A wind turbine global control method and system based on stability constraints

Abstract

The present disclosure relates to a wind turbine global control method and system based on stability constraints, and relates to the technical field of wind turbine intelligent control, comprising: collecting multi-source data in the operation process of the wind turbine and constructing global state variables, disturbance variables and actuator residual regulation variables; predicting multiple stability responses and the stability margin of each stability response according to the global state variables and the disturbance variables; establishing a coupling instability chain according to the lag correlation coefficient between the multiple stability margins, the prediction sensitivity and the normalized coefficient of the actuator residual regulation capacity; generating a chain constraint feasible region according to the first launch depletion margin, the link propagation time, the model prediction residual and the actuator residual regulation variable; solving the candidate global control variable in the chain constraint feasible region; when the candidate global control variable touches the boundary of the chain constraint feasible region, allocating a control repair variable along the reverse recovery path of the coupling instability chain, and determining the target global control variable. Thus, the global stability of the wind turbine under complex conditions is improved.

Description

Technical Field

[0001] This disclosure relates to the field of intelligent control technology for wind turbine generators, and more specifically, to a global control method and system for wind turbine generators based on stability constraints. Background Technology

[0002] As the capacity of individual wind turbine units continues to increase, along with the continuous expansion of rotor diameter, tower height, and flexible structure dimensions, the operational stability of wind turbines under complex turbulence, wind shear, tower shadow effect, gust impact, weak grid disturbances, and frequent power dispatch conditions, is becoming increasingly prominent. Traditional wind turbine control often employs a hierarchical control architecture. For example, pitch control maintains rated speed, generator torque control achieves maximum power point tracking or power limiting, yaw control reduces yaw error, and converter control regulates active power, reactive power, and grid-connected current. While this type of control can meet the power regulation requirements under normal operating conditions, the different control links are usually relatively independent, making it difficult to promptly identify the coupling propagation relationships between blade loads, tower vibration, drive train torsional vibration, speed fluctuations, converter current margin, and grid synchronization margin. When a certain stability domain first experiences margin depletion, its impact may propagate step by step along paths such as aerodynamic loads, structural vibration, drive train impact, electrical side current limitations, and grid synchronization errors, resulting in the inability of local compensation in a single control link to suppress the risk of global instability. Summary of the Invention

[0003] The purpose of this disclosure is to provide a global control method and system for wind turbine generators based on stability constraints.

[0004] According to a first aspect of the present disclosure, a global control method for wind turbine generators based on stability constraints is provided, comprising: Collect multi-source data during the operation of the wind turbine, and construct global state variables, disturbance variables, and actuator residual regulation variables based on the multi-source data; the multi-source data includes at least one of the following: incoming flow data, impeller data, tower data, drive train data, converter data, and grid connection status data; Multiple stable responses are predicted based on the global state variables and the disturbance variables, and the stability margin of each stable response is determined based on the prediction results, resulting in multiple stability margins; the multiple stability margins include at least one of the following: aeroelastic stability margin, tower vibration stability margin, transmission chain torsional vibration stability margin, speed stability margin, converter current stability margin, and grid synchronization stability margin. A coupled instability chain is established based on the hysteresis correlation coefficients between each pair of the multiple stability margins, the normalized coefficients of the prediction sensitivity and the remaining adjustment capability of the actuator; Based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator, the chain constraint feasible region of the coupled unstable chain is generated. Solve for the candidate global control variables of the wind turbine within the feasible region of the chain constraints; When the candidate global control quantity touches the boundary of the feasible region of the chain constraint, a control repair quantity is allocated along the reverse recovery path of the coupled unstable chain, and a target global control quantity is determined based on the candidate global control quantity and the control repair quantity.

[0005] Optionally, the global state variables include at least one of the following: wind turbine speed, generator speed, blade root bending moment, tower forward and backward acceleration, tower lateral acceleration, low-speed shaft torque, high-speed shaft torque, converter active current, converter reactive current, grid connection point voltage, grid connection point frequency, and phase angle deviation. The disturbance includes at least one of wind speed, wind direction, wind shear intensity, turbulence intensity, grid voltage drop amplitude, and grid frequency offset. The remaining adjustment amount of the actuator includes at least one of the following: remaining pitch angle stroke, remaining pitch rate, remaining generator torque adjustment, remaining yaw angle adjustment, remaining converter active current capacity, and remaining converter reactive current capacity.

[0006] Optionally, the step of predicting multiple stable responses based on the global state quantity and the disturbance quantity, and determining the stability margin of each stable response based on the prediction results, includes: A first predicted stable response is determined based on the global state quantity and the disturbance quantity; the first stable response is any one of the plurality of stable responses, and the first predicted stable response is the predicted stable response corresponding to the first stable response. The stability margin of the first stable response is determined based on the distance between the stability boundary of the first stable response and the first predicted stable response.

[0007] Optionally, establishing the coupled instability chain based on the hysteresis correlation coefficients, predictive sensitivity, and normalized coefficients of the remaining adjustment capability of the actuators among the pairwise relationships of the plurality of stability margins includes: The link impact coefficient is determined based on the hysteresis correlation coefficient between each pair of the multiple stability margins, the prediction sensitivity, and the normalization coefficient. The coupling instability chain is established based on the link influence coefficient.

[0008] Optionally, establishing the coupling instability chain based on the link influence coefficient includes: When the link influence coefficient between any two stability margins is greater than the preset link determination threshold, an instability propagation edge is established between the two stability margins, pointing from the previous stability margin to the next stability margin. Determine the first exhaustion margin among the plurality of stability margins, wherein the first exhaustion margin is the first stability margin among the plurality of stability margins that is less than the corresponding stability margin threshold in the prediction time domain. Starting from the initial exhaustion margin, the subsequent stability margin is searched sequentially along the unstable propagation edge, and the main propagation path is selected from large to small according to the link influence coefficient to obtain the coupled unstable chain.

[0009] Optionally, generating the chain constraint feasible region of the coupled unstable chain based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator includes: The link propagation time is determined by the sum of the propagation times required for two adjacent stability margins in the coupled unstable chain; The dynamic tightening boundary of each stable response is determined based on the stability margin, stability margin exhaustion rate, model prediction residual, remaining actuator adjustment, and position in the coupled instability chain corresponding to each stable response. The chain-constrained stabilizing tube is determined based on the dynamically tightening boundary of each stable response; The chain-constrained feasible region is generated based on the chain-constrained stabilizing tube, the global feasible region of the coupled unstable chain, the initial exhaustion margin, and the link propagation time.

[0010] Optionally, solving for the candidate global control variables of the wind turbine within the feasible region of the chain constraints includes: Using power tracking deviation, the penalty for the decrease in stability margin of each stability margin in the coupled instability chain, tower vibration penalty, transmission chain torsional vibration penalty, and actuator action change penalty as joint objectives, and the feasible region of the chain constraint as constraint conditions, the candidate global control quantities are solved. The candidate global control quantities include at least one of the following: collective pitch angle, independent pitch correction of each blade of the wind turbine, generator torque, yaw offset angle, converter active current command, and converter reactive current command.

[0011] According to a second aspect of the present disclosure, a global control system for wind turbine generators based on stability constraints is provided. The system includes: a data acquisition module, a global state construction module, a stability margin prediction module, a coupled instability chain generation module, a chain constraint generation module, a global optimization module, a reverse repair allocation module, and an execution interface module. The data acquisition module is used to collect multi-source data during the operation of the wind turbine, and construct global state variables, disturbance variables, and actuator residual regulation variables based on the multi-source data; the multi-source data includes at least one of the following: incoming flow data, impeller data, tower data, transmission chain data, converter data, and grid connection status data; The global state construction module is used to predict multiple stable responses based on the global state quantity and the disturbance quantity, and to determine the stability margin of each stable response based on the prediction results, thereby obtaining multiple stability margins; the multiple stability margins include at least one of the following: aeroelastic stability margin, tower vibration stability margin, transmission chain torsional vibration stability margin, speed stability margin, converter current stability margin, and grid synchronization stability margin. The coupled instability chain generation module is used to establish coupled instability chains based on the hysteresis correlation coefficients, prediction sensitivity, and normalized coefficients of the remaining adjustment capability of the actuators between each pair of the multiple stability margins. The chain constraint generation module is used to generate the feasible region of chain constraints for the coupled unstable chain based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator. The global optimization module is used to solve for the candidate global control variables of the wind turbine within the feasible region of the chain constraint. The reverse repair allocation module is used to allocate a control repair amount along the reverse recovery path of the coupled unstable chain when the candidate global control amount touches the boundary of the feasible region of the chain constraint, and to determine the target global control amount based on the candidate global control amount and the control repair amount.

[0012] Optionally, the coupling unstable chain generation module includes: a link influence coefficient calculation unit and a coupling unstable chain generation unit; The link impact coefficient calculation unit is used to determine the link impact coefficient based on the lag correlation coefficient between each pair of the plurality of stability margins, the prediction sensitivity, and the normalization coefficient. The coupling unstable chain generation unit is used to establish the coupling unstable chain based on the link influence coefficient.

[0013] Optionally, the system further includes: an execution interface module; The execution interface module is used to output control commands to at least one of the pitch system, generator torque control system, yaw system and converter control system, so that the pitch system, the generator torque control system, the yaw system and the converter control system output the corresponding control quantity in the target global control quantity.

[0014] The above technical solution collects multi-source data during the operation of the wind turbine, and constructs global state variables, disturbance variables, and actuator residual regulation variables based on this multi-source data. This multi-source data includes at least one of the following: incoming flow data, impeller data, tower data, drive train data, converter data, and grid connection status data. Multiple stable responses are predicted based on the global state variables and disturbance variables, and the stability margin of each stable response is determined based on the prediction results, resulting in multiple stability margins. These multiple stability margins include: aeroelastic stability margin, tower vibration stability margin, drive train torsional vibration stability margin, speed stability margin, converter current stability margin, and grid connection synchronization stability margin. The method involves determining at least one of the stability margins; establishing a coupled instability chain based on the hysteresis correlation coefficients, prediction sensitivity, and normalized coefficients of the actuator's residual regulation capacity between each pair of these multiple stability margins; generating a chain-constrained feasible region for the coupled instability chain based on the initial exhaustion margin, the link propagation time of the coupled instability chain, the model prediction residual, and the actuator's residual regulation; solving for candidate global control variables for the wind turbine within this chain-constrained feasible region; and when the candidate global control variable touches the boundary of the chain-constrained feasible region, allocating a control repair amount along the reverse recovery path of the coupled instability chain, and determining the target global control variable based on the candidate global control variable and the control repair amount. This method can improve the global stability of wind turbines under complex conditions and is beneficial for extending the lifespan of key components of wind turbines and improving the long-term power generation revenue of the turbine.

[0015] Other features and advantages of this disclosure will be described in detail in the following detailed description section. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the present disclosure and form part of the specification. They are used together with the following detailed description to explain the present disclosure, but do not constitute a limitation thereof. In the drawings: Figure 1 This is one of the flowcharts illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment.

[0017] Figure 2 This is the second flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment.

[0018] Figure 3 This is the third flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment.

[0019] Figure 4 This is the fourth flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment.

[0020] Figure 5This is the fifth flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment.

[0021] Figure 6 This is a schematic diagram of a global control system for a wind turbine based on stability constraints, according to an exemplary embodiment.

[0022] Figure 7 This is a block diagram illustrating an electronic device according to an exemplary embodiment. Detailed Implementation

[0023] The specific embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit this disclosure.

[0024] It should be noted that all actions involving the acquisition of signals, information, or data in this disclosure are performed with authorization from the owner of the relevant device.

[0025] Figure 1 This is one of the flowcharts illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment. Figure 1 As shown, the method includes the following steps: In step S11, multi-source data during the operation of the wind turbine are collected, and global state variables, disturbance variables, and actuator residual regulation variables are constructed based on the multi-source data. The multi-source data includes at least one of the following: incoming flow data, impeller data, tower data, transmission chain data, converter data, and grid connection status data.

[0026] For example, the stability of wind turbines during operation is affected by a variety of factors, thus requiring multi-dimensional data support for global state analysis and global control of wind turbines. Therefore, multi-source data (at least one of the following: incoming flow data, rotor data, tower data, transmission chain torque sensor, generator speed measurement device, converter sampling unit, and grid connection point voltage and frequency sampling unit) during the operation of the wind turbine can be collected through corresponding sensors in the wind turbine (wind speed sensor, wind direction sensor, blade root load sensor, tower acceleration sensor, transmission chain data, converter data, and grid connection status data). Among them, the incoming flow data can include: wind speed, wind direction, wind shear intensity, and turbulence intensity, etc.; the rotor data can include: rotor speed, blade root bending moment, and blade azimuth angle, etc.; the tower data can include: tower forward acceleration, tower backward acceleration, tower lateral acceleration, and nacelle acceleration, etc.; the transmission chain data can include: low-speed shaft torque, high-speed shaft torque, generator speed, and torque change rate, etc.; the converter data can include: active current, reactive current, DC bus voltage, and current limiting status, etc.; and the grid connection status data can include: grid connection point voltage, grid connection point frequency, phase angle deviation, and grid frequency offset, etc.

[0027] In one possible embodiment, the wind turbine's main control system can operate with a fixed sampling period, such as 20 milliseconds, 50 milliseconds, or 100 milliseconds. Within each sampling period, multi-source data can be acquired through corresponding sensors. However, since the sampling frequencies of the sensors acquiring this multi-source data may differ, time synchronization, outlier removal, and filtering are also required. For example, based on the current sampling period, high-frequency data from the multi-source data can be extracted using window averaging, peak hold, or bandpass filtering; low-frequency data can be aligned to the current control time using hold or linear interpolation; and amplitude or energy within the corresponding mode frequency band can be extracted for tower acceleration and drive train torsional vibration signals using bandpass filtering. For example, for data with different sampling frequencies, such as tower acceleration sampling frequency being higher than wind speed sampling frequency, or converter current sampling frequency being higher than yaw angle sampling frequency, a sliding window resampling method can be used to unify the data to the current sampling period.

[0028] Optionally, the global state variables include at least one of the following: wind turbine speed, generator speed, blade root bending moment, tower forward and backward acceleration, tower lateral acceleration, low-speed shaft torque, high-speed shaft torque, converter active current, converter reactive current, grid connection point voltage, grid connection point frequency, and phase angle deviation. The disturbance includes at least one of the following: wind speed, wind direction, wind shear intensity, turbulence intensity, grid voltage drop amplitude, and grid frequency offset. The remaining adjustment of the actuator includes at least one of the following: remaining pitch angle stroke, remaining pitch rate, remaining generator torque adjustment, remaining yaw angle adjustment, remaining converter active current capacity, and remaining converter reactive current capacity.

[0029] For example, after acquiring the multi-source data, global state variables, disturbance variables, and actuator residual regulation variables can be constructed based on this data. The global state variables can be a set of data from the multi-source data that reflects the operating states of the wind turbine's aerodynamic, structural, transmission chain, electrical, and grid-connected sides. These global state variables may include: rotor speed, generator speed, blade root bending moment, tower forward acceleration, tower backward acceleration, tower lateral acceleration, low-speed shaft torque, high-speed shaft torque, converter active current, converter reactive current, grid-connected voltage, grid-connected frequency, and phase angle deviation, etc. The global state variables can be used to provide a unified state basis for subsequent stability margin prediction and control variable optimization. The disturbance refers to external disturbances or environmental variables that may cause changes in the stability margin of the wind turbine. These disturbances can include wind speed, wind direction, wind shear intensity, turbulence intensity, grid voltage drop, and grid frequency offset. It is understood that these disturbances can be directly collected by anemometers, lidar, nacelle wind vanes, and grid monitoring devices, or estimated from historical sampling data of the wind turbine. For example, if the wind turbine is not equipped with lidar, short-term disturbance trends can be estimated as disturbances based on historical wind speed sequences, rotor speed variations, and blade root moment variations. The remaining actuator regulation refers to the residual capacity of each control actuator of the wind turbine unit that can still be used for stability recovery or power regulation at the current control moment. This remaining actuator regulation can include: remaining pitch angle travel, remaining pitch rate, remaining generator torque regulation, remaining yaw angle regulation, remaining converter active current capacity, and remaining converter reactive current capacity. The remaining actuator regulation prevents the issuance of unexecutable or insufficiently restorative control commands when the actuator is nearing saturation. This remaining actuator regulation can be calculated based on the current actuator position, upper and lower allowable limits, maximum rate of change, and control cycle of each actuator. For example, the remaining pitch angle travel can be determined based on the distance from the current pitch angle to the maximum and minimum allowable pitch angles; the remaining pitch rate can be determined based on the distance from the current pitch rate to the maximum allowable pitch rate; the remaining generator torque regulation can be determined based on the distance from the current torque to the upper and lower torque limits; and the remaining converter active and reactive current capacities can be determined based on the distance from the current vector to the converter current limit.

[0030] In step S12, multiple stable responses are predicted based on the global state quantity and the disturbance quantity, and the stability margin of each stable response is determined based on the prediction results, resulting in multiple stability margins. The multiple stability margins include at least one of the following: aeroelastic stability margin, tower vibration stability margin, transmission chain torsional vibration stability margin, speed stability margin, converter current stability margin, and grid synchronization stability margin.

[0031] For example, the stability response refers to the dynamic response of the wind turbine to the stability of the wind turbine under different disturbances, and the stability margin can refer to the safety margin between a certain stability response and its stability boundary. After determining the stability margin corresponding to each stability response, a corresponding margin stability margin vector can be generated based on the multiple stability margins. The stability response may include at least one of the following: aerodynamic response, structural response, transmission chain response, speed response, current response, and grid connection response; the stability margin may include at least one of the following: aeroelastic stability margin, tower vibration stability margin, transmission chain torsional vibration stability margin, speed stability margin, converter current stability margin, and grid connection synchronization stability margin; the stability response and the stability margin correspond one-to-one. For example, aeroelastic stability margin can be used to characterize the distance between blade root bending moment, blade vibration, or aeroelastic response and the allowable boundary; tower vibration stability margin can be used to characterize the distance between tower acceleration or tower displacement and the allowable boundary; drive train torsional vibration stability margin can be used to characterize the distance between the torsional vibration response of low-speed or high-speed shaft and the allowable boundary; speed stability margin can be used to characterize the distance between wind turbine speed or generator speed and the overspeed boundary; converter current stability margin can be used to characterize the distance between converter current and the limiting boundary; grid-connected synchronization stability margin can be used to characterize the distance between phase angle deviation, grid connection point frequency deviation, or synchronization control state and the allowable boundary.

[0032] This stability margin can be achieved through, for example... Figure 2 The method shown determines that, Figure 2 This is a second flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment. Figure 2 As shown, step S12 includes the following steps: In step S121, a first predicted stable response is determined based on the global state quantity and the disturbance quantity; the first stable response is any one of the plurality of stable responses, and the first predicted stable response is the predicted stable response corresponding to the first stable response.

[0033] For example, a first predicted stable response can be determined based on the global state variable and the perturbation variable. This first predicted stable response can be calculated by a prediction model based on a preset response. The prediction model can be a neural network model or a linear state-space model, with the global state variable and the perturbation variable as inputs and the first predicted stable response as output. This first stable response can be any one of the multiple stable responses.

[0034] In step S122, the stability margin of the first stable response is determined based on the distance between the stability boundary of the first stable response and the first predicted stable response.

[0035] For example, the first Class stability margin can be calculated as follows:

[0036] in, Current control time The Class stability margin; For the first The stability boundary corresponding to the class-steady response at the current control moment; To predict step size The end of the first Predicted stable response (i.e., predicted stable response); To predict the step size; To avoid positive numbers with a denominator of zero; The stability margin is assigned a number. A smaller stability margin indicates that the corresponding stable response is closer to the stability boundary. When the stability margin is less than a preset lower limit, the corresponding stable response may be considered to have entered a risky state. The first stable response can be the [number]th stable response among the multiple stable responses. Stable response.

[0037] For example, the aeroelastic stability margin can be calculated based on the predicted values ​​of blade root bending moment, blade vibration amplitude, or blade load change rate and the allowable stability boundary; the tower vibration stability margin can be calculated based on the predicted values ​​of tower forward and backward acceleration, lateral acceleration, or tower displacement and the allowable stability boundary; the transmission chain torsional vibration stability margin can be calculated based on the predicted values ​​of low-speed shaft torque oscillation amplitude, high-speed shaft torque oscillation amplitude, or torsional vibration velocity and the allowable stability boundary; the speed stability margin can be calculated based on the predicted values ​​of wind turbine speed or generator speed and the overspeed boundary; the converter current stability margin can be calculated based on the predicted values ​​of active current, reactive current, or current vector amplitude and the current limit; and the grid-connected synchronization stability margin can be calculated based on the predicted values ​​of phase angle deviation, frequency deviation, or phase-locked loop synchronization state and the allowable stability boundary.

[0038] In step S13, a coupled instability chain is established based on the hysteresis correlation coefficient between each pair of the multiple stability margins, the normalized coefficient of the prediction sensitivity and the remaining adjustment capability of the actuator.

[0039] For example, the coupled instability chain refers to an ordered link formed by the multiple stability margins according to the direction of instability propagation. The coupled instability chain is not a simple set of constraints, but rather reflects the propagation relationship in which a certain stability margin first decreases and further induces other stability margins to decrease.

[0040] Alternatively, the coupled unstable chain can be achieved through, for example... Figure 3 The method shown determines that, Figure 3 This is the third flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment. Figure 3 As shown, step S13 includes the following steps: In step S131, the link impact coefficient is determined based on the hysteresis correlation coefficient between each pair of the plurality of stability margins, the prediction sensitivity, and the normalization coefficient.

[0041] For example, the link influence coefficient between any two stability margins can be expressed as:

[0042] in, For the first The stability margin points to the first The link impact coefficient with a stability margin; For the first The stability margin relative to the first The lag correlation coefficient of the lagged position with a stability margin is used to characterize the lagged position with a stability margin of . A change in stability margin leads or lags the first The degree of correlation between changes in the stability margin; For the first The change in the stability margin affects the... Predictive sensitivity for changes in stability margin; For the first Each stability margin corresponds to a normalized coefficient of the actuator's residual regulation capability; and All are stability margin numbers, and Not equal to .

[0043] The lag correlation coefficient can be calculated using a sliding time window. For example, within several recent control periods, the change sequences of two stability margins are recorded, and the correlation is calculated under different time lags; when the... The decline in the stability margin leads the time frame of the first... If the stability margin decreases and the correlation between the two is high, then the first stability margin can be considered to be decreasing. The stability margin may affect the first... The stability margin has a propagation effect. This predictive sensitivity can be obtained by perturbing the predicted response corresponding to one stability margin or the related control variable and observing the change in the predicted value of another stability margin. The normalized coefficient of the actuator's residual regulation capability is used to reflect whether the subsequent stability region is easily recovered. When the corresponding actuator's residual regulation capability is low, even if the intensity of the link effect is the same, attention should be paid to this instability propagation side, because once the subsequent stability margin decreases, the margin available for recovery by the controller is small.

[0044] In step S132, the coupling instability chain is established based on the link influence coefficient.

[0045] For example, simply calculating the stability margin itself may still be understood as the parallel monitoring of multiple constraints, while this disclosure identifies the direction of instability propagation and establishes the coupled instability chain by calculating the link influence coefficient based on the hysteresis correlation coefficient, the predictive sensitivity, and the normalized coefficient of the actuator's residual adjustment capability between any two stability margins.

[0046] Optionally, this step S132 can be performed by, for example... Figure 4 The method shown determines that, Figure 4 This is the fourth flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment. Figure 4 As shown, step S132 includes the following steps: In step S1321, when the link influence coefficient between any two stability margins is greater than the preset link determination threshold, an instability propagation edge is established between the two stability margins, pointing from the previous stability margin to the next stability margin.

[0047] In step S1322, the first exhaustion margin among the plurality of stability margins is determined. The first exhaustion margin is the first stability margin among the plurality of stability margins that is less than the corresponding stability margin threshold in the prediction time domain.

[0048] In step S1323, starting from the initial exhaustion margin, the subsequent stability margin is searched sequentially along the unstable propagation edge, and the main propagation path is selected from large to small according to the link influence coefficient to obtain the coupled unstable chain.

[0049] For example, when the link influence coefficient is greater than a preset link judgment threshold, an instability propagation edge is established pointing from the previous stability margin to the next stability margin. Subsequently, in the prediction time domain, the stability margin that first falls below the preset lower limit is identified as the initial exhaustion margin; starting from this initial exhaustion margin, subsequent stability margins are sequentially searched along the instability propagation edge, and the main propagation path is selected according to the link influence coefficient from largest to smallest, thereby obtaining the coupled instability chain. Here, the preset lower limit is the judgment threshold for the corresponding stability margin to enter a risk state, and the main propagation path is the propagation path that contributes the most to the subsequent decrease in stability margin in the prediction time domain.

[0050] For example, under sudden gusts of wind, the aeroelastic stability margin may decrease first, subsequently inducing a decrease in the tower vibration stability margin, then inducing a decrease in the torsional vibration stability margin of the transmission chain, and finally causing a decrease in the rotational speed stability margin. These stability margins can constitute a coupled instability chain on the mechanical side. As another example, under weak grid voltage dips, the converter current stability margin decreases first, further affecting the grid synchronization stability margin, thus forming an electrical-side coupled instability chain pointing from the converter current stability margin to the grid synchronization stability margin.

[0051] In step S14, the chain constraint feasible region of the coupled unstable chain is generated based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator.

[0052] For example, based on the link position of the initial exhaustion margin in the coupled unstable chain, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator, the original stable boundary of the stable response is dynamically tightened to generate a corresponding chain-constrained stable tube. Then, the chain-constrained feasible region is generated by tightening the global feasible region (the set of original stable boundaries of the multiple stable responses) based on the chain-constrained stable tube.

[0053] Optionally, step S14 can be performed as follows: Figure 5 The method shown determines that, Figure 5 This is the fifth flowchart illustrating a global control method for wind turbines based on stability constraints, according to an exemplary embodiment. Figure 5 As shown, step S14 includes the following steps: In step S141, the propagation time of the link is determined based on the sum of the propagation times of two adjacent stability margins in the coupled unstable link.

[0054] For example, the link propagation time can be the sum of the time required for an anomalous change in one stability margin to propagate to an anomalous change in the next stability margin in a coupled instability chain. The shorter the link propagation time, the faster the instability propagates, and the earlier the corresponding controller needs to take conservative control measures.

[0055] In step S142, the dynamic tightening boundary of each stable response is determined based on the stability margin, stability margin exhaustion rate, model prediction residual, remaining adjustment of the actuator, and position in the coupled unstable chain.

[0056] For example, the dynamic tightening boundary of each stable response can be determined based on the maximum tightening amplitude and chain tightening coefficient corresponding to each stable response in the coupled unstable chain. The dynamic tightening boundary of a stable response can be expressed as:

[0057] in, For the first A dynamic tightening boundary for a stable response; For the first The original stability boundary of a stable response; For the first The chain tightening coefficient of a stable response; For the first The maximum tightening range of a stable response; This is the current control moment.

[0058] The chain tightening coefficient can be determined based on the stability margin, stability margin exhaustion rate, model prediction residual, actuator residual adjustment, and the first... The location of a stable response within a coupled unstable chain can be determined. For example, the following normalized combination method can be used:

[0059] in, For the first The chain tightening coefficient of a stable response; This is a limiting function used to restrict the chain tightening coefficient within a preset range; For the first A normalized value with a stability margin; For the first A normalized value of a stable margin exhaustion rate; For the first The normalized value of the predicted residuals for a stable response model; For the first The normalized value of the remaining adjustment of the actuator corresponds to each stable response; For the first A stable response in a coupled unstable chain is a link position factor; , , , and These are non-negative weighting coefficients. Among them, the smaller the stability margin, the larger the exhaustion rate, the larger the model prediction residual, the smaller the actuator residual adjustment, and the closer the stable response is to the initial exhaustion margin, the larger the chain tightening coefficient, and the more obvious the tightening of the corresponding stability boundary.

[0060] Understandably, the model prediction residual refers to the difference between the measured value and the predicted value of the stable response. The larger the model prediction residual, the worse the adaptability of the prediction model to the current operating conditions, and the stronger the stability protection should be. The smaller the residual adjustment of the actuator, the weaker the subsequent recovery capability of the corresponding controller, and the boundary should be tightened in advance.

[0061] In step S143, the chain-constrained stabilizing tube is determined based on the dynamic tightening boundary of each stable response.

[0062] For example, a chain-constrained stabilizing tube can refer to a set of constraints formed by dynamically tightening the original stability boundary of the stable response based on the initial exhaustion margin, link propagation time, model prediction residuals, and actuator residual regulation in the coupled unstable chain. The chain-constrained stabilizing tube is used to restrict candidate control variables in advance before the instability propagation is complete, preventing the controller from exacerbating subsequent stability margin exhaustion in pursuit of power tracking. For example, in a mechanically coupled unstable chain, if the aeroelastic stability margin is the initial exhaustion margin and the link propagation time is short, the blade root bending moment boundary, tower vibration boundary, and drive train torsional vibration boundary can be tightened preferentially. Thus, when solving for subsequent candidate control variables, the controller will not continue to adopt control strategies that might increase blade root loads or tower excitation, such as increasing generator torque too quickly or pursuing the power reference too aggressively. In the electrical-side coupling instability chain, if the converter current stability margin is the initial exhaustion margin and the model prediction residual increases, the converter current boundary and grid synchronization boundary can be tightened first to reduce the active current demand or adjust the reactive current distribution, so as to avoid grid synchronization deterioration after the current limiting trigger.

[0063] In step S144, the chain-constrained feasible region is generated based on the chain-constrained stabilizing tube, the global feasible region of the coupled unstable chain, the initial exhaustion margin, and the link propagation time.

[0064] For example, after the chain-constrained stabilizing transistor is generated, the global feasible region of the coupled unstable chain can be tightened into a chain-constrained feasible region based on the chain-constrained stabilizing transistor. For instance, the intersection of the chain-constrained stabilizing transistor and the global feasible region is calculated, and the initial exhaustion margin is dynamically scaled using its position in the coupled unstable chain and the propagation time of the link to generate the chain-constrained feasible region. The chain-constrained feasible region can be jointly defined by the chain-constrained stabilizing transistor, actuator amplitude constraints, actuator rate constraints, speed constraints, load constraints, converter current constraints, and grid synchronization constraints. Compared with the conventional feasible region, the chain-constrained feasible region is not a fixed set of safety boundaries, but a dynamic feasible region that changes in real time with the coupled unstable chain. Its essence is that when a certain stability margin decrease may propagate rapidly along the link, the control actions related to the propagation chain are restricted in advance to avoid taking protective actions only after the subsequent stability response exceeds the limit.

[0065] In step S15, the candidate global control variables of the wind turbine are solved within the feasible region of the chain constraint.

[0066] For example, model predictive control, rolling time-domain optimization, quadratic programming, sequential quadratic programming, or constrained fast optimization algorithms can be used to solve for the candidate global control variables of the wind turbine within the feasible domain of the chain constraints.

[0067] Optionally, step S15 may include the following steps: using the power tracking deviation, the penalty for the decrease in stability margin of each stability margin in the coupled instability chain, the tower vibration penalty, the transmission chain torsional vibration penalty, and the actuator action change penalty as joint objectives, and using the feasible region of the chain constraint as the constraint condition, to solve for the candidate global control quantity; the candidate global control quantity includes at least one of the following: collective pitch angle, independent pitch correction of each blade of the wind turbine, generator torque, yaw offset angle, converter active current command, and converter reactive current command.

[0068] For example, candidate global control variables are solved within the feasible region of the chain constraint. These candidate global control variables include at least one of the following: collective pitch angle, independent pitch correction for each blade of the wind turbine, generator torque, yaw offset angle, converter active current command, and converter reactive current command.

[0069] For example, the candidate global control quantity can be solved by taking the power tracking deviation, the penalty for the decrease in stability margin of each stability margin in the coupled instability chain, the tower vibration penalty, the torsional vibration penalty of the transmission chain, and the penalty for the change in actuator action as joint objectives, and using the feasible region of the chain constraint as the constraint condition. The joint objective can be expressed by the following objective function:

[0070] in, The joint objective function; Number the prediction step; To predict the step size; For the first Predicted active power for each prediction step; For the first Active power reference value for each prediction step; For power tracking weights; For the first A penalty weight with a stable margin; The function is a penalty function for decreasing stability margin; For the first The amplitude of the tower vibration response in each predicted step; For the first The amplitude of the torsional vibration response of the transmission chain in each predicted step; Weighting for tower vibration penalties; Weight for torsional vibration penalty of the transmission chain; The penalty weight for changes in actuator actions; for No. The change in control quantity in each prediction step relative to the previous prediction step; It is a 2-norm.

[0071] Understandably, when the coupled instability chain has not formed or the stability margins are all in a safe state, the weights of power point tracking and actuator action changes can be increased; when the coupled instability chain forms and the initial exhaustion margin is close to the risk boundary, the penalty weights of the relevant stability margins in the coupled instability chain can be increased, making the optimization results tend to reduce load, vibration, torsional vibration, current, or synchronization risks. Therefore, this disclosure can achieve the following under normal operating conditions without sacrificing power generation performance, and under risky operating conditions, it can actively relinquish some power point tracking performance in exchange for the overall stability of the wind turbine.

[0072] Furthermore, this method can also be executed using a rolling time-domain approach. That is, within each control cycle, using the current global state as the initial state, the stable response and power response for the next few steps are predicted, the future control sequence is solved, only the first set of control variables for the current cycle is executed, and data is reacquired and the control variables are solved again in the next control cycle. This rolling optimization method can adapt to rapid changes in wind conditions and grid conditions.

[0073] In step S16, when the candidate global control quantity touches the boundary of the feasible region of the chain constraint, a control repair quantity is allocated along the reverse recovery path of the coupled unstable chain, and the target global control quantity is determined based on the candidate global control quantity and the control repair quantity.

[0074] For example, the reverse recovery path refers to the control repair sequence opposite to the instability propagation direction of the coupled instability chain. This path is used to determine the repair priority of different actuators, so that the control repair quantity is no longer allocated according to the fixed limiting rule, but is allocated along the reverse recovery path of the coupled instability chain, and the target global control quantity is output. Unlike traditional solutions that only limit the amplitude of each actuator, this disclosure can determine the control repair sequence according to the instability propagation direction.

[0075] For example, in a mechanically coupled instability chain, if the coupled instability chain represents the sequential depletion of aeroelastic stability margin, tower vibration stability margin, drivetrain torsional vibration stability margin, and speed stability margin, it indicates that the instability risk gradually propagates from aerodynamic loads to structural vibration, drivetrain torsional vibration, and speed overshoot. In this case, the reverse recovery path is opposite to the propagation direction, but the control repair does not simply start from the last margin; instead, it comprehensively considers the upstream suppression effect of the recovery action. Preferably, firstly, the blade root moment imbalance and tower excitation are reduced by independent pitch correction, thus decreasing the instability energy at the aerodynamic input end; then, the drivetrain torsional vibration excitation is reduced by limiting the generator torque slope or adjusting the generator torque command; finally, the control speed is adjusted by collective pitch angle to bring the turbine speed and generator speed back to a safe range. This repair sequence can reduce the adverse situations where simply reducing speed through collective pitch leads to a sudden increase in tower load or where simply limiting torque leads to a continued increase in speed.

[0076] For example, in a coupled instability chain between the electrical and grid-connected sides, if the coupled instability chain indicates that the converter current stability margin and the grid synchronization stability margin are successively exhausted, it suggests that insufficient current capacity may further affect grid synchronization. In this case, priority should be given to adjusting the allocation ratio of the converter's active current command and reactive current command. For example, while meeting the grid support requirements, reduce the active current demand and increase the necessary reactive power support capacity; at the same time, limit the generator torque rise rate to reduce the power surge from the mechanical side to the electrical side and avoid further deterioration of the grid synchronization margin due to converter current limiting.

[0077] In one possible embodiment, the control repair quantity can be allocated according to the recovery sensitivity of each actuator to the triggered stability margin. For cases where candidate global control quantities touch the feasible domain boundary of the chain constraint, the stability response number corresponding to the touching boundary can be determined first, then the upstream propagation node and downstream affected node can be determined according to the coupled instability chain, and finally, a priority actuator can be selected from a preset actuator repair table. For example, aeroelastic stability margin and tower vibration stability margin preferentially correspond to independent pitch control; transmission chain torsional vibration stability margin preferentially corresponds to generator torque slope limitation; speed stability margin preferentially corresponds to collective pitch control; converter current stability margin preferentially corresponds to active and reactive current redistribution; and grid-connected synchronization stability margin preferentially corresponds to reactive power support and torque rise rate limitation.

[0078] In one possible embodiment, the target global control variable can be determined jointly by the candidate global control variable and the control repair variable. For example:

[0079] in, The target is the global control variable; Candidate global control variables; The control repair amount is allocated along the reverse recovery path; This is the current control moment.

[0080] Furthermore, to avoid abrupt actuator actions caused by this control correction, the control correction can be rate-limited and smoothed. For example, a maximum rate of change can be set for the pitch correction, a slope limit can be set for the generator torque correction, and current vector limits can be set for active and reactive current commands. Finally, after determining the target global control quantity, it can be sent to the wind turbine's pitch system, generator torque control system, yaw system, and converter control system, so that these systems output the corresponding control quantities from the target global control quantity.

[0081] The above technical solution collects multi-source data during the operation of the wind turbine, and constructs global state variables, disturbance variables, and actuator residual regulation variables based on this multi-source data. This multi-source data includes at least one of the following: incoming flow data, impeller data, tower data, drive train data, converter data, and grid connection status data. Multiple stable responses are predicted based on the global state variables and disturbance variables, and the stability margin of each stable response is determined based on the prediction results, resulting in multiple stability margins. These multiple stability margins include: aeroelastic stability margin, tower vibration stability margin, drive train torsional vibration stability margin, speed stability margin, converter current stability margin, and grid connection synchronization stability margin. The method involves determining at least one of the stability margins; establishing a coupled instability chain based on the hysteresis correlation coefficients, prediction sensitivity, and normalized coefficients of the actuator's residual regulation capacity between each pair of these multiple stability margins; generating a chain-constrained feasible region for the coupled instability chain based on the initial exhaustion margin, the link propagation time of the coupled instability chain, the model prediction residual, and the actuator's residual regulation; solving for candidate global control variables for the wind turbine within this chain-constrained feasible region; and when the candidate global control variable touches the boundary of the chain-constrained feasible region, allocating a control repair amount along the reverse recovery path of the coupled instability chain, and determining the target global control variable based on the candidate global control variable and the control repair amount. This method can improve the global stability of wind turbines under complex conditions and is beneficial for extending the lifespan of key components of wind turbines and improving the long-term power generation revenue of the turbine.

[0082] Figure 6This is a schematic diagram illustrating a global control system for a wind turbine based on stability constraints, according to an exemplary embodiment. Figure 6 As shown, the system includes: a data acquisition module, a global state construction module, a stability margin prediction module, a coupled unstable chain generation module, a chain constraint generation module, a global optimization module, and a reverse repair allocation module; This data acquisition module is used to collect multi-source data during the operation of the wind turbine. The multi-source data includes at least one of the following: incoming flow data, rotor data, tower data, transmission chain data, converter data, and grid connection status data. This global state construction module is used to construct global state quantities, disturbance quantities, and actuator residual adjustment quantities based on the multi-source data; The stability margin prediction module is used to predict multiple stable responses based on the global state quantity and the disturbance quantity, and to determine the stability margin of each stable response based on the prediction results, thereby obtaining multiple stability margins; the multiple stability margins include at least one of the following: aeroelastic stability margin, tower vibration stability margin, transmission chain torsional vibration stability margin, speed stability margin, converter current stability margin, and grid synchronization stability margin. The coupled instability chain generation module is used to establish coupled instability chains based on the hysteresis correlation coefficients, prediction sensitivity, and normalized coefficients of the remaining adjustment capability of the actuators between each pair of the multiple stability margins. The chain constraint generation module is used to generate the feasible region of chain constraints for the coupled unstable chain based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator. This global optimization module is used to solve for candidate global control variables of the wind turbine within the feasible region of the chain constraint. The reverse repair allocation module is used to allocate a control repair amount along the reverse recovery path of the coupled unstable chain when the candidate global control amount touches the boundary of the feasible region of the chain constraint, and to determine the target global control amount based on the candidate global control amount and the control repair amount.

[0083] Optionally, the coupling unstable chain generation module includes: a link influence coefficient calculation unit and a coupling unstable chain generation unit; The link impact coefficient calculation unit is used to determine the link impact coefficient based on the lag correlation coefficient between each pair of the multiple stability margins, the prediction sensitivity, and the normalization coefficient. The coupling unstable chain generation unit is used to establish the coupling unstable chain based on the link influence coefficient.

[0084] Optionally, the system may also include: an execution interface module; The execution interface module is used to output control commands to at least one of the pitch system, generator torque control system, yaw system and converter control system, so that the pitch system, generator torque control system, yaw system and converter control system output the corresponding control quantity in the target global control quantity.

[0085] The above technical solution collects multi-source data during the operation of the wind turbine, and constructs global state variables, disturbance variables, and actuator residual regulation variables based on this multi-source data. This multi-source data includes at least one of the following: incoming flow data, impeller data, tower data, drive train data, converter data, and grid connection status data. Multiple stable responses are predicted based on the global state variables and disturbance variables, and the stability margin of each stable response is determined based on the prediction results, resulting in multiple stability margins. These multiple stability margins include: aeroelastic stability margin, tower vibration stability margin, drive train torsional vibration stability margin, speed stability margin, converter current stability margin, and grid connection synchronization stability margin. The method involves determining at least one of the stability margins; establishing a coupled instability chain based on the hysteresis correlation coefficients, prediction sensitivity, and normalized coefficients of the actuator's residual regulation capacity between each pair of these multiple stability margins; generating a chain-constrained feasible region for the coupled instability chain based on the initial exhaustion margin, the link propagation time of the coupled instability chain, the model prediction residual, and the actuator's residual regulation; solving for candidate global control variables for the wind turbine within this chain-constrained feasible region; and when the candidate global control variable touches the boundary of the chain-constrained feasible region, allocating a control repair amount along the reverse recovery path of the coupled instability chain, and determining the target global control variable based on the candidate global control variable and the control repair amount. This method can improve the global stability of wind turbines under complex conditions and is beneficial for extending the lifespan of key components of wind turbines and improving the long-term power generation revenue of the turbine.

[0086] Regarding the system in the above embodiments, the specific ways in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0087] Figure 7 This is a block diagram illustrating an electronic device 700 according to an exemplary embodiment. Figure 7 As shown, the electronic device 700 may include a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an input / output (I / O) interface 704, and a communication component 705.

[0088] The processor 701 controls the overall operation of the electronic device 700 to complete all or part of the steps in the aforementioned global control method for wind turbine generators based on stability constraints. The memory 702 stores various types of data to support the operation of the electronic device 700. This data may include, for example, instructions for any application or method operating on the electronic device 700, and application-related data such as contact data, sent and received messages, pictures, audio, video, etc. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. Multimedia component 703 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in memory 702 or transmitted via communication component 705. The audio component also includes at least one speaker for outputting audio signals. I / O interface 704 provides an interface between processor 701 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual or physical buttons. Communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IoT, eMTC, or other 5G technologies, or combinations thereof, is not limited here. Therefore, the corresponding communication component 705 may include: a Wi-Fi module, a Bluetooth module, an NFC module, etc.

[0089] In an exemplary embodiment, the electronic device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the above-described global control method for wind turbine generators based on stability constraints.

[0090] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the stability-constraint-based global control method for wind turbines described above. For example, the computer-readable storage medium may be the memory 702 including program instructions described above, which may be executed by the processor 701 of the electronic device 700 to complete the stability-constraint-based global control method for wind turbines described above.

[0091] In another exemplary embodiment, a computer program product is also provided, comprising a computer program executable by a programmable device, the computer program having a code portion for performing the above-described global control method for wind turbines based on stability constraints when executed by the programmable device.

[0092] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and all such simple modifications fall within the protection scope of this disclosure. It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0093] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

Claims

1. A global control method for wind turbine generators based on stability constraints, characterized in that, include: Collect multi-source data during the operation of the wind turbine, and construct global state variables, disturbance variables, and actuator residual regulation variables based on the multi-source data; the multi-source data includes: incoming flow data, impeller data, tower data, drive train data, converter data, and grid connection status data; Multiple stable responses are predicted based on the global state variables and the disturbance variables, and the stability margin of each stable response is determined based on the prediction results, resulting in multiple stability margins; the multiple stability margins include: aeroelastic stability margin, tower vibration stability margin, transmission chain torsional vibration stability margin, speed stability margin, converter current stability margin, and grid synchronization stability margin. A coupled instability chain is established based on the hysteresis correlation coefficients between each pair of the multiple stability margins, the normalized coefficients of the prediction sensitivity and the remaining adjustment capability of the actuator; Based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator, the chain constraint feasible region of the coupled unstable chain is generated. Solve for the candidate global control variables of the wind turbine within the feasible region of the chain constraints; When the candidate global control quantity touches the boundary of the feasible region of the chain constraint, a control repair quantity is allocated along the reverse recovery path of the coupled unstable chain, and a target global control quantity is determined based on the candidate global control quantity and the control repair quantity. The step of establishing a coupled instability chain based on the hysteresis correlation coefficients, predictive sensitivity, and normalized coefficients of the remaining adjustment capability of the actuators among the pairwise stability margins includes: The link impact coefficient is determined based on the hysteresis correlation coefficient between each pair of the multiple stability margins, the prediction sensitivity, and the normalization coefficient. When the link influence coefficient between any two stability margins is greater than the preset link determination threshold, an instability propagation edge is established between the two stability margins, pointing from the previous stability margin to the next stability margin. Determine the first exhaustion margin among the plurality of stability margins, wherein the first exhaustion margin is the first stability margin among the plurality of stability margins that is less than the corresponding stability margin threshold in the prediction time domain. Starting from the initial exhaustion margin, the subsequent stability margin is searched sequentially along the unstable propagation edge, and the main propagation path is selected from large to small according to the link influence coefficient to obtain the coupled unstable chain.

2. The method according to claim 1, characterized in that, The global state variables include: wind turbine speed, generator speed, blade root bending moment, tower forward and backward acceleration, tower lateral acceleration, low-speed shaft torque, high-speed shaft torque, converter active current, converter reactive current, grid connection point voltage, grid connection point frequency, and phase angle deviation. The disturbance quantities include wind speed, wind direction, wind shear intensity, turbulence intensity, grid voltage drop amplitude, and grid frequency offset. The remaining adjustment of the actuator includes: remaining pitch angle stroke, remaining pitch rate, remaining generator torque adjustment, remaining yaw angle adjustment, remaining active current capacity of the converter, and remaining reactive current capacity of the converter.

3. The method according to claim 1, characterized in that, The step of predicting multiple stable responses based on the global state variables and the perturbation variables, and determining the stability margin of each stable response based on the prediction results, includes: A first predicted stable response is determined based on the global state quantity and the disturbance quantity; the first stable response is any one of the plurality of stable responses, and the first predicted stable response is the predicted stable response corresponding to the first stable response. The stability margin of the first stable response is determined based on the distance between the stability boundary of the first stable response and the first predicted stable response.

4. The method according to claim 1, characterized in that, The step of generating the chain constraint feasible region of the coupled unstable chain based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator includes: The link propagation time is determined by the sum of the propagation times required for two adjacent stability margins in the coupled unstable chain; The dynamic tightening boundary of each stable response is determined based on the stability margin, stability margin exhaustion rate, model prediction residual, remaining actuator adjustment, and position in the coupled instability chain corresponding to each stable response. The chain-constrained stabilizing tube is determined based on the dynamically tightening boundary of each stable response; The chain-constrained feasible region is generated based on the chain-constrained stabilizing tube, the global feasible region of the coupled unstable chain, the initial exhaustion margin, and the link propagation time.

5. The method according to claim 1, characterized in that, Solving for the candidate global control variables of the wind turbine within the feasible region of the chain constraints includes: Using power tracking deviation, the penalty for the decrease in stability margin of each stability margin in the coupled instability chain, tower vibration penalty, transmission chain torsional vibration penalty, and actuator action change penalty as joint objectives, and the feasible region of the chain constraint as constraint conditions, the candidate global control quantities are solved. The candidate global control quantities include: collective pitch angle, independent pitch correction of each blade of the wind turbine, generator torque, yaw offset angle, converter active current command, and converter reactive current command.

6. A global control system for wind turbine generators based on stability constraints, characterized in that, The system includes: a data acquisition module, a global state construction module, a stability margin prediction module, a coupled unstable chain generation module, a chain constraint generation module, a global optimization module, and a reverse repair allocation module; The data acquisition module is used to collect multi-source data during the operation of the wind turbine. The multi-source data includes: incoming flow data, impeller data, tower data, transmission chain data, converter data, and grid connection status data. The global state construction module is used to construct global state quantities, disturbance quantities, and actuator residual adjustment quantities based on the multi-source data. The stability margin prediction module is used to predict multiple stable responses based on the global state quantity and the disturbance quantity, and to determine the stability margin of each stable response based on the prediction results, thereby obtaining multiple stability margins; the multiple stability margins include: aeroelastic stability margin, tower vibration stability margin, transmission chain torsional vibration stability margin, speed stability margin, converter current stability margin, and grid synchronization stability margin. The coupled instability chain generation module is used to establish coupled instability chains based on the hysteresis correlation coefficients, prediction sensitivity, and normalized coefficients of the remaining adjustment capability of the actuators between each pair of the multiple stability margins. The chain constraint generation module is used to generate the feasible region of chain constraints for the coupled unstable chain based on the initial exhaustion margin, the link propagation time of the coupled unstable chain, the model prediction residual, and the remaining adjustment of the actuator. The global optimization module is used to solve for the candidate global control variables of the wind turbine within the feasible region of the chain constraint. The reverse repair allocation module is used to allocate a control repair amount along the reverse recovery path of the coupled unstable chain when the candidate global control amount touches the boundary of the feasible region of the chain constraint, and to determine the target global control amount based on the candidate global control amount and the control repair amount. The step of establishing a coupled instability chain based on the hysteresis correlation coefficients, predictive sensitivity, and normalized coefficients of the remaining adjustment capability of the actuators among the pairwise stability margins includes: The link impact coefficient is determined based on the hysteresis correlation coefficient between each pair of the multiple stability margins, the prediction sensitivity, and the normalization coefficient. When the link influence coefficient between any two stability margins is greater than the preset link determination threshold, an instability propagation edge is established between the two stability margins, pointing from the previous stability margin to the next stability margin. Determine the first exhaustion margin among the plurality of stability margins, wherein the first exhaustion margin is the first stability margin among the plurality of stability margins that is less than the corresponding stability margin threshold in the prediction time domain. Starting from the initial exhaustion margin, the subsequent stability margin is searched sequentially along the unstable propagation edge, and the main propagation path is selected from large to small according to the link influence coefficient to obtain the coupled unstable chain.

7. The system according to claim 6, characterized in that, The coupling unstable chain generation module includes: a link influence coefficient calculation unit and a coupling unstable chain generation unit; The link impact coefficient calculation unit is used to determine the link impact coefficient based on the lag correlation coefficient between each pair of the plurality of stability margins, the prediction sensitivity, and the normalization coefficient. The coupling unstable chain generation unit is used to establish the coupling unstable chain based on the link influence coefficient.

8. The system according to claim 6, characterized in that, The system also includes: an execution interface module; The execution interface module is used to output control commands to the pitch system, generator torque control system, yaw system and converter control system, so that the pitch system, the generator torque control system, yaw system and converter control system output the corresponding control quantity in the target global control quantity.

Images