C is a high-precision torque sensor that can be used to monitor changes in blade root bending moment, with good linearity and sensitivity.

[0025] The specific method for obtaining the initial amplitude of load oscillation corresponding to the standard pulse command, as well as the vibration characteristic parameters of the wind turbine, is as follows: For any wind turbine, based on the blade root flapping moment time sequence of any blade, the blade root flapping moment time sequence of the blade is high-pass filtered to obtain the filtered signal. The maximum value of the absolute peak value is determined in the filtered signal, and the maximum value of the absolute peak value is recorded as the initial amplitude of the load oscillation of the blade corresponding to the standard pulse command. It should be noted that high-pass filters can be implemented in various ways, commonly including digital filters such as Butterworth filters, Chebyshev filters, and FIR (Finite Impulse Response) filters; the appropriate filter type should be selected based on the characteristics of the signal and the application requirements. Simultaneously, select the cutoff frequency, specifically determined based on the signal's spectral characteristics. The cutoff frequency should be chosen above the frequencies of important components in the signal to remove low-frequency noise. Next, determine the filter order; a higher order results in a more significant filtering effect, but may also introduce phase delay. Design a high-pass filter based on the selected filter type and parameters. The designed high-pass filter is applied to the original blade root flapping moment time series signal; an appropriate filtering method is used to ensure that the phase of the signal is not distorted during the filtering process, generally a zero-phase filtering method is adopted.

[0026] Iterate through all blades of the wind turbine and take the initial amplitude of the average load oscillation corresponding to the standard pulse command for each blade as the initial amplitude of the load oscillation corresponding to the standard pulse command for the wind turbine.

[0027] The specific method for obtaining the vibration characteristic parameters of the wind turbine is as follows: For any blade, the maximum absolute peak value and several subsequent peak values ​​in the same direction are extracted from the blade root flapping moment time series to form a decay peak sequence. Based on the amplitude ratio of adjacent peaks in the attenuation peak sequence, the logarithmic reduction, which characterizes the degree of vibration attenuation, is calculated, and the damping ratio of the blade is determined based on a preset damping conversion relationship; the specific formula used to calculate the damping ratio is as follows: In the formula, Let be the damping ratio of the q-th blade. This represents the logarithmic reduction of the q-th blade, where q is the index of the wind turbine blade. It should be noted that the damping ratio is a dimensionless value used to describe the degree of energy dissipation in a system; it is the ratio between the damping force and the critical damping force in a vibrating system, reflecting the system's ability to suppress vibration; the formula is based on vibration theory, in which the behavior of vibration amplitude decaying with time in a vibrating system is described by an exponential function; when the system is subjected to damping, the vibration amplitude decays exponentially, therefore logarithmic decrease is used to quantify the damping ratio; Among them, the logarithmic decrease of the leaf The specific formula used for the calculation is as follows: In the formula, This represents the i-th peak value in the attenuation peak sequence of the q-th blade. Let be the (i-1)th peak in the q-th blade attenuation peak sequence, n be the total number of peaks in the blade attenuation peak sequence, and i be the index of the peak in the blade attenuation peak sequence; It should be noted that logarithmic attenuation reflects the degree of amplitude decay between adjacent peaks. By taking the logarithm, the amplitude ratio is converted into an additive operation, which simplifies the calculation of the attenuation of multiple peaks. When the system is subjected to an impact, the vibration will gradually decay, and the amplitude of the peak will gradually decrease. At the same time, in the vibration system, the behavior of vibration amplitude decaying with time is described by an exponential function. Therefore, logarithmic attenuation is set to quantify the degree of decay.

[0028] Based on the time interval between adjacent peaks in the attenuation peak sequence, the damped natural frequency of the blade is determined, and the damped natural frequency is converted using the damping ratio to obtain the undamped natural frequency of the blade; the specific formula used to calculate the undamped natural frequency is as follows: In the formula, Let be the undamped natural frequency of the q-th blade. Let q be the damped natural frequency of the q-th blade; It should be noted that the formula originates from the damped vibration model in vibration theory, which describes the relationship between damped and undamped systems. In actual vibration systems, damping leads to a decrease in vibration frequency. The damping ratio is a quantity that describes the degree of energy loss in the system. The larger the damping ratio, the more significant the energy loss of the system and the lower the vibration frequency. Therefore, the influence of damping on the natural frequency should be considered.

[0029] This includes the damped natural frequency. The specific formula used for the calculation is as follows: In the formula, is the average time interval between adjacent peaks in the attenuation peak sequence of the q-th blade; For any wind turbine, the damping ratio and undamped natural frequency of the wind turbine are characterized by the average damping ratio and average undamped natural frequency of its blades.

[0030] S102: Obtain the candidate control command issued by the current wind farm coordinating controller, extract the duration of the candidate control command and the command rate at each moment, set the instantaneous excitation intensity function based on the command rate and the ambient wind speed, and superimpose the excitation intensity function over the duration of the command to obtain the effective excitation equivalent of the candidate control command.

[0031] The candidate control commands specifically refer to the yaw rate commands of each wind turbine in the wind farm; ensure that the current candidate control commands are obtained from the wind farm's collaborative controller; the candidate control commands are usually recorded in the form of a time series, containing the yaw rate control requirements at each moment; Analyzing the acquired time-series data, the time point of the first non-zero or valid yaw rate command is identified as the start time, and the time point of the last valid command is identified as the end time. By subtracting the start and end times, the duration of the candidate control commands is obtained. The logic behind setting the instantaneous excitation intensity function is as follows: Extract the duration and command rate of candidate control commands. For any wind turbine, within the duration of the candidate control command, determine the active power of the wind turbine after the corresponding control at different times, and the average wind speed at the hub height of the wind turbine. The average wind speed specifically refers to the historical average wind speed within a preset detection window. Map the historical average wind speed to different times within the candidate control command to obtain the corresponding ambient wind speed. Based on the ambient wind speed and active power at the corresponding time, calculate the wind energy utilization coefficient of the wind turbine at that time. The ratio of the wind energy utilization coefficient to the maximum wind energy utilization coefficient of the wind turbine is recorded as the relative aerodynamic efficiency of the wind turbine at the corresponding time. The formula for calculating the relative aerodynamic efficiency is as follows: In the formula, Let be the relative aerodynamic efficiency of the j-th wind turbine at time t under the candidate control command. The time variable for candidate control instructions. Let be the wind energy utilization coefficient of the j-th wind turbine at time t under the candidate control command. Let be the maximum wind energy utilization coefficient of the j-th wind turbine, where j is the index of the wind turbine. Calculating the wind energy utilization factor of any wind turbine at the time corresponding to the candidate control command includes: The active power of the wind turbine after the corresponding regulation is obtained, and the theoretical wind energy input value at the corresponding time is determined by combining the ambient wind speed, air density and rotor sweep area at the corresponding time. Based on the correspondence between the active power and the theoretical wind energy input value, the wind energy utilization coefficient of the wind turbine at the corresponding time is calculated to characterize the aerodynamic energy conversion efficiency of the wind turbine under the current operating conditions; wherein the wind energy utilization coefficient of the j-th wind turbine at time t of the candidate control command. The formula used for the calculation is: In the formula, Let be the active power of the j-th wind turbine after adjustment at time t of the candidate control command. This is the theoretical wind energy input value. Let be the ambient wind speed of the j-th wind turbine at time t of the candidate control command. air density, Let be the blade radius of the j-th wind turbine unit; It should be noted that this formula is a direct transformation of the standard wind turbine aerodynamic power equation. Its physical meaning is the ratio of the actual wind energy captured to the total power of the incoming flow. Wind energy capture and conversion are the core of wind turbine performance evaluation. By using the theoretical wind energy input value as a benchmark, the energy conversion capability of the wind turbine under specific wind speed and environmental conditions can be intuitively expressed. The cubic relationship between wind energy and wind speed is a basic theory of fluid mechanics, reflecting the strong influence of wind speed changes on wind energy capture.

[0032] The active power of the wind turbine after corresponding regulation is predicted based on a deep learning model. The input of the deep learning model is the ambient wind speed, wind direction, candidate regulation commands, and corresponding yaw angle at the corresponding time. The output is the active power of the candidate regulation commands at the corresponding time. The deep learning model is specifically built based on an LSTM model, selecting an activation function and an optimization algorithm. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm for the LSTM model. The formula for the Tanh function is: In the formula, Represents the Tanh function, with the independent variable... This represents the weighted sum of the neuron's inputs, that is, the result of the weighted sum of the inputs received by the neuron from the previous layer; Simultaneously, the hyperparameters of the LSTM model are set, including: number of network layers, number of iterations, learning rate, batch size, number of training iterations, number of batches, and number of hidden layer neurons. The network is set to a 3-layer structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 32, the number of training iterations is set to 100, the number of batches is set to 256, and the number of hidden layer neurons is set to 32. The ambient wind speed at the corresponding moment is specifically characterized based on the ambient wind speed at the previous moment. The ambient wind speed is specifically characterized by the average wind speed at the wheel hub height within a preset time window. The preset time window is generally set to 1-5 minutes, with the current moment as the end point of the time window.

[0033] Collect historical data, including historical wind speed, wind direction, control commands, and corresponding yaw angles, and map the active power at the corresponding times to form a dataset. Divide the dataset into training and testing sets, typically using 80% for training and 20% for testing. Use the active power at the corresponding times as labels for training. Select an appropriate loss function, such as mean squared error, to evaluate the difference between the predicted and actual values. Input the training data into the model in batches; perform forward propagation in each batch to calculate the predicted value; calculate the loss, perform backpropagation, and update the model weights; iterate until the set number of iterations is reached; please refer to [link to relevant documentation]. Figure 2 , Figure 2 This shows the percentage of prediction error when the trained deep learning model is predicting active power, with the highest percentage of prediction error being... The prediction accuracy is satisfactory and can be used as a reference for active power in actual environments.

[0034] The ratio of the ambient wind speed to the preset rated wind speed at the corresponding moment of the wind turbine is normalized to obtain the base wind speed term. The product of the base wind speed term and the relative aerodynamic efficiency is used as the joint weighting factor for the wind turbine at the corresponding moment. This joint weighting factor is multiplied by the absolute value of the yaw rate of the candidate control command at the corresponding moment to obtain the instantaneous excitation intensity function of the wind turbine. The instantaneous excitation intensity function of the wind turbine is specifically expressed as follows: In the formula, Let be the instantaneous excitation intensity value of the j-th wind turbine at time t under the candidate control command. The yaw rate of the j-th wind turbine at time t under the candidate control command. The joint weighting factor for the j-th wind turbine at time t under the candidate control command; It should be noted that yaw rate It is the core variable in the candidate control instructions, and its absolute value reflects the control instruction at time t. The intensity of the excitation; the greater the yaw rate, the more significant the response characteristics of the wind turbine. Therefore, the absolute value of the yaw rate directly affects the instantaneous excitation intensity. Joint weighting factor It is the product of aerodynamic efficiency and normalized wind speed, reflecting the comprehensive performance of the wind turbine under current wind speed and aerodynamic conditions; it plays a weighting role in the instantaneous excitation intensity, used to adjust the final excitation intensity of the control command on the dynamic response of the wind turbine. Among them, the joint weighting factor The specific expression is: In the formula, Let be the base wind speed term of the j-th wind turbine at time t under the candidate control command; It should be noted that by combining relative aerodynamic efficiency and base wind speed, a composite factor is formed, which more comprehensively reflects the overall performance of the wind turbine under specific wind speed conditions and aerodynamic states. The influence of control commands is dynamically adjusted according to changes in real-time wind speed and aerodynamic efficiency, ensuring that the unit can adapt to wind speed fluctuations under different operating conditions.

[0035] The basic wind speed term is specifically represented as follows: In the formula, The rated wind speed of the wind turbine unit; It should be noted that the square relationship stems from the fact that aerodynamic torque is proportional to the square of wind speed; this formula ensures that the base wind speed term increases with increasing wind speed when the wind speed is below the rated wind speed, and remains at 1 when the wind speed is above the rated wind speed, avoiding excessive amplification. Additionally, if the current ambient wind speed... If the wind speed exceeds 1.1 times the rated wind speed, stop the operation of the wind turbine to avoid damage to the wind turbine. Rated wind speed is generally set according to the design objectives of the wind turbine, the site wind resource conditions, and the technical and economic requirements, and is determined through the technical manual of the wind turbine.

[0036] S103: Determine the relative impact intensity based on the ratio of the effective excitation equivalent of the candidate control command to the pre-stored reference excitation equivalent of the standard pulse command. Based on the relative impact intensity and the initial amplitude of the load oscillation, determine the amplitude of the load oscillation excited by the candidate control command.

[0037] For any wind turbine, the temporal forgetting coefficient is determined based on the undamped natural frequency of the wind turbine, and the exponential forgetting weight corresponding to each discrete moment is determined based on the time interval between each discrete moment during the duration of the candidate control command and the end moment of the candidate control command. The instantaneous excitation intensity function value at each discrete moment is weighted with the corresponding exponential forgetting weight to obtain the weighted excitation value at each discrete moment. The weighted excitation values ​​at all discrete moments are then integrated and accumulated to obtain the effective excitation equivalent of the candidate control command for the wind turbine. The exponential forgetting weight is used to characterize the contribution of instantaneous excitation at different discrete moments during the duration of the candidate control command to the load oscillation response. The closer the discrete moment is to the end of the candidate control command, the greater the corresponding exponential forgetting weight. The specific formula used to calculate the effective excitation equivalent of the candidate control command for the wind turbine is as follows: In the formula, Let be the effective excitation equivalent of the j-th wind turbine under the candidate control command. This is the temporal forgetting coefficient. This is the end time of the candidate control instruction; This is the exponential forgetting weight; It should be noted that in practical engineering, the dynamic response of a system typically weakens gradually over time; especially in mechanical systems, early excitations usually have a greater impact on the system's response, while the impact of subsequent excitations gradually decreases over time; the exponential forgetting weight is achieved through a function This dynamic characteristic is realized, which decreases over time, reflecting the trend of gradually weakening influence on past stimuli; it smoothly describes the nonlinear cumulative effect that occurs later and contributes more. Among them, the time-series forgetting coefficient It is generally set to 0.5 times the undamped natural frequency of the corresponding wind turbine; by using the forgetting factor The measured undamped natural frequency of the unit is fixed. The weight is 0.5 times that of the previous weight, so that the time constant of the weight is exactly at the optimal balance point between filtering high-frequency noise and preserving resonant energy.

[0038] Replace the candidate control command with the standard pulse command, and use the same weighted integration method as the candidate control command to calculate the reference effective excitation equivalent of the standard pulse command for the wind turbine; specifically, the trapezoidal method can be used for numerical integration. The relative impact intensity of a candidate control command is determined based on its effective excitation equivalent and the reference effective excitation equivalent of a standard pulse command. The relative impact intensity characterizes the excitation impact level of the candidate control command relative to the standard pulse command. The relative impact intensity is positively correlated with the effective excitation equivalent of the candidate control command and negatively correlated with the reference effective excitation equivalent of the standard pulse command. The specific formula used to calculate the relative impact intensity is as follows: In the formula, Let be the relative impact intensity of the j-th wind turbine under the candidate control command; this dimensionless coefficient reflects the impact capability of the current command relative to the standard pulse, and the larger the value, the stronger the impact of the candidate control command on the wind turbine.

[0039] Based on the relative impact intensity, the initial amplitude of the load oscillation corresponding to the standard pulse command, and the preset safety factor, the amplitude of the load oscillation excited by the candidate control command is determined; the specific formula used to calculate the load oscillation amplitude is as follows: In the formula, Let be the load oscillation amplitude of the j-th wind turbine unit excited by the candidate control command. Let be the initial amplitude of the load oscillation corresponding to the j-th wind turbine under the standard pulse command, where For safety factors, it is generally set between 0.8 and 1.5, with 1.1 being preferred. The linear assumption may underestimate the actual load under high impact intensity, and the uncertainty of the model is compensated by a factor of 1.1.

[0040] S104: Based on the load oscillation amplitude of the candidate control command and combined with the vibration characteristic parameters of the wind turbine, construct the envelope function of the load oscillation amplitude of each wind turbine under the candidate control. For any wind turbine, solve for the earliest safe execution time required for the envelope to decay to the preset safety threshold when executing the candidate control command.

[0041] Based on the measured blade root bending moment and its variation characteristics at the current moment, the equivalent amplitude corresponding to the current oscillation state is determined; the specific formula used to calculate the equivalent amplitude corresponding to the current oscillation state is as follows: In the formula, Let be the equivalent amplitude of the j-th wind turbine at the current moment. Let be the blade root bending moment of the j-th wind turbine at the current moment, which is specifically characterized by the average value of the blade root bending moments of each blade of the j-th wind turbine. Let be the rate of change of the blade root bending moment of the j-th wind turbine. Let be the undamped natural frequency of the j-th wind turbine. Indicates the current moment; It should be noted that the physical basis of this formula is the conservation of energy in simple harmonic motion, where the total mechanical energy is proportional to the square of the amplitude, and the total energy equals the current potential energy, which is proportional to... proportional to kinetic energy sum; The rate of change of blade root bending moment of the j-th wind turbine The specific acquisition method is as follows: In the last control cycle before executing the candidate instruction, read the current value and the previous value of the measured blade root bending moment under any blade, and calculate the difference. Based on the time difference between the difference and the previous value and the current value, calculate the current blade root bending moment change rate of the blade. Calculate the average value of the current blade root bending moment change rates of all blades of the wind turbine, and use it as the blade root bending moment change rate of the wind turbine.

[0042] The equivalent amplitude is superimposed with the load oscillation amplitude excited by the candidate control command, and compared with a preset safety threshold; the preset safety threshold is generally set according to the design limit value of the blade root flapping moment of the wind turbine. If the sum of the results is less than the preset security threshold, then the earliest safe execution time is determined to be 0. If the superposition result is greater than or equal to the preset safety threshold, an amplitude attenuation envelope is constructed based on the damping ratio and undamped natural frequency of the wind turbine, and the earliest safe execution time is determined according to the time it takes for the amplitude attenuation envelope to decay to the preset safety threshold. The amplitude attenuation envelope is specifically represented as follows: In the formula, To ensure the earliest safe execution time, Let j be the damping ratio of the j-th wind turbine. This is a safety threshold; It should be noted that the system's amplitude decreases over time, and this attenuation is closely related to the system's damping ratio and natural frequency. The decrease in amplitude over time can be expressed as... This indicates that the system gradually returns to an equilibrium state after being stimulated; Where the damping ratio Used to describe the energy dissipation capacity of a system; the larger the damping ratio, the faster the system's vibration decays and the shorter the time required to return to static equilibrium; undamped natural frequency. This represents the vibration frequency of the system when it is undamped, and it affects the vibration decay rate. The higher the natural frequency, the shorter the vibration period of the system, and the faster the decay process will be. This formula shows that when the amplitude of the wind turbine is... Reduce to equal to or less than the preset safety threshold At that time, the wind turbine will reach a safe state; by setting this equation, the earliest safe execution time can be solved. .

[0043] S105: Based on the rated blade bending moment load of each wind turbine, set the maximum waiting time for each wind turbine, compare the earliest safe execution time of each wind turbine with the maximum waiting time, and obtain the executability judgment result of the candidate control command on the corresponding wind turbine based on the comparison result.

[0044] The maximum waiting time for each wind turbine is proportional to its rated blade bending moment load. The method for setting the maximum waiting time for each wind turbine is as follows: The rated blade bending moment load refers to the maximum bending moment that the wind turbine blade can withstand under rated operating conditions. This value is an important parameter that must be considered when designing wind turbines, as it directly affects the safety and reliability of the wind turbines. The fatigue characteristics and dynamic response capability of the blade material are important factors in determining the maximum waiting time. The maximum waiting time is set as the maximum excitation duration that the blade can withstand under the rated bending moment load, ensuring that the blade will not be damaged at this time. That is, the maximum waiting time is proportional to the magnitude of the rated bending moment load. Based on the corresponding comparison results, the specific logic for determining the executability of candidate control commands on the corresponding wind turbine is as follows: for any wind turbine, if its earliest safe execution time is greater than 0 and less than the corresponding maximum waiting time, the candidate control command is executed after delaying the earliest safe execution time. If the earliest safe execution time exceeds the corresponding maximum waiting time, the candidate control command is determined to be unexecutable on the wind turbine. It should be noted that the earliest safe execution time is greater than 0 and less than the corresponding maximum waiting time, which means that the wind turbine can work safely under the current conditions, but has not yet reached its maximum tolerance limit. By delaying execution, potential risks caused by unreasonable incentives can be avoided while ensuring performance, and instructions can be executed reasonably. The earliest safe execution time has exceeded the corresponding maximum waiting time. In this case, the design safety boundary has been exceeded. Exceeding the maximum waiting time means that the unit's current operating state may cause fatigue damage or structural problems to the blades. Therefore, it is necessary to avoid executing control commands at this time to protect the structural integrity of the wind turbine.

[0045] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0046] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0047] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for rapid load assessment of wind turbine units for coordinated wind farm control, characterized in that, include: A standard pulse-shaped control command is applied to each wind turbine. For any wind turbine, the oscillation decay signal of the wind turbine is collected after the control command ends, so as to extract the vibration characteristic parameters of the wind turbine and the initial amplitude of the load oscillation corresponding to the standard pulse command. The candidate control commands issued by the current wind farm coordinating controller are obtained, the duration of the candidate control command and the command rate at each moment are extracted, and an instantaneous excitation intensity function is set based on the command rate and the ambient wind speed. The excitation intensity function is superimposed over the duration of the command to obtain the effective excitation equivalent of the candidate control command. The relative impact intensity is determined based on the ratio of the effective excitation equivalent of the candidate control command to the pre-stored reference excitation equivalent of the standard pulse command. Based on this relative impact intensity and the initial amplitude of the load oscillation, the amplitude of the load oscillation excited by the candidate control command is determined. Based on the load oscillation amplitude of the candidate control command and combined with the vibration characteristic parameters of the wind turbine, the envelope function of the load oscillation amplitude of each wind turbine under the candidate control is constructed. For any wind turbine, the earliest safe execution time required for the envelope to decay to the preset safety threshold when executing the candidate control command is solved. Based on the rated blade bending moment load of each wind turbine, the maximum waiting time for each wind turbine is set. The earliest safe execution time of each wind turbine is compared with the maximum waiting time. Based on the comparison results, the feasibility judgment result of the candidate control command on the corresponding wind turbine is obtained.

2. The method according to claim 1, characterized in that, The standard pulse shape control command is the standard yaw rate command for wind turbine generators; The oscillation attenuation signal of the wind turbine is the time sequence of the blade root flapping moment of each blade of the wind turbine, and the vibration characteristic parameters include the damping ratio and undamped natural frequency of the wind turbine.

3. The method according to claim 2, characterized in that, The specific method for obtaining the initial amplitude of load oscillation corresponding to the standard pulse command, as well as the vibration characteristic parameters of the wind turbine, is as follows: For any wind turbine, based on the blade root flapping moment time sequence of any blade, the blade root flapping moment time sequence of the blade is high-pass filtered to obtain the filtered signal. The maximum value of the absolute peak value is determined in the filtered signal, and the maximum value of the absolute peak value is recorded as the initial amplitude of the load oscillation of the blade corresponding to the standard pulse command. Iterate through all blades of the wind turbine and take the initial amplitude of the average load oscillation corresponding to the standard pulse command for each blade as the initial amplitude of the load oscillation corresponding to the standard pulse command for the wind turbine.

4. The method according to claim 3, characterized in that, The specific method for obtaining the vibration characteristic parameters of the wind turbine is as follows: For any blade, the maximum absolute peak value and several subsequent peak values ​​in the same direction are extracted from the blade root flapping moment time series to form a decay peak sequence. Based on the amplitude ratio of adjacent peaks in the attenuation peak sequence, the logarithmic reduction, which characterizes the degree of vibration attenuation, is calculated, and the damping ratio of the blade is determined based on a preset damping conversion relationship. Based on the time interval between adjacent peaks in the attenuation peak sequence, the damped natural frequency of the blade is determined, and the damped natural frequency is converted by combining the damping ratio to obtain the undamped natural frequency of the blade.

5. The method according to claim 1, characterized in that, The candidate control command specifically refers to the yaw rate command of each wind turbine in the wind farm; the logic for setting the instantaneous excitation intensity function is as follows: extract the duration of the candidate control command and the command rate at each moment; for any wind turbine, within the duration of the candidate control command, determine the active power of the wind turbine after the corresponding control at different moments, and the average wind speed at the hub height of the wind turbine. The average wind speed specifically refers to the historical average wind speed within the preset detection window. Map the historical average wind speed to different moments within the candidate control command as the ambient wind speed at the corresponding moment. Based on the ambient wind speed and active power at the corresponding moment, calculate the wind energy utilization coefficient of the wind turbine at the corresponding moment. The ratio of the wind energy utilization coefficient to the maximum wind energy utilization coefficient of the wind turbine is recorded as the relative aerodynamic efficiency of the wind turbine at the corresponding moment. The ratio of the ambient wind speed to the preset rated wind speed at the corresponding moment of the wind turbine is normalized to obtain the basic wind speed term. The product of the basic wind speed term and the relative aerodynamic efficiency is used as the joint weighting factor of the wind turbine at the corresponding moment. The joint weighting factor is multiplied by the absolute value of the yaw rate of the candidate control command at the corresponding moment to obtain the instantaneous excitation intensity function of the wind turbine.

6. The method according to claim 5, characterized in that, Calculating the wind energy utilization factor of any wind turbine at the time corresponding to the candidate control command includes: The active power of the wind turbine after the corresponding regulation is obtained, and the theoretical wind energy input value at the corresponding time is determined by combining the ambient wind speed, air density and rotor sweep area at the corresponding time. Based on the correspondence between the active power and the theoretical wind energy input value, the wind energy utilization coefficient of the wind turbine at the corresponding moment is calculated to characterize the aerodynamic energy conversion efficiency of the wind turbine under the current operating conditions. The active power of the wind turbine after corresponding regulation is predicted based on a deep learning model. The input of the deep learning model is the ambient wind speed, wind direction, candidate regulation command and corresponding yaw angle at the corresponding time, and the output is the active power of the candidate regulation command at the corresponding time.

7. The method according to claim 6, characterized in that, When calculating the effective excitation equivalent of the candidate control command, for any wind turbine, the temporal forgetting coefficient is determined based on the undamped natural frequency of the wind turbine, and the exponential forgetting weight corresponding to each discrete moment is determined based on the time interval between each discrete moment during the duration of the candidate control command and the end moment of the candidate control command. The instantaneous excitation intensity function value at each discrete moment is weighted with the corresponding exponential forgetting weight to obtain the weighted excitation value at each discrete moment. The weighted excitation values ​​at all discrete moments are then integrated and accumulated to obtain the effective excitation equivalent of the candidate control command for the wind turbine. The exponential forgetting weight is used to characterize the contribution of instantaneous excitation at different discrete moments during the duration of the candidate control command to the load oscillation response. The closer the discrete moment is to the end of the candidate control command, the greater the corresponding exponential forgetting weight.

8. The method according to claim 7, characterized in that, The candidate control command is replaced with the standard pulse command, and the same weighted integral method as the candidate control command is used to calculate the reference effective excitation equivalent of the standard pulse command for the wind turbine. The relative impact intensity corresponding to the candidate control command is determined based on the effective excitation equivalent of the candidate control command and the reference effective excitation equivalent of the standard pulse command. The relative impact intensity is used to characterize the excitation impact level of the candidate control command relative to the standard pulse command. The relative impact intensity is positively correlated with the effective excitation equivalent of the candidate control command and negatively correlated with the reference effective excitation equivalent of the standard pulse command.

9. The method according to claim 8, characterized in that, The logic for determining the earliest safe execution time is as follows: Based on the relative impact intensity, the initial amplitude of the load oscillation corresponding to the standard pulse command, and the preset safety factor, the amplitude of the load oscillation excited by the candidate control command is determined. Based on the measured leaf root bending moment and its variation characteristics at the current moment, determine the equivalent amplitude corresponding to the current oscillation state; The equivalent amplitude is superimposed with the load oscillation amplitude excited by the candidate control command, and compared with a preset safety threshold. If the sum of the results is less than the preset security threshold, then the earliest safe execution time is determined to be 0. If the superposition result is greater than or equal to the preset safety threshold, an amplitude attenuation envelope is constructed based on the damping ratio and undamped natural frequency of the wind turbine, and the earliest safe execution time is determined according to the time it takes for the amplitude attenuation envelope to decay to the preset safety threshold.

10. The method according to claim 8, characterized in that, The maximum waiting time for each wind turbine is proportional to its rated blade bending moment load. For any wind turbine, if its earliest safe execution time is greater than 0 and less than the corresponding maximum waiting time, then the candidate control command is executed after delaying the earliest safe execution time. If the earliest safe execution time exceeds the corresponding maximum waiting time, the candidate control command is determined to be unexecutable on the wind turbine.

Images