Wind driven generator large-torque transmission chain resistance increasing control method and system

By combining real-time spectrum analysis and state-space prediction models with filter banks and extended state observers, the problem of insufficient vibration suppression under high torque loads in traditional drive train resistance control methods has been solved. This enables adaptive control for different unit configurations, improving the operational stability and lifespan of wind turbines.

CN120946504APending Publication Date: 2025-11-14CHINA HUANENG INT ENG & TECH CO LTD +1
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202511345954.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-09-19
Publication Date
2025-11-14

AI Technical Summary

Technical Problem

Traditional transmission chain resistance control methods are ineffective when faced with multi-frequency composite vibrations caused by high torque loads, and cannot adapt to different unit configurations, leading to increased control complexity. They are also difficult to effectively attenuate vibrations in high wind speed ranges and have poor anti-interference performance under complex operating conditions.

Method used

By collecting wind turbine parameters in real time, performing spectrum analysis to lock the dominant vibration frequency, constructing a state-space prediction model, building a filter bank for filtering compensation, and using an extended state observer to estimate the total disturbance, an electromagnetic torque command with phase lead compensation is generated to achieve active suppression of transmission chain vibration.

Benefits of technology

It improves the vibration suppression effect and operational stability of the transmission chain, enhances the adaptability to different unit configurations, improves safety and reliability under complex working conditions, reduces engineering commissioning and maintenance costs, and extends equipment service life.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120946504A_ABST
    Figure CN120946504A_ABST
Patent Text Reader

Abstract

The invention discloses a wind driven generator large-torque transmission chain resistance increasing control method and system, and relates to the technical field of wind driven generator resistance increasing control. The method comprises the steps that operation parameters of a wind driven generator are collected in real time and preprocessed; calculating a rotation fluctuation frequency set and a corresponding amplitude spectrum of the current time window, and tracking and locking the dominant vibration frequency; constructing a fan state space prediction model containing the torsional vibration state of the transmission chain, outputting the predicted variable quantity of the torque of the transmission chain in the next control period, constructing a filter bank based on the dominant vibration frequency, performing filtering compensation on the predicted variable quantity, and generating an electromagnetic torque given value; and performing phase lead compensation on the electromagnetic torque given value, generating a control instruction, sending the control instruction to the wind power converter, and controlling the wind power generator to output the corresponding electromagnetic torque by using the wind power converter. Through systematic multi-link cooperative processing, the vibration suppression effect and the operation stability of the transmission chain can be comprehensively improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of wind turbine generator resistance control technology, and in particular to a method and system for resistance control of a high-torque transmission chain in a wind turbine generator. Background Technology

[0002] As wind turbines develop towards higher power and longer blades, the torque load on their transmission systems is increasing, posing a more severe challenge to torsional vibration suppression technology in the transmission chain. In particular, high-power onshore wind turbines, with the continuous increase in blade length, exhibit more frequency characteristics in the transmission chain control range below 10Hz. Traditional transmission chain damping techniques largely rely on whole-system simulation to determine the system's vibration modes, and then apply damping at a specific characteristic frequency when designing the overall control strategy.

[0003] Such methods are inadequate when dealing with complex vibrations of multiple frequencies that may coexist, caused by high torque loads. Using a single drivetrain resistance architecture and parameters is insufficient to effectively attenuate all drivetrain frequency signals, and may even lead to amplification of vibrations at some drivetrain frequencies. Furthermore, existing solutions are typically implemented in a main control programmable logic controller (PLC), which has a relatively long control cycle. The speed signal needs to be transmitted from the converter to the main control system via communication, introducing unavoidable communication transmission delays. This results in phase delays when detecting speed characteristic frequency fluctuations above 0.5Hz, ultimately causing phase lag in the execution of the resistance torque, severely affecting the resistance effect and making it difficult to reliably suppress overall machine speed oscillations.

[0004] Another significant limitation is its lack of adaptability. Since different blade, gearbox, tower and generator configurations will cause the unit to exhibit different oscillation modes, the fixed single characteristic frequency impedance strategy cannot adapt to the diverse unit configurations. This means that the characteristic frequency needs to be set separately for different configurations of the unit, which greatly increases the complexity of the engineering configuration and hinders the realization of platform-based management of the control software.

[0005] Furthermore, when the unit operates in the high wind speed range near or above the rated wind speed, the torsional vibration energy of the drive train is significantly enhanced. Traditional methods are insufficient to effectively attenuate all high-frequency energy, which may lead to a gradual increase in the oscillation amplitude of signals such as generator speed, electromagnetic torque, and turbine output power until they diverge. Especially when facing complex operating conditions such as nonlinear disturbances in the system, high-speed oscillations in the drive train, and sudden changes in external loads, the dynamic anti-interference performance of traditional methods is poor, the resistance effect is often unsatisfactory, and continuous and effective suppression cannot be guaranteed. Summary of the Invention

[0006] Therefore, it is necessary to provide a method and system for controlling the high torque transmission chain resistance of wind turbines to address the aforementioned technical problems.

[0007] In a first aspect, the present invention provides a method for controlling the resistance of a high-torque transmission chain in a wind turbine generator, comprising: S1. Real-time acquisition and preprocessing of wind turbine operating parameters; S2. Perform real-time spectrum analysis on the preprocessed operating parameters, calculate the rotational fluctuation frequency set and corresponding amplitude spectrum of the current time window, and track and lock the dominant vibration frequency. S3. Construct a wind turbine state-space prediction model that includes the torsional vibration state of the transmission chain, output the predicted change in the transmission chain torque in the next control cycle, and build a filter bank based on the dominant vibration frequency to filter and compensate the predicted change to generate the electromagnetic torque setpoint. S4. Perform phase advance compensation on the electromagnetic torque setpoint, generate control commands and send them to the wind power converter, and use the wind power converter to control the wind turbine to output the corresponding electromagnetic torque in order to achieve active suppression of transmission chain vibration. The wind turbine state-space prediction model, which includes the torsional vibration state of the drive train, is constructed, and the predicted change in drive train torque during the next control cycle is output, including: S31. Using the changes in generator speed, impeller speed, and transmission chain torsion angle as state variables, and the changes in electromagnetic torque and pitch angle as control variables, construct a wind turbine state-space prediction model. Based on the control cycle, set a finite-time domain optimization objective, and combine the tracking error of the wind turbine generator to solve for the predicted change in transmission chain torque in the next control cycle.

[0008] Furthermore, real-time acquisition and preprocessing of wind turbine operating parameters include: S11. Use a sensor network to synchronously collect the operating parameters of the wind turbine generator, including the generator speed, rotor speed, and the turbine-side voltage and turbine-side current of the wind power converter. S12. The generator speed and impeller speed are used as the original speed signals and input to the anti-aliasing low-pass filter to filter out high-frequency noise. An adaptive Kalman filter is used to perform linear unbiased estimation on the filtered speed signal to eliminate measurement noise and random interference. S13. Monitor the generator speed after filtering in real time, calculate the rate of change and amplitude within the preset interval, and compare it with the preset threshold range. If there is a speed exceeding the limit, mark the corresponding data stream as abnormal. If there is no speed exceeding the limit, mark it as healthy. S14. Combine the pre-processed generator speed, impeller rotation speed, converter side voltage, and side current with the data status identifier, and assign a unified timestamp to form a standardized data frame.

[0009] Furthermore, real-time spectrum analysis is performed on the preprocessed operating parameters to calculate the set of rotational fluctuation frequencies and corresponding amplitude spectra within the current time window, and to track and lock the dominant vibration frequencies, including: S21. Use the Gaussian window function to window the generator speed, and perform fast Fourier transform calculations in each time window to obtain the set of rotational fluctuation frequencies and the corresponding amplitude spectrum in the current time window. S22. Search for the point with the maximum amplitude in the amplitude spectrum, and set the rotational fluctuation frequency value corresponding to the point with the maximum amplitude as the candidate value of the vibration frequency that needs to be suppressed first in the current control cycle. S23. Compare the candidate vibration frequency obtained in the current control cycle with the dominant vibration frequency of the previous control cycle, calculate the absolute difference between the two, and compare the absolute difference with the preset reasonable frequency change threshold. If the absolute difference is less than or equal to the preset reasonable frequency change threshold, mark the candidate vibration frequency as the dominant vibration frequency. If the absolute difference is greater than the preset reasonable frequency change threshold, mark the candidate vibration frequency as an abnormal frequency. S24. Determine whether the dominant vibration frequency of the current control cycle after marking is within the physical frequency range of the transmission chain. If it is within the physical frequency range of the transmission chain, output the dominant vibration frequency. If it is not within the physical frequency range of the transmission chain or is marked as an abnormal frequency, output the dominant vibration frequency of the previous control cycle as the dominant vibration frequency of the current control cycle and trigger the warning mark of the abnormal frequency.

[0010] Furthermore, a filter bank is built based on the dominant vibration frequency to filter and compensate the predicted changes, generating the electromagnetic torque setpoint, including: S32. Based on the dominant vibration frequency, set the cutoff frequency of multiple low-pass filters in the filter bank, and use multiple low-pass filters to generate torque components that suppress low-frequency vibration and high-frequency vibration respectively. S33. Based on the extended state observer, the total disturbance of the generator set is estimated in real time. The total disturbance is estimated and the compensation torque is output based on the speed difference between the impeller speed and the equivalent generator speed. S34. The predicted change in transmission chain torque, torque component, compensation torque, and base electromagnetic torque generated by the upper controller are superimposed to form the primary electromagnetic torque setpoint. S35. Calculate the energy index between the primary electromagnetic torque setpoint and the base electromagnetic torque in real time, and compare the energy index with the preset energy index threshold. If the energy index is greater than the preset energy index threshold, mark the primary electromagnetic torque setpoint as incremental over-limit, output the electromagnetic torque setpoint and add the incremental over-limit mark; if the energy index is less than or equal to the preset energy threshold, mark the primary electromagnetic torque setpoint as normal, output the electromagnetic torque setpoint and add the healthy torque mark.

[0011] Furthermore, a state-space prediction model for the wind turbine is constructed. Based on the control cycle, a finite-time domain optimization objective is set. Combining the tracking error of the wind turbine generator, the predicted change in the transmission chain torque for the next control cycle is calculated, including: S311. Read the generator speed, impeller speed and system state estimate of the previous control cycle in the current control cycle, update and output the full state vector estimate of the wind turbine state space prediction model in the current control cycle using the extended state observer. S312. Based on the operating target of the wind turbine, calculate the continuous value of the power reference trajectory of the generator output power in the preset time domain, and set the torque reference trajectory of the transmission chain torque change. S313. Construct the finite-time domain optimization objective and cost function for the model predictive controller; S314. Using a numerical optimization algorithm, under the premise of satisfying all set constraints, the target control sequence in the preset control time domain is solved, and the optimal control increment of the first time step is extracted from the target control sequence as the predicted change in the current control cycle.

[0012] Furthermore, the finite-time optimization objective and cost function for the model predictive controller are constructed as follows: The finite-time domain optimization objectives are to minimize the deviation between the predicted state of the state variables and the torque reference trajectory, and to minimize the deviation between the predicted generator output power and the power reference trajectory, thus forming a quadratic cost function.

[0013] Furthermore, based on the dominant vibration frequency, the cutoff frequencies of multiple low-pass filters within the filter bank are set, and torque components that suppress low-frequency and high-frequency vibrations are generated using these multiple low-pass filters, including: S321. Based on the dominant vibration frequency, set the low-frequency suppression bandwidth and the high-frequency suppression bandwidth, and according to the low-frequency suppression bandwidth and the high-frequency suppression bandwidth, set the cutoff frequencies of the first-stage low-pass filter and the second-stage low-pass filter respectively. S322. Input the real-time acquired generator speed signal to the first-stage low-pass filter and the second-stage low-pass filter respectively. Use the first-stage low-pass filter to output the low-frequency speed fluctuation signal and use the second-stage low-pass filter to output the high-frequency speed fluctuation signal. S323. Multiply the low-frequency speed fluctuation signal with a preset first proportional gain to generate a torque component for suppressing low-frequency vibration in the transmission chain; multiply the high-frequency speed fluctuation signal with a preset second proportional gain to generate a torque component for suppressing high-frequency vibration in the transmission chain.

[0014] Furthermore, based on the extended state observer, the total disturbance of the generator set is estimated in real time. Using the speed difference between the impeller speed and the equivalent generator speed as input, the total disturbance is predicted and the compensation torque is output, including: S331. Set up a physical dynamic model of the wind turbine drive train, using the speed difference and derivative as control variables, and set the total disturbance as a control variable that is not directly measured, and establish an augmented state space model that includes the real state and the extended state. S332. Based on the augmented state space model, a linear extended state observer is established, and the linear extended state observer is discretized. The estimated value of the augmented state is calculated recursively, and the estimated value of the total disturbance is extracted from the augmented state. The compensation torque is generated based on the total disturbance.

[0015] Furthermore, phase lead compensation is performed on the electromagnetic torque setpoint to generate control commands and send them to the wind power converter. The wind power converter then controls the wind turbine to output the corresponding electromagnetic torque, thereby achieving active suppression of drivetrain vibration, including: S41. Read the electromagnetic torque setpoint and status indicator. If the status indicator of the electromagnetic torque setpoint is normal, proceed to step S42. If the status indicator of the electromagnetic torque setpoint is that the increment exceeds the limit, trigger the safety strategy to limit the electromagnetic torque setpoint within the preset safety range. S42. Calculate the phase parameters required for advance phase compensation using the dominant vibration frequency, take the verified electromagnetic torque setpoint as input, process it through the phase advance filter, and output the phase-compensated electromagnetic torque command. S43. Perform safety arbitration on the electromagnetic torque command. If the electromagnetic torque command is within the electromagnetic torque range that the wind turbine converter and wind turbine generator can safely execute, then send the electromagnetic torque command as a control command to the wind turbine converter. If the electromagnetic torque command is not within the electromagnetic torque range that the wind turbine converter and wind turbine generator can safely execute, then limit the electromagnetic torque command to the nearest boundary value and send it as a control command to the wind turbine converter.

[0016] Secondly, the present invention also provides a high-torque transmission chain resistance control system for wind turbines, the system comprising: The data acquisition and preprocessing module is used to acquire and preprocess the operating parameters of the wind turbine in real time. The frequency tracking and locking module is used to perform real-time spectrum analysis on the preprocessed operating parameters, calculate the set of rotational fluctuation frequencies and the corresponding amplitude spectrum of the current time window, and track and lock the dominant vibration frequency. The prediction compensation output module is used to construct a wind turbine state space prediction model that includes the torsional vibration state of the transmission chain, output the predicted change in the transmission chain torque in the next control cycle, and build a filter bank based on the dominant vibration frequency to filter and compensate the predicted change to generate the electromagnetic torque setpoint. The drive execution control module is used to perform phase advance compensation on the electromagnetic torque setpoint, generate control commands and send them to the wind power converter, and use the wind power converter to control the wind turbine to output the corresponding electromagnetic torque, so as to achieve active suppression of transmission chain vibration. The acquisition and preprocessing module, frequency tracking and locking module, prediction compensation output module, and drive execution control module are connected in sequence.

[0017] The beneficial effects of this invention are as follows: 1. Through systematic multi-stage collaborative processing, the vibration suppression effect and operational stability of the transmission chain can be comprehensively improved. By comprehensively utilizing strategies such as real-time spectrum analysis, model predictive control, extended state observer and phase compensation, it effectively addresses the core challenges faced by modern high-power, long-blade wind turbines under complex operating conditions, such as wide-frequency vibration, nonlinear disturbance and control delay, and significantly enhances the safety and reliability of the unit in high wind speed range and dynamic operation.

[0018] 2. It effectively improves the adaptability to different unit configurations and changing operating conditions. By dynamically locking the dominant vibration frequency through real-time spectrum analysis, and adaptively adjusting the filter bank parameters accordingly, it breaks the limitation of the traditional method of relying on fixed characteristic frequency resistance. This allows the same control strategy to be flexibly applied to wind turbines equipped with different blades, gearboxes, towers and generators, greatly enhancing the versatility and platform management level of the control software, and reducing the engineering commissioning and maintenance costs caused by characteristic changes due to differences in unit configuration or component aging.

[0019] 3. It exhibits strong robustness in the face of unmodeled dynamic and nonlinear factors and external disturbances. By introducing an extended state observer to estimate and compensate for the total disturbance in real time, combined with the multi-objective optimization capability of the model predictive controller, the control system can still maintain excellent vibration suppression effect when encountering complex operating conditions such as sudden wind speed changes and power grid fluctuations. It effectively suppresses the oscillation of transmission chain torque, improves power generation quality, and protects key mechanical components such as gearboxes and bearings from fatigue damage, thus extending the service life of the equipment. Attached Figure Description

[0020] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this invention, illustrate exemplary embodiments of the invention and are used to explain the invention, but do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of a high-torque transmission chain resistance control method for wind turbines according to an embodiment of the present invention; Figure 2 This is a system principle block diagram of a wind turbine high torque transmission chain resistance control system according to an embodiment of the present invention.

[0021] Reference numerals: 1. Acquisition and preprocessing module; 2. Frequency tracking and locking module; 3. Prediction and compensation output module; 4. Drive execution and control module. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0023] Please see Figure 1 A method for controlling the resistance of a high-torque drive train in a wind turbine is provided, comprising: S1. Real-time acquisition and preprocessing of wind turbine operating parameters.

[0024] In the description of this invention, real-time acquisition and preprocessing of wind turbine operating parameters includes: S11. Use a sensor network to synchronously collect the operating parameters of the wind turbine generator, including generator speed, rotor speed, and the turbine-side voltage and turbine-side current of the wind power converter.

[0025] Specifically, a sensor network consisting of high-precision encoders, voltage / current sensors, and other components is used to synchronously collect key physical quantities during the operation of the wind turbine generator. Generator speed ω g and impeller speed ω r It is the core raw data for calculating the torsional vibration of the transmission chain; the machine-side voltage and current can be used to assist in calculating the generator torque or as a supplementary verification source for speed data.

[0026] S12. The generator speed and impeller speed are used as the original speed signals and input to the anti-aliasing low-pass filter to filter out high-frequency noise. An adaptive Kalman filter is then used to perform linear unbiased estimation on the filtered speed signal to eliminate measurement noise and random interference.

[0027] Specifically, the raw speed signal first passes through an anti-aliasing low-pass filter to remove high-frequency noise, such as power electronic switch noise, preventing spectral aliasing in subsequent analysis. Then, an adaptive Kalman filter is used to recursively predict and update the system state, providing an optimal linear unbiased estimate of the true speed state even in the presence of noise and uncertainty. This effectively suppresses random interference and significantly improves signal quality. Compared to traditional fixed-parameter filters, the adaptive Kalman filter better adapts to dynamic changes in the system and the uncertainty of noise statistics, providing a cleaner and more reliable low-frequency speed signal.

[0028] S13. Monitor the generator speed after filtering in real time, calculate the rate of change and amplitude within a preset interval, and compare them with the preset threshold range. If there is a speed exceeding the limit, mark the corresponding data stream as abnormal. If there is no speed exceeding the limit, mark it as healthy.

[0029] Specifically, by moving the health monitoring function forward and embedding it into the data collection stage, the initial state perception is achieved, providing early and basic state information. This serves as the first line of defense for achieving full-process safety monitoring, avoiding protective actions that may affect stability in this early stage.

[0030] S14. Combine the pre-processed generator speed, impeller rotation speed, converter side voltage, and side current with the data status identifier, and assign a unified timestamp to form a standardized data frame.

[0031] Specifically, standardized output ensures the standardization and consistency of data transmission, providing a unified, high-quality data interface; timestamp alignment analysis ω g With ω r The phase relationship between them is a prerequisite for accurately calculating the torsional vibration of the transmission chain.

[0032] S2. Perform real-time spectrum analysis on the preprocessed operating parameters, calculate the set of rotational fluctuation frequencies and the corresponding amplitude spectrum for the current time window, and track and lock the dominant vibration frequency.

[0033] In the description of this invention, real-time spectrum analysis is performed on the preprocessed operating parameters to calculate the set of rotational fluctuation frequencies and the corresponding amplitude spectrum for the current time window, and to track and lock the dominant vibration frequency, including: S21. Apply a Gaussian window function to the generator speed and perform a fast Fourier transform calculation within each time window to obtain the set of rotational fluctuation frequencies and the corresponding amplitude spectrum within the current time window.

[0034] Specifically, a Short-Time Fourier Transform (STFT) is applied to the preprocessed generator speed. To achieve STFT, a Gaussian window function is first used to window the speed data, reducing spectral leakage caused by directly performing FFT on finite-length data and improving frequency resolution and accuracy. Then, a Fast Fourier Transform (FFT) is performed on the windowed data within each time window. The FFT converts the time-domain speed signal into a frequency-domain representation, ultimately obtaining the set of rotational fluctuation frequencies and their corresponding amplitude spectra within the current time window. The amplitude spectra visually demonstrate the energy strength of different frequency vibration components.

[0035] S22. Search for the point with the maximum amplitude in the amplitude spectrum, and set the rotational fluctuation frequency value corresponding to the point with the maximum amplitude as the candidate value of the vibration frequency that needs to be suppressed first in the current control cycle.

[0036] Specifically, in the obtained amplitude spectrum, all frequency points are traversed to search for the point with the maximum amplitude. The frequency corresponding to this point is determined as the candidate vibration frequency that needs to be prioritized for suppression in the current control cycle. f candidate The frequency component with the largest amplitude usually means that it excites the strongest vibration energy and is most harmful to the stability of the transmission chain and the load on the components. Therefore, it should be the primary target for suppression.

[0037] S23. Compare the candidate vibration frequency obtained in the current control cycle with the dominant vibration frequency of the previous control cycle, calculate the absolute difference between the two, and compare the absolute difference with the preset reasonable frequency change threshold. If the absolute difference is less than or equal to the preset reasonable frequency change threshold, mark the candidate vibration frequency as the dominant vibration frequency. If the absolute difference is greater than the preset reasonable frequency change threshold, mark the candidate vibration frequency as an abnormal frequency.

[0038] S24. Determine whether the dominant vibration frequency of the current control cycle after marking is within the physical frequency range of the transmission chain. If it is within the physical frequency range of the transmission chain, output the dominant vibration frequency. If it is not within the physical frequency range of the transmission chain or is marked as an abnormal frequency, output the dominant vibration frequency of the previous control cycle as the dominant vibration frequency of the current control cycle and trigger the abnormal frequency warning flag.

[0039] Specifically, regardless of whether the frequency is marked as "abnormal," a final physical rationality verification is required to determine whether the currently obtained dominant vibration frequency value is within the possible frequency range of the wind turbine drivetrain physical system. f min , f max If the frequency value is within [ ] f min , fmax If the frequency value is within [ ], then the dominant vibration frequency and its status indicator will be output. If the frequency value exceeds [ ], then the dominant vibration frequency and its status indicator will be output. f min , f max If the frequency falls within the specified range, this frequency data will be marked as an abnormal frequency, and the reliable frequency value of the previous cycle will be output. f dominant_prev As the output of this cycle, it also triggers an anomaly warning flag, which can prevent the control algorithm from outputting absurd frequency commands due to extreme abnormal signals, and is crucial for ensuring the safety of the unit.

[0040] S3. Construct a wind turbine state-space prediction model that includes the torsional vibration state of the transmission chain, output the predicted change in transmission chain torque in the next control cycle, and build a filter bank based on the dominant vibration frequency to filter and compensate the predicted change, thereby generating the electromagnetic torque setpoint.

[0041] In the description of this invention, a wind turbine state-space prediction model including the torsional vibration state of the transmission chain is constructed, outputting the predicted change in transmission chain torque in the next control cycle, and a filter bank is built based on the dominant vibration frequency to filter and compensate the predicted change, generating the electromagnetic torque setpoint including: S31. Using the changes in generator speed, impeller speed, and transmission chain torsion angle as state variables, and the changes in electromagnetic torque and pitch angle as control variables, construct a wind turbine state-space prediction model. Based on the control cycle, set a finite-time domain optimization objective, and combine the tracking error of the wind turbine generator to solve for the predicted change in transmission chain torque in the next control cycle.

[0042] In the description of this invention, a wind turbine state-space prediction model is constructed. A finite-time domain optimization objective is set based on the control cycle. Combining the tracking error of the wind turbine generator, the predicted change in the transmission chain torque for the next control cycle is calculated, including: S311. Read the generator speed, impeller speed and system state estimate of the previous control cycle in the current control cycle, update and output the full state vector estimate of the wind turbine state space prediction model in the current control cycle using the extended state observer.

[0043] Specifically, a fully real-time full-state estimation of the system is achieved through an Extended State Observer (ESO). Specifically, the system reads the generator speed ω for the current period. g and impeller speed ω rThe measured values, combined with the state estimates from the previous cycle, are used to update and output the full state vector estimate for the current cycle, which includes the torsional vibration state of the transmission chain, through the recursive algorithm of ESO. The state vector includes not only directly measurable states, but also estimates key states that cannot be directly measured, such as the torsional angular velocity of the transmission chain and the total unknown disturbance of the lumped system, including unmodeled dynamics, nonlinear friction, and external wind disturbance.

[0044] The wind turbine state-space prediction model is a mathematical model that describes the dynamic behavior of the wind turbine drivetrain. Based on current and historical states, it can predict the future behavior of the system. This model represents the wind turbine's operating states, such as generator speed, impeller speed, and drivetrain torsion angle, as a series of state variables. It uses a set of differential or difference equations to describe how these state variables change over time and how they are affected by control variables. The core function of this model is prediction. Combined with model predictive control (MPC) algorithms, it can predict the trend of drivetrain torque changes over a period of time, based on the currently estimated state and the future control sequence, in each control cycle.

[0045] Extended State Observer (ESO) is a core component of Active Disturbance Rejection Control (ADRC). It is a disturbance observer that does not rely on a precise mathematical model. ESO can treat all unknown factors, such as unmodeled dynamics and nonlinear characteristics within the system as well as external wind disturbances, as a single total disturbance and observe it in real time as a new extended state of the system. ESO estimates the system state and total disturbance that cannot be directly measured through the system's control inputs and measurable outputs.

[0046] "Full State Vector Estimation" refers to the set of real-time estimates of all state variables of the system (including the original state and extended state) obtained through ESO.

[0047] S312. Based on the operating target of the wind turbine, calculate the continuous value of the power reference trajectory of the generator output power in the preset time domain, and set the torque reference trajectory of the transmission chain torque change.

[0048] Specifically, based on the wind turbine's operational objectives, such as maximum power point tracking or power dispatch commands, the generator's output power is calculated within a preset prediction time domain. N p Power reference trajectory within P ref Simultaneously, based on the target of suppressing torsional vibration in the transmission chain, a reference trajectory Δ for the torque variation is set. T ref For example, the torque reference trajectory can be set to a smooth trajectory that approaches zero to directly suppress torque fluctuations.

[0049] By simultaneously tracking two reference trajectories, power and torque, the system no longer solely pursues power generation or solely suppresses vibration. It can intelligently balance multiple objectives, thereby ensuring power generation performance while actively and smoothly suppressing transmission chain vibration, thus improving the economy and safety of the entire machine operation.

[0050] S313. Construct the finite-time domain optimization objective and cost function for the model predictive controller.

[0051] Specifically, the finite-time optimization objective and quadratic cost function of the model predictive controller (MPC) are constructed. J The cost function is designed to minimize two core deviations: 1) the predicted state of the state variables and the torque reference trajectory Δ. T ref 1) Deviation; 2) The predicted value of generator output power differs from the power reference trajectory. P ref The deviation. At the same time, the cost function also includes the control increment (Δ). u The penalty term, in mathematical form, is: ; And it satisfies the following constraints: Model constraints: ; Control constraints: ; Control Incremental Constraints: ; Output constraints: ; In the formula, N p For the prediction time domain, Δ represents the number of forward prediction steps in MPC; T pred The predicted change in torque of the transmission chain is one of the state tracking objectives in the optimization problem; Δ T ref The reference trajectory for the torque variation in the transmission chain is typically desired to be smooth or approach zero to suppress vibration; Δ u ( k | t ) for at time t The predicted future control increment sequence, i.e., the optimization variable; Q , R This is the weight matrix. Q It is the penalty weight for state tracking error; the larger the weight, the closer the tracking is required. R It is the penalty weight for the control increment or control quantity. The larger the weight, the smoother the control action and the lower the energy consumption. u min , u maxThe minimum and maximum values ​​of the control quantity; Δ u min Δ u max To control the minimum and maximum values ​​of the increment; y min , y max These are the minimum and maximum values ​​of the output quantity; k This is a discrete-time index, representing the current time. t The Beginning of the Future k One control cycle.

[0052] In the description of this invention, the finite-time domain optimization objective and cost function for constructing the model predictive controller include: The finite-time domain optimization objectives are to minimize the deviation between the predicted state of the state variables and the torque reference trajectory, and to minimize the deviation between the predicted generator output power and the power reference trajectory, thus forming a quadratic cost function.

[0053] S314. Using a numerical optimization algorithm, under the premise of satisfying all set constraints, the target control sequence in the preset control time domain is solved, and the optimal control increment of the first time step is extracted from the target control sequence as the predicted change in the current control cycle.

[0054] Specifically, efficient numerical optimization algorithms, such as quadratic programming QP solvers, are used to solve the aforementioned finite-time optimization problem online, while satisfying all set constraints, such as electromagnetic torque change rate limits and generator speed safety ranges.

[0055] The solution yields a future control time domain. N c The optimal control increment sequence {Δ u ( k ), Δ u ( k +1| k ), ..., Δ u ( k + N c -1| k Following the "rolling optimization" principle of model predictive control, only the first element in the sequence—that is, the optimal control increment Δu (for the current control cycle)—is considered. k The extracted values ​​are used as the predicted change in transmission chain torque to drive the next cycle.

[0056] S32. Based on the dominant vibration frequency, set the cutoff frequency of multiple low-pass filters in the filter bank, and use multiple low-pass filters to generate torque components that suppress low-frequency vibration and high-frequency vibration respectively.

[0057] In the description of this invention, setting the cutoff frequency of multiple low-pass filters within a filter bank based on the dominant vibration frequency, and using the multiple low-pass filters to generate torque components that suppress low-frequency and high-frequency vibrations respectively includes: S321. Based on the dominant vibration frequency, set the low-frequency suppression bandwidth and the high-frequency suppression bandwidth, and according to the low-frequency suppression bandwidth and the high-frequency suppression bandwidth, set the cutoff frequencies of the first-stage low-pass filter and the second-stage low-pass filter respectively.

[0058] Specifically, it needs to be based on the dominant vibration frequency identified in real time. f d This is used to set the low-frequency suppression bandwidth and high-frequency suppression bandwidth. The low-frequency suppression bandwidth is usually set to [ fd -Δ f l , f d The high-frequency suppression bandwidth is set to []. f d , f d +Δ f h ], where Δ f l , and Δ f d This is based on the physical characteristics of the transmission chain and the preset frequency offset of the control target. Based on these bandwidths, the cutoff frequency of the first-stage low-pass filter (used to extract low-frequency fluctuation components) is set to... f d -Δ f l The cutoff frequency of the second-stage low-pass filter (used to extract high-frequency fluctuation components) is set to... f d +Δ f h The purpose of this design is to enable the suppression frequency band of the filter bank to dynamically follow the changes in the dominant vibration frequency during wind turbine operation, thereby achieving self-adaptation to changing operating conditions and different unit configurations, and avoiding the problem of poor suppression effect of traditional fixed frequency filters under parameter changes.

[0059] S322. The real-time acquired generator speed signal is input to the first-stage low-pass filter and the second-stage low-pass filter respectively. The first-stage low-pass filter outputs the low-frequency speed fluctuation signal, and the second-stage low-pass filter outputs the high-frequency speed fluctuation signal.

[0060] Specifically, the real-time acquired generator speed signal is simultaneously input to two parallel low-pass filters. The first low-pass filter (with a cutoff frequency of...) f d -Δ fl This is used to filter out components above its cutoff frequency, outputting a pure low-frequency speed fluctuation signal Δω. low The second low-pass filter (cutoff frequency is...) f d +Δ f h It will allow wideband signals below its cutoff frequency to pass through, and the output signal Δω wide It contains both low-frequency and high-frequency components. To separate the purely high-frequency fluctuation signal Δω... high Δω needs to be wide With Δω low Perform subtraction (Δω) high =Δω wide -Δω low This process effectively decomposes the complex speed fluctuation signal into two independent components: low-frequency and high-frequency components. This lays the foundation for subsequent generation of targeted torque suppression. Furthermore, by using a low-pass filter for separation, the phase information of the original signal can be well preserved. S323. Multiply the low-frequency speed fluctuation signal with a preset first proportional gain to generate a torque component for suppressing low-frequency vibration in the transmission chain; multiply the high-frequency speed fluctuation signal with a preset second proportional gain to generate a torque component for suppressing high-frequency vibration in the transmission chain.

[0061] S33. Based on the extended state observer, the total disturbance of the generator set is estimated in real time. The total disturbance is estimated and the compensation torque is output based on the speed difference between the impeller speed and the equivalent generator speed.

[0062] In the description of this invention, the total disturbance of the generator set is estimated in real time based on the extended state observer. Using the speed difference between the impeller speed and the equivalent generator speed as input, the total disturbance is estimated and a compensation torque is output, including: S331. Set up a physical dynamic model of the wind turbine drive train, using the speed difference and derivative as control variables, and set the total disturbance as a non-directly measured control variable to establish an augmented state space model that includes the real state and the extended state.

[0063] Specifically, a physical dynamic model of the wind turbine drivetrain is established. Based on Newton's second law, the transmission system is simplified into a mass-spring-damped system. The speed difference between the impeller and generator speeds (torsional angular velocity) and its derivative (torsional angular acceleration) are selected as state variables. All uncertainties in the system that are difficult to model or measure, such as nonlinear friction, parameter variations, and aerodynamic torque fluctuations, are uniformly considered as a "total disturbance" and expanded into a new state variable. Thus, the original state space of the system is augmented, forming an augmented state space model that includes the actual mechanical state and this expanded disturbance state, typically expressed as a set of differential equations. The core technical effect of this step is that, through model reconstruction, the complex and uncertain nonlinear system is transformed into a linear system that is formally simpler and easier for observer design. This lays the theoretical foundation for subsequent high-precision observation and compensation, and greatly enhances the inclusiveness of unmodeled internal dynamics and unknown external disturbances.

[0064] S332. Based on the augmented state space model, a linear extended state observer is established, and the linear extended state observer is discretized. The estimated value of the augmented state is calculated recursively, and the estimated value of the total disturbance is extracted from the augmented state. The compensation torque is generated based on the total disturbance.

[0065] Specifically, a linear extended state observer (LESO) is constructed. The observer takes directly measurable system signals, such as electromagnetic torque and speed difference, as input, and uses the observer gain matrix to recursively calculate and update the estimated value of the augmented state vector in real time.

[0066] In practical digital controllers, this continuous-time observer model needs to be discretized, such as using the Euler method or the zero-order hold method, to transform it into a difference equation form suitable for microprocessor execution. Through recursive calculations at each step, the observer can quickly track and output an estimate of the total disturbance term.

[0067] Finally, based on this estimate, an equivalent but reversed compensation torque is generated, which is directly added to the controller output as a feedforward. The core technical effect of this step is that it enables real-time dynamic estimation and active compensation of the total disturbance, transforming the originally complex and disturbed system compensation into a simple linear feedback problem. This significantly improves the response speed, anti-interference ability, and robustness of the control system, enabling the system to maintain high-performance vibration suppression even in the presence of significant model uncertainties and external disturbances.

[0068] S34. The predicted change in transmission chain torque, torque component, compensation torque, and base electromagnetic torque generated by the upper controller are superimposed to form the primary electromagnetic torque setpoint.

[0069] S35. Calculate the energy index between the primary electromagnetic torque setpoint and the base electromagnetic torque in real time. Compare the energy index with a preset energy index threshold. If the energy index is greater than the preset energy index threshold, mark the primary electromagnetic torque setpoint as exceeding the incremental limit and output the electromagnetic torque setpoint with the incremental limit exceeding mark. If the energy index is less than or equal to the preset energy threshold, mark the primary electromagnetic torque setpoint as normal and output the electromagnetic torque setpoint with the healthy torque mark.

[0070] S4. Perform phase advance compensation on the electromagnetic torque setpoint, generate control commands and send them to the wind power converter, and use the wind power converter to control the wind turbine to output the corresponding electromagnetic torque in order to achieve active suppression of transmission chain vibration.

[0071] In the description of this invention, phase lead compensation is performed on the electromagnetic torque setpoint to generate a control command and send it to the wind power converter. The wind power converter controls the wind turbine to output the corresponding electromagnetic torque, thereby achieving active suppression of transmission chain vibration, including: S41. Read the electromagnetic torque setpoint and status indicator. If the status indicator of the electromagnetic torque setpoint is normal, proceed to step S42. If the status indicator of the electromagnetic torque setpoint is that the increment exceeds the limit, trigger the safety strategy to limit the electromagnetic torque setpoint within the preset safety range.

[0072] Specifically, the electromagnetic torque setpoint and its status indicator are read. If the status indicator is "normal," the process continues; if the indicator is "incremental over-limit," it means that the setpoint may have exceeded the allowable instantaneous change range due to external disturbances or internal calculation anomalies, and a safety strategy will be triggered immediately. This strategy uses a preset safety range, usually dynamically calculated or obtained from a table based on the real-time power and speed of the wind turbine and the short-term overload capacity of the converter, to limit the setpoint and strictly restrict it within the safety range. This avoids the impact on the drivetrain caused by drastic changes in torque command, and even damage to critical components such as the pitch system and gearbox, ensuring the safety principle of the control system.

[0073] S42. Calculate the phase parameters required for advance phase compensation using the dominant vibration frequency, take the verified electromagnetic torque setpoint as input, process it through the phase advance filter, and output the phase-compensated electromagnetic torque command.

[0074] Specifically, by utilizing the dominant vibration frequency identified in real time ( f d This is used to calculate the key parameters required for the phase lead compensator. Its transfer function is usually in the form of... Gc ( s )= Ts +1 αTs +1 ( α >1), where the time constant isT =2 πfdα 1, while the leading coefficient α Then the phase quantity that needs to be compensated Decide, α =1−sin 1+sin 4. Phase amount to be compensated It is based on the vibration frequency of the entire system (including signal transmission, calculation, converter execution, etc.). f d The total phase lag is accurately assessed to determine the output. The electromagnetic torque setpoint, verified by S41, is used as input and processed by this phase-lead filter. The output is a phase-compensated electromagnetic torque command. The core technical effect of this step lies in actively canceling the phase lag in the control loop, ensuring that the generated suppression torque is precisely out of phase with the actual vibration phase of the transmission chain. This achieves optimal vibration suppression and solves the core problem of reduced resistance or even excitation caused by control delay.

[0075] S43. Perform safety arbitration on the electromagnetic torque command. If the electromagnetic torque command is within the electromagnetic torque range that the wind turbine converter and wind turbine generator can safely execute, then send the electromagnetic torque command as a control command to the wind turbine converter. If the electromagnetic torque command is not within the electromagnetic torque range that the wind turbine converter and wind turbine generator can safely execute, then limit the electromagnetic torque command to the nearest boundary value and send it as a control command to the wind turbine converter.

[0076] Specifically, the electromagnetic torque command, after phase compensation, undergoes final safety arbitration to determine whether it falls within the electromagnetic torque range that the wind turbine converter and generator can safely and reliably execute. This range comprehensively considers the generator's thermal capacity, torque limit, and the converter's maximum current output capability. If the command is within the safe range, it is directly sent to the wind turbine converter as the final control command; if the command exceeds the safe range, it is limited to the nearest boundary value (upper or lower limit), and this limited value is sent as the final control command.

[0077] This step constitutes the last safety barrier of the control system. While pursuing vibration suppression performance, it absolutely ensures that all executed commands are within the physical safety limits of the equipment, preventing damage to the generator and converter from over-torque and guaranteeing the long-term reliable operation of the equipment.

[0078] Please see Figure 2 Furthermore, a high-torque transmission chain resistance control system for wind turbines is also provided, the system comprising: The data acquisition and preprocessing module 1 is used to acquire and preprocess the operating parameters of the wind turbine in real time.

[0079] Frequency tracking and locking module 2 is used to perform real-time spectrum analysis on the preprocessed operating parameters, calculate the set of rotational fluctuation frequencies and the corresponding amplitude spectrum of the current time window, and track and lock the dominant vibration frequency.

[0080] The prediction compensation output module 3 is used to construct a wind turbine state space prediction model that includes the torsional vibration state of the transmission chain, output the predicted change in the transmission chain torque in the next control cycle, and build a filter bank based on the dominant vibration frequency to filter and compensate the predicted change to generate the electromagnetic torque setpoint.

[0081] The drive execution control module 4 is used to perform phase advance compensation on the electromagnetic torque setpoint, generate control commands and send them to the wind power converter, and use the wind power converter to control the wind turbine to output the corresponding electromagnetic torque, so as to achieve active suppression of transmission chain vibration.

[0082] Among them, the acquisition and preprocessing module 1, the frequency tracking and locking module 2, the prediction compensation output module 3, and the drive execution control module 4 are connected in sequence.

[0083] In summary, by utilizing the technical solutions described above, the vibration suppression effect and operational stability of the transmission chain can be comprehensively improved through systematic multi-stage collaborative processing. By comprehensively employing strategies such as real-time spectrum analysis, model predictive control, extended state observers, and phase compensation, the core challenges faced by modern high-power, long-bladed wind turbines under complex operating conditions, including wide-frequency domain vibration, nonlinear disturbances, and control delays, are effectively addressed. This significantly enhances the safety and reliability of the unit in high wind speed ranges and during dynamic operation. Furthermore, it effectively improves the adaptability to different unit configurations and changing operating conditions. By dynamically locking the dominant vibration frequency through real-time spectrum analysis and adaptively adjusting the filter bank parameters accordingly, it breaks through the limitations of traditional methods that rely on fixed characteristic frequencies for impedance adjustment. This allows the same control strategy to be flexibly applied to wind turbines equipped with different blades, gearboxes, towers, and generators, greatly enhancing the versatility and platform management level of the control software and reducing engineering commissioning and maintenance costs caused by characteristic changes due to differences in unit configuration or component aging. It exhibits strong robustness in the face of unmodeled dynamic and nonlinear factors and external disturbances. By introducing an extended state observer to estimate and compensate for the total disturbance in real time, combined with the multi-objective optimization capability of the model predictive controller, the control system can still maintain excellent vibration suppression effect when encountering complex operating conditions such as sudden wind speed changes and power grid fluctuations. It effectively suppresses the oscillation of transmission chain torque, improves power generation quality, and protects key mechanical components such as gearboxes and bearings from fatigue damage, thus extending the service life of the equipment.

[0084] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

Claims

1. A method for controlling the resistance of a high-torque transmission chain in a wind turbine generator, characterized in that, include: S1. Real-time acquisition and preprocessing of wind turbine operating parameters; S2. Perform real-time spectrum analysis on the preprocessed operating parameters, calculate the rotational fluctuation frequency set and corresponding amplitude spectrum of the current time window, and track and lock the dominant vibration frequency. S3. Construct a wind turbine state-space prediction model that includes the torsional vibration state of the transmission chain, output the predicted change in the transmission chain torque in the next control cycle, and build a filter bank based on the dominant vibration frequency to filter and compensate the predicted change to generate the electromagnetic torque setpoint. S4. Perform phase advance compensation on the electromagnetic torque setpoint, generate control commands and send them to the wind power converter, and use the wind power converter to control the wind turbine to output the corresponding electromagnetic torque in order to achieve active suppression of transmission chain vibration. The construction of the wind turbine state-space prediction model, which includes the torsional vibration state of the transmission chain, and the output of the predicted change in transmission chain torque for the next control cycle, includes: S31. Using the changes in generator speed, impeller speed, and transmission chain torsion angle as state variables, and the changes in electromagnetic torque and pitch angle as control variables, construct a wind turbine state-space prediction model. Based on the control cycle, set a finite-time domain optimization objective, and combine the tracking error of the wind turbine generator to solve for the predicted change in transmission chain torque in the next control cycle.

2. The method for controlling the resistance of a high-torque transmission chain in a wind turbine generator according to claim 1, characterized in that, The real-time acquisition and preprocessing of wind turbine operating parameters includes: S11. Synchronously collect the operating parameters of the wind turbine using a sensor network. The operating parameters include the generator speed, rotor speed, and the turbine-side voltage and turbine-side current of the wind power converter. S12. The generator speed and impeller speed are used as the original speed signals and input to the anti-aliasing low-pass filter to filter out high-frequency noise. An adaptive Kalman filter is used to perform linear unbiased estimation on the filtered speed signal to eliminate measurement noise and random interference. S13. Monitor the generator speed after filtering in real time, calculate the rate of change and amplitude within the preset interval, and compare it with the preset threshold range. If there is a speed exceeding the limit, mark the corresponding data stream as abnormal. If there is no speed exceeding the limit, mark it as healthy. S14. Combine the pre-processed generator speed, impeller rotation speed, converter side voltage, and side current with the data status identifier, and assign a unified timestamp to form a standardized data frame.

3. The method for controlling the resistance of a high-torque transmission chain in a wind turbine according to claim 1, characterized in that, The step of performing real-time spectrum analysis on the preprocessed operating parameters, calculating the set of rotational fluctuation frequencies and the corresponding amplitude spectrum for the current time window, and tracking and locking the dominant vibration frequency includes: S21. Use the Gaussian window function to window the generator speed, and perform fast Fourier transform calculations in each time window to obtain the set of rotational fluctuation frequencies and the corresponding amplitude spectrum in the current time window. S22. Search for the point with the maximum amplitude in the amplitude spectrum, and set the rotational fluctuation frequency value corresponding to the point with the maximum amplitude as the candidate value of the vibration frequency that needs to be suppressed first in the current control cycle. S23. Compare the candidate vibration frequency obtained in the current control cycle with the dominant vibration frequency of the previous control cycle, calculate the absolute difference between the two, and compare the absolute difference with the preset reasonable frequency change threshold. If the absolute difference is less than or equal to the preset reasonable frequency change threshold, mark the candidate vibration frequency as the dominant vibration frequency. If the absolute difference is greater than the preset reasonable frequency change threshold, mark the candidate vibration frequency as an abnormal frequency. S24. Determine whether the dominant vibration frequency of the current control cycle after marking is within the physical frequency range of the transmission chain. If it is within the physical frequency range of the transmission chain, output the dominant vibration frequency. If it is not within the physical frequency range of the transmission chain or is marked as an abnormal frequency, output the dominant vibration frequency of the previous control cycle as the dominant vibration frequency of the current control cycle and trigger the warning mark of the abnormal frequency.

4. The method for controlling the resistance of a high-torque transmission chain in a wind turbine generator according to claim 1, characterized in that, The process of building a filter bank based on the dominant vibration frequency to filter and compensate the predicted changes and generate the electromagnetic torque setpoint includes: S32. Based on the dominant vibration frequency, set the cutoff frequency of multiple low-pass filters in the filter bank, and use multiple low-pass filters to generate torque components that suppress low-frequency vibration and high-frequency vibration respectively. S33. Based on the extended state observer, the total disturbance of the generator set is estimated in real time. The total disturbance is estimated and the compensation torque is output based on the speed difference between the impeller speed and the equivalent generator speed. S34. The predicted change in transmission chain torque, torque component, compensation torque, and base electromagnetic torque generated by the upper controller are superimposed to form the primary electromagnetic torque setpoint. S35. Calculate the energy index between the primary electromagnetic torque setpoint and the base electromagnetic torque in real time, and compare the energy index with the preset energy index threshold. If the energy index is greater than the preset energy index threshold, mark the primary electromagnetic torque setpoint as incremental over-limit, output the electromagnetic torque setpoint and add the incremental over-limit mark; if the energy index is less than or equal to the preset energy threshold, mark the primary electromagnetic torque setpoint as normal, output the electromagnetic torque setpoint and add the healthy torque mark.

5. The method for controlling the resistance of a high-torque transmission chain in a wind turbine according to claim 1, characterized in that, The construction of the wind turbine state-space prediction model involves setting a finite-time domain optimization objective based on the control cycle, and combining this with the tracking error of the wind turbine generator to solve for the predicted change in the transmission chain torque during the next control cycle, including: S311. Read the generator speed, impeller speed and system state estimate of the previous control cycle in the current control cycle, update and output the full state vector estimate of the wind turbine state space prediction model in the current control cycle using the extended state observer. S312. Based on the operating target of the wind turbine, calculate the continuous value of the power reference trajectory of the generator output power in the preset time domain, and set the torque reference trajectory of the transmission chain torque change. S313. Construct the finite-time domain optimization objective and cost function of the model predictive controller; S314. Using a numerical optimization algorithm, under the premise of satisfying all set constraints, the target control sequence in the preset control time domain is solved, and the optimal control increment of the first time step is extracted from the target control sequence as the predicted change in the current control cycle.

6. The method for controlling the resistance of a high-torque transmission chain in a wind turbine generator according to claim 5, characterized in that, The finite-time optimization objective and cost function for constructing the model predictive controller include: The finite-time domain optimization objectives are to minimize the deviation between the predicted state of the state variables and the torque reference trajectory, and to minimize the deviation between the predicted generator output power and the power reference trajectory, thus forming a quadratic cost function.

7. The method for controlling the resistance of a high-torque transmission chain in a wind turbine according to claim 4, characterized in that, The step of setting the cutoff frequency of multiple low-pass filters within the filter bank based on the dominant vibration frequency, and using multiple low-pass filters to generate torque components that suppress low-frequency and high-frequency vibrations respectively, includes: S321. Based on the dominant vibration frequency, set the low-frequency suppression bandwidth and the high-frequency suppression bandwidth, and according to the low-frequency suppression bandwidth and the high-frequency suppression bandwidth, set the cutoff frequencies of the first-stage low-pass filter and the second-stage low-pass filter respectively. S322. Input the real-time acquired generator speed signal to the first-stage low-pass filter and the second-stage low-pass filter respectively. Use the first-stage low-pass filter to output the low-frequency speed fluctuation signal and use the second-stage low-pass filter to output the high-frequency speed fluctuation signal. S323. Multiply the low-frequency speed fluctuation signal with a preset first proportional gain to generate a torque component for suppressing low-frequency vibration in the transmission chain; multiply the high-frequency speed fluctuation signal with a preset second proportional gain to generate a torque component for suppressing high-frequency vibration in the transmission chain.

8. The method for controlling the resistance of a high-torque transmission chain in a wind turbine according to claim 4, characterized in that, The method of estimating the total disturbance of the generator set in real time based on the extended state observer, using the speed difference between the impeller speed and the equivalent generator speed as input, to predict the total disturbance and output the compensation torque includes: S331. Set up a physical dynamic model of the wind turbine drive train, using the speed difference and derivative as control variables, and set the total disturbance as a control variable that is not directly measured, and establish an augmented state space model that includes the real state and the extended state. S332. Based on the augmented state space model, a linear extended state observer is established, and the linear extended state observer is discretized. The estimated value of the augmented state is calculated recursively, and the estimated value of the total disturbance is extracted from the augmented state. The compensation torque is generated based on the total disturbance.

9. The method for resistance control of a high-torque transmission chain in a wind turbine generator according to claim 1, characterized in that, The process of performing phase lead compensation on the electromagnetic torque setpoint, generating control commands and sending them to the wind power converter, and using the wind power converter to control the wind turbine to output the corresponding electromagnetic torque to achieve active suppression of transmission chain vibration includes: S41. Read the electromagnetic torque setpoint and status indicator. If the status indicator of the electromagnetic torque setpoint is normal, proceed to step S42. If the status indicator of the electromagnetic torque setpoint is that the increment exceeds the limit, trigger the safety strategy to limit the electromagnetic torque setpoint within the preset safety range. S42. Calculate the phase parameters required for advance phase compensation using the dominant vibration frequency, take the verified electromagnetic torque setpoint as input, process it through the phase advance filter, and output the phase-compensated electromagnetic torque command. S43. Perform safety arbitration on the electromagnetic torque command. If the electromagnetic torque command is within the electromagnetic torque range that the wind turbine converter and wind turbine generator can safely execute, then send the electromagnetic torque command as a control command to the wind turbine converter. If the electromagnetic torque command is not within the electromagnetic torque range that the wind turbine converter and wind turbine generator can safely execute, then limit the electromagnetic torque command to the nearest boundary value and send it as a control command to the wind turbine converter.

10. A high-torque transmission chain resistance control system for a wind turbine generator, used to implement the high-torque transmission chain resistance control method for a wind turbine generator as described in any one of claims 1-9, characterized in that, The system includes: The data acquisition and preprocessing module is used to acquire and preprocess the operating parameters of the wind turbine in real time. The frequency tracking and locking module is used to perform real-time spectrum analysis on the preprocessed operating parameters, calculate the set of rotational fluctuation frequencies and the corresponding amplitude spectrum of the current time window, and track and lock the dominant vibration frequency. The prediction compensation output module is used to construct a wind turbine state space prediction model that includes the torsional vibration state of the transmission chain, output the predicted change in the transmission chain torque in the next control cycle, and build a filter bank based on the dominant vibration frequency to filter and compensate the predicted change to generate the electromagnetic torque setpoint. The drive execution control module is used to perform phase advance compensation on the electromagnetic torque setpoint, generate control commands and send them to the wind power converter, and use the wind power converter to control the wind turbine to output the corresponding electromagnetic torque, so as to achieve active suppression of transmission chain vibration. The acquisition and preprocessing module, the frequency tracking and locking module, the prediction compensation output module, and the drive execution control module are sequentially connected.

Citation Information

Cited By

  • Solid powder homogenizing equipment and mixing control method thereof

    CN121669075A

  • Tension Adaptive Control System for Rocket Nozzle Diver Section Winding

    CN122401865A