Adaptive variable pitch control method and system based on laser wind lidar

Abstract

The application relates to the technical field of wind turbine control, and discloses an adaptive variable pitch control method and system based on a laser wind measurement radar, which comprises the following steps: calculating an initial thrust coefficient in combination with a two-dimensional interpolation table of aerodynamic thrust coefficients, determining an optimal reference distance gate and a far-end reference distance gate in a radar distance gate sequence according to the initial thrust coefficient; extracting synchronous measurement wind speeds of the two gates to construct a first-order difference model, extrapolating and calculating an equivalent incoming flow wind speed, and obtaining a smooth wind wheel plane effective wind speed through axial induction factor correction; limiting the effective wind speed to generate a controlled input wind speed, combining electromechanical operation state parameters, performing aerodynamic performance degradation diagnosis and thrust coefficient compensation update based on torque residual error, and finally generating a total variable pitch instruction. The application actively avoids the wind wheel blocking effect to improve the prediction accuracy of the feedforward wind speed, can online adaptively compensate the blade aerodynamic performance attenuation, effectively reduces the mechanical fatigue load of the wind turbine, and guarantees the operation stability of the unit.

Description

Technical Field

[0001] This invention relates to the field of wind turbine control technology, specifically to an adaptive pitch control method and system based on laser wind radar. Background Technology

[0002] To reduce fatigue loads and improve operational stability of wind turbine generators, control systems typically incorporate nacelle-mounted laser wind-measuring radar for feedforward pitch control. The laser radar can detect the incoming wind speed upstream of the rotor in advance, allowing the main control system to adjust the pitch angle before gusts arrive, thus mitigating mechanical shock.

[0003] However, the rotation of the wind turbine generates aerodynamic induction effects on the forward flow field, causing wind speed to decrease nonlinearly as it approaches the turbine plane. Existing radar feedforward control methods typically select measurement data at a fixed range threshold, failing to fully incorporate the real-time aerodynamic thrust of the turbine to dynamically define and avoid areas affected by flow field blockage. This static extraction method makes it difficult to accurately establish a spatial evolution model of wind speed, resulting in inherent biases in the calculated effective wind speed at the turbine plane.

[0004] Meanwhile, during long-term operation, wind turbine blades inevitably experience contamination, icing, or wear, leading to a gradual degradation of aerodynamic performance. Existing pitch control strategies largely rely on factory-preset aerodynamic parameter models, lacking online diagnostics and parameter adaptation capabilities. When the actual aerodynamic characteristics of the blades deviate from the theoretical model, the feedforward gain and correction factor calculated based on the original model will fail to match the actual operating conditions. This parameter mismatch not only weakens the load-reducing effect of feedforward control but also causes excessive or insufficient pitch, increasing wear on the drivetrain.

[0005] Furthermore, in complex wind conditions, alignment deviations or signal distortions can easily occur between radar wind measurement data and the actual wind speed sensed by the engine nacelle. If the feedforward control system lacks real-time verification of data correlation and anomaly interlocking mechanisms, directly inputting wind speed signals that deviate from the actual values ​​into the control loop can easily lead to erroneous pitch command outputs, thereby affecting the safety and stability of the entire engine operation. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides an adaptive pitch control method and system based on laser wind radar. This solves the problems in existing laser radar-based feedforward pitch control technology, which fails to fully consider the dynamic impact of wind turbine aerodynamic induction effects on upstream wind speed evolution, leading to deviations in the wind speed prediction of the wind turbine plane. Furthermore, when the blades experience aerodynamic performance degradation, the fixed feedforward control parameters cannot adapt to the actual operating conditions, resulting in decreased control accuracy and increased mechanical fatigue load.

[0007] To achieve the above objectives, the present invention provides the following technical solution: The first aspect of the present invention provides an adaptive pitch control method based on a laser wind radar, comprising:

[0008] The initial thrust coefficient is calculated using a two-dimensional interpolation table of aerodynamic thrust coefficients.

[0009] Determine the optimal reference range gate and the far-end reference range gate in the range gate sequence based on the initial thrust coefficient;

[0010] Extract the synchronously measured wind speeds on the optimal reference distance gate and the far-end reference distance gate, and construct a first-order difference model to calculate the equivalent incoming wind speed;

[0011] The equivalent incoming wind speed is corrected by using the axial induction factor corresponding to the initial thrust coefficient, and the effective wind speed in the smooth rotor plane is calculated.

[0012] A controlled input wind speed is generated by limiting the effective wind speed in the smooth wind turbine plane, and aerodynamic performance degradation diagnosis and compensation update are performed in combination with electromechanical operating status parameters;

[0013] The total pitch command is generated based on the effective wind speed in the smooth rotor plane and the controlled input wind speed.

[0014] Furthermore, by combining the two-dimensional interpolation table of aerodynamic thrust coefficients, the initial thrust coefficient is calculated, including: obtaining the actual pitch angle feedback value and the current tip speed ratio;

[0015] The current tip speed ratio is calculated based on the generator speed feedback value, gearbox transmission ratio, impeller radius, and equivalent incoming wind speed of the previous control cycle; the actual pitch angle feedback value and the current tip speed ratio are limited, and four adjacent known nodes surrounding the two are determined in the two-dimensional interpolation table of aerodynamic thrust coefficient;

[0016] Calculate the lateral and longitudinal distance weights of the actual pitch angle feedback value and the current tip speed ratio relative to the rectangular grid boundary formed by the four nodes; use the lateral and longitudinal distance weights to perform a weighted summation of the preset thrust coefficient values ​​corresponding to the four nodes to obtain the initial thrust coefficient.

[0017] Furthermore, the optimal reference range gate and the far-end reference range gate are determined in the range gate sequence based on the initial thrust coefficient, including: calculating the estimated forward extension depth of the aerodynamic induced zone based on the initial thrust coefficient and the impeller diameter, and outputting the forward boundary distance of the aerodynamic induced zone after limiting the amplitude.

[0018] Read the distance gate sequence and the corresponding data quality status; include distance gates whose spatial distance is not less than the sum of the forward boundary distance of the aerodynamic induction zone and the preset safety margin into the candidate set;

[0019] After removing invalid distance gates from the candidate set that are below the preset confidence threshold, the distance gate with the smallest distance difference from the forward boundary of the aerodynamic induction zone is selected as the optimal reference distance gate.

[0020] Among the effective distance gates whose spatial distance is greater than the spatial distance corresponding to the optimal reference distance gate and whose spacing is not less than the minimum span threshold, the preset distance gate is selected as the far-end reference distance gate.

[0021] By dynamically calculating the forward boundary of the aerodynamic induction zone using the thrust coefficient, the distance gate for subsequent wind speed extraction can actively avoid the near-end flow field region affected by the wind turbine blockage effect, thereby obtaining a representative free-flow wind speed.

[0022] Furthermore, the synchronously measured wind speeds at the optimal reference distance gate and the far reference distance gate are extracted, and a first-order difference model is constructed to calculate the equivalent incoming wind speed, including: obtaining the first spatial distance and the second spatial distance corresponding to the optimal reference distance gate and the far reference distance gate;

[0023] Radar radial wind speeds on the optimal reference range gate and the far-end reference range gate are converted into axial wind speeds. Adjacent historical data points are locked in the historical cache queue in combination with the transport time, and interpolation is used to obtain the first matching wind speed and the second matching wind speed. The difference between the second spatial distance and the first spatial distance is calculated as the distance difference, and the difference between the second matching wind speed and the first matching wind speed is calculated as the wind speed difference. The wind speed difference is divided by the distance difference to obtain the wind speed evolution gradient.

[0024] Based on the first spatial distance, the second spatial distance, and the corresponding matching wind speed, a spatial linear equation is constructed. Using the wind speed evolution gradient as the slope, the intercept at the origin is solved to obtain the equivalent incoming wind speed and limit its amplitude. This process utilizes the spatial gradient characteristics detected by multiple range gates of the radar to establish a spatial linear difference model and solve for the origin intercept, realizing the extrapolation calculation of wind speed from the radar spatial measurement section to the wind turbine plane.

[0025] Furthermore, the equivalent incoming wind speed is corrected using an axial induction factor corresponding to the initial thrust coefficient, and the effective wind speed in the smoothed rotor plane is calculated, including:

[0026] Perform amplitude limiting verification on the initial thrust coefficient and output the verification thrust coefficient; calculate the difference between value 1 and the verification thrust coefficient and take the arithmetic square root; subtract the arithmetic square root from value 1 to obtain the numerator; divide the numerator by value 2 to obtain the estimated value; and output the axial induction factor after amplitude limiting of the estimated value.

[0027] The attenuation coefficient is constructed by subtracting the axial induction factor from the numerical value of 1. The effective wind speed in the rotor plane is obtained by multiplying the equivalent incoming wind speed by the attenuation coefficient.

[0028] The filtering weight coefficients are calculated based on the mechanical response inertia time. Discrete difference iterative calculation is then performed on the effective wind speed in the wind turbine plane using the filtering weight coefficients to output a smoothed effective wind speed in the wind turbine plane.

[0029] Furthermore, a controlled input wind speed is generated by limiting the effective wind speed in the smoothed impeller plane, and aerodynamic performance degradation diagnosis and compensation updates are performed in conjunction with electromechanical operating status parameters, including:

[0030] Establish a data window for the effective wind speed in the smooth wind turbine plane and calculate the statistical variance; when the variance is greater than the preset dead zone threshold, generate a feedforward control enable flag and perform abnormal jump amplitude limiting on the latest input effective wind speed to generate a controlled input wind speed.

[0031] Align the controlled input wind speed with the estimated wind speed sequence in the nacelle by interpolation, calculate the time bias based on the current average wind speed of the estimated wind speed sequence in the nacelle, and extract the cached controlled input wind speed sequence by translation; calculate the bias correlation coefficient between the aligned controlled input wind speed sequence and the estimated wind speed sequence in the nacelle.

[0032] When the bias correlation coefficient is not greater than the correlation locking threshold within a continuous preset period, the interlock flag is set to logic 1 and the set correlation correction coefficient is set to 0.

[0033] Furthermore, the aerodynamic performance degradation diagnosis and compensation update combined with electromechanical operating status parameters also includes: extracting electromechanical operating status parameters for logical diagnosis, and determining whether the state machine enters the monitoring or frozen state by combining interlock flags.

[0034] When in monitoring mode, the current tip speed ratio is calculated based on the impeller angular velocity, impeller radius and equivalent incoming wind speed, and the torque residual is calculated by combining the difference between the actual aerodynamic torque and the model aerodynamic torque; the torque residual is low-pass filtered, and when the absolute value of the filtered residual is continuously greater than the preset judgment threshold and the signs are consistent, it is determined that there is aerodynamic performance degradation.

[0035] When degradation exists, the actual pitch angle is obtained. The current tip speed ratio and the actual pitch angle are used as coordinates. The compensation sensitivity coefficient is extracted and multiplied with the torque residual to obtain the thrust coefficient compensation amount. The thrust coefficient compensation amount is multiplied by the learning rate and accumulated into the four nodes surrounding the current coordinates in the preset thrust coefficient compensation table according to the bilinear interpolation weight to complete the update.

[0036] The above-mentioned diagnosis and compensation update process performs logical diagnosis by extracting the residual between the actual operating torque of the generator and the torque of the theoretical model, quantitatively identifying the aerodynamic performance degradation caused by blade contamination or wear, and updating the aerodynamic thrust coefficient table online when the state machine permits, so that the feedforward control algorithm has self-learning and adaptive compensation capabilities.

[0037] Furthermore, the total pitch command is generated based on the effective wind speed in the smoothed rotor plane and the controlled input wind speed, including: using the effective wind speed in the smoothed rotor plane after sliding filtering as the reference wind speed, extracting the basic feedforward gain coefficient from the preset one-dimensional gain table; calculating the real-time air density, and calculating the density and thrust correction factor in combination with the initial thrust coefficient.

[0038] The basic feedforward gain coefficient is multiplied by the density and thrust correction factors and the correlation correction factor in sequence to generate the dynamic gain coefficient; the controlled input wind speed is subtracted from the rated wind speed to obtain the wind speed deviation, and the wind speed deviation is multiplied by the dynamic gain coefficient to obtain the additional pitch angle.

[0039] When the feedforward control enable flag is logic 1, the interlock flag is logic 0, and the state machine is not frozen, the pitch rate limit is applied to the additional pitch angle to generate a valid feedforward additional pitch angle; the valid feedforward additional pitch angle is added to the acquired basic absolute pitch angle command, and after limiting, the total pitch command is generated.

[0040] A second aspect of the present invention provides an adaptive pitch control system based on a laser wind radar, comprising:

[0041] The status acquisition module is distributed and installed on the drive chain and electrical circuit of the wind turbine generator set to collect electromechanical operating status parameters.

[0042] The radar measurement module is rigidly mounted on the top of the nacelle of the wind turbine generator and is used to output wind speed data on multiple spatial depth sections.

[0043] The pitch drive module is used to drive the blades to perform pitch angle adjustment actions;

[0044] The main control and computing module is communicatively connected to the above-mentioned module and is used to execute the adaptive pitch control method as described in the first aspect of the present invention, and generate a total pitch command to be sent to the pitch drive module.

[0045] Furthermore, the main control computing module opens multiple ring buffers in the internal memory and pushes in the electromechanical operating status parameters with hardware clock timestamps; taking the current task trigger pulse as the reference time point, it traverses the timestamps, extracts the adjacent historical data points with the smallest time difference, performs linear interpolation, and obtains the synchronized and aligned electromechanical operating status parameters.

[0046] This invention provides an adaptive pitch control method and system based on laser wind radar. It has the following beneficial effects:

[0047] 1. This invention dynamically defines the boundary of the aerodynamic induction zone by introducing an aerodynamic thrust coefficient, adaptively selects the optimal reference range gate and the far-end reference range gate from the radar range gate sequence, and constructs a first-order difference model by combining the spatial span of the two and the synchronous wind speed. This feature enables the wind measurement data extraction process to actively avoid the interference of the near-end blockage effect of the wind turbine, realizes the extrapolation calculation of wind speed from the radar spatial measurement section to the wind turbine plane, and improves the accuracy of incoming wind speed prediction and the reference accuracy of feedforward control.

[0048] 2. This invention generates a torque residual by comparing the actual operating torque of the generator with the model aerodynamic torque, and performs online logical diagnosis of the wind turbine's aerodynamic performance by combining electromechanical operating state parameters. When performance degradation of the blades due to contamination or wear is detected, the system extracts a compensation sensitivity coefficient and uses a bilinear interpolation algorithm to update the two-dimensional interpolation table of aerodynamic thrust coefficients online. This mechanism enables the pitch control algorithm to adaptively track the gradual change in blade aerodynamic characteristics, avoiding long-term operational accuracy degradation caused by fixed control parameters.

[0049] 3. This invention designs an enabling logic based on the effective wind speed variance and an interlocking mechanism based on the bias correlation coefficient, and dynamically generates a feedforward gain coefficient by combining real-time air density and thrust correction factor. When the correlation between radar measurement data and nacelle-estimated wind speed decreases or when abnormal wind conditions occur, the state machine is frozen in time and the output of the additional pitch angle is limited by an interlocking flag. This prevents erroneous pitch changes caused by wind speed measurement distortion or spatial alignment deviation, reduces the mechanical fatigue load on the wind turbine drivetrain, and ensures the stability of the unit's operation. Attached Figure Description

[0050] Figure 1 This is a system architecture diagram of the present invention;

[0051] Figure 2 This is a flowchart of the method of the present invention;

[0052] Figure 3 This is a flowchart of the extrapolation calculation of the equivalent incoming air velocity using the first-order difference model of the present invention.

[0053] Figure 4 This is a flowchart of the aerodynamic attenuation correction and wind turbine plane effective wind speed reconstruction of the present invention;

[0054] Figure 5 This is a flowchart of the multi-condition state machine diagnosis and online thrust coefficient update of the present invention;

[0055] Figure 6 This is a flowchart of the aerodynamic inverse model solution and the fusion of feedforward and feedback total pitch command of the present invention;

[0056] Figure 7The following is a time-series curve diagram of the feedforward compensation calculation process under the sudden wind speed condition of the present invention, wherein (a) is a radar measured wind speed trajectory diagram, and (b) is a comparison diagram of feedback pitch command and total pitch command.

[0057] Figure 8 The graphs are line graphs comparing the electromechanical operating parameters under extreme gust conditions of the present invention. (a) is a comparison graph of generator speed, (b) is a comparison graph of actual pitch angle trajectory, and (c) is a comparison graph of longitudinal bending moment at the tower base. Detailed Implementation

[0058] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0059] See attached document Figure 1 The present invention provides an adaptive pitch control system based on laser wind radar, which may include: a power generation execution module, a radar measurement module, a status acquisition module, a main control calculation module, and a pitch drive module.

[0060] The power generation module comprises an impeller, mechanical transmission chain, generator, converter, and nacelle structure. The radar measurement module is rigidly mounted on the top of the nacelle of the power generation module. The radar measurement module synchronously outputs wind speed data and optical carrier-to-noise ratio data from multiple spatial depth sections according to a set time sampling period and distance threshold.

[0061] The status acquisition modules are distributed and installed on the mechanical transmission chain and electrical circuits of the generator execution module. The status acquisition modules collect actual pitch angle, generator speed, generator torque, grid voltage RMS value, grid connection status, converter status, and nacelle ambient temperature, air pressure, and humidity parameters.

[0062] The main control module establishes communication connections with the radar measurement module, the status acquisition module, and the pitch drive module. The main control module receives output data from each front-end device and performs control logic calculations. The pitch drive module receives pitch control commands from the main control module and drives the blades to perform corresponding pitch angle adjustments.

[0063] See attached document Figure 2 This invention provides an adaptive pitch control method based on laser wind radar, comprising the following steps:

[0064] S100, the main control computing module collects the electromechanical operating status parameters output by the status acquisition module, calls the internally stored two-dimensional interpolation table of aerodynamic thrust coefficient, and calculates the initial thrust coefficient of the power generation execution module under the current operating cycle.

[0065] S200: The main control computing module estimates the forward extension depth of the aerodynamic induced zone in front of the impeller based on the initial thrust coefficient, and performs a traversal and screening in the range gate sequence output by the radar measurement module to determine the optimal reference range gate that meets the anti-blocking boundary conditions.

[0066] S300: The main control and computing module extracts the synchronous wind speed measured by the radar measurement module at the optimal reference distance gate and the far-end fixed section, constructs a first-order difference model to calculate the wind flow spatial distribution gradient, and extrapolates the estimated value of the undisturbed incoming wind speed based on the outer section data.

[0067] S400, the main control computing module calculates the axial induction factor corresponding to the initial thrust coefficient based on momentum theory, and uses the axial induction factor to perform aerodynamic attenuation correction on the estimated value of the undisturbed incoming wind speed, and calculates the effective wind speed actually acting on the wind turbine plane.

[0068] The S500 main control computing module performs variance dead zone limitation on the effective wind speed sequence, performs time offset correlation calculation on the controlled input wind speed sequence and the estimated wind speed sequence of the nacelle, and performs aerodynamic performance degradation diagnosis in combination with radar data quality, grid connection status, power grid status and converter status; when the degradation diagnosis conditions are met, the thrust coefficient compensation amount is calculated based on the generator torque residual, and the aerodynamic thrust coefficient compensation table is updated online.

[0069] The S600 main control computing module schedules the basic feedforward gain based on the effective wind speed, and corrects and interlocks the basic feedforward gain based on the bias correlation coefficient; it inputs the effective wind speed into the locally linearized aerodynamic inverse model, solves the feedforward additional pitch angle, and merges it with the proportional-integral feedback pitch command to generate the final pitch adjustment command.

[0070] The various technical solutions provided by the present invention will be described in detail below with reference to the specific accompanying drawings and control logic.

[0071] Within the set control cycle, the main control computing module acquires the real-time operating status parameters of each underlying subsystem of the wind turbine generator set, and establishes the basic thrust parameters based on these parameters. This process includes the following sub-steps:

[0072] S101, the main control computing module reads multi-dimensional operating status parameters distributed by the status acquisition module through the industrial fieldbus. Within one bus communication cycle, the main control computing module reads the actual pitch angle feedback value, generator speed feedback value, generator torque feedback value, grid voltage RMS value, and converter status word from the pitch drive module, generator encoder, and converter grid-side controller, respectively.

[0073] S102, the main control computing module performs timing synchronization and data alignment processing on the read multi-dimensional operating status parameters. The main control computing module allocates multiple sets of circular buffers in its internal memory, and continuously pushes the received status parameters with hardware clock timestamps into the corresponding circular buffers.

[0074] The main control module uses the task trigger pulse of the current control cycle as the reference time point, traverses the data timestamps in each ring buffer, extracts the two adjacent historical data points with the smallest time difference from the reference time point, and performs a linear interpolation algorithm according to the time deviation ratio to calculate the state parameter value corresponding to the reference time point.

[0075] The main control module obtains the electromechanical operating status parameters synchronized in the current control cycle through timing synchronization and data alignment processing. When the time deviation of a certain sensor data exceeds the maximum allowable time difference threshold, the main control module continues to use the valid data from the previous cycle, but the number of consecutive cycles cannot exceed the preset maximum number; if it exceeds this limit, the corresponding data will be marked as invalid, and feedforward control and online updates of the thrust coefficient will be prohibited.

[0076] S103, the main control calculation module uses the actual pitch angle feedback value after alignment and the current tip speed ratio to call the two-dimensional interpolation table of aerodynamic thrust coefficient to calculate the initial thrust coefficient of the current control cycle.

[0077] The current tip speed ratio is obtained by multiplying the rotor angular velocity (calculated by converting the generator speed through the gearbox transmission ratio) and the rotor radius, and then dividing by the equivalent incoming wind speed of the previous control cycle. For direct-drive wind turbine generators, the gearbox transmission ratio is set to 1. Before calculation, the equivalent incoming wind speed of the previous control cycle is subject to a lower limit. When it is lower than the preset minimum cut-in wind speed threshold, the minimum cut-in wind speed threshold is used to avoid the denominator being zero or too small.

[0078] The two-dimensional interpolation table for aerodynamic thrust coefficient includes a read-only base table and an online compensation table with an initial value of zero. Both tables are indexed by the actual pitch angle and the current tip speed ratio calculated based on the equivalent incoming wind speed. The base table is generated by offline aerodynamic simulation, and the current initial thrust coefficient is the sum of the interpolation results from the two tables, limited to a preset physical range. The main control calculation module uses the aligned actual pitch angle feedback value as the horizontal index coordinate and the current tip speed ratio as the vertical index coordinate, performing bilinear interpolation calculations within this numerical matrix.

[0079] During bilinear interpolation, if the input horizontal or vertical index coordinates exceed the boundary range set by the numerical matrix, the main control module takes the closest boundary node value as the final interpolation result. The target matrix node values ​​extracted by the main control module through bilinear interpolation are used as the initial thrust coefficient of the wind turbine generator under the current operating state.

[0080] After obtaining the aligned actual pitch angle feedback value and the current tip speed ratio, the main control computing module extracts the initial thrust coefficient from the two-dimensional interpolation table of aerodynamic thrust coefficients using its internal bilinear interpolation logic. This extraction process includes the following sub-steps:

[0081] S104, the main control calculation module determines the valid boundary nodes in the two-dimensional interpolation table of aerodynamic thrust coefficients. The two-dimensional interpolation table of aerodynamic thrust coefficients has preset horizontal and vertical axes. The horizontal axis consists of a monotonically increasing sequence of discrete blade pitch angle reference values, and the vertical axis consists of a monotonically increasing sequence of discrete blade tip speed ratio reference values.

[0082] The main control calculation module performs boundary limiting judgment on the actual pitch angle feedback value and the current tip speed ratio. When the actual pitch angle feedback value is less than the minimum reference value or greater than the maximum reference value on the horizontal axis, the main control calculation module sets it to the corresponding minimum or maximum reference value.

[0083] The current blade tip speed ratio limiting logic is the same as the actual pitch angle feedback value. After the limiting process is completed, the main control calculation module compares the limited actual pitch angle feedback value with the horizontal axis sequence to determine the two adjacent pitch angle reference values ​​surrounding the actual pitch angle feedback value, which are divided into the lower limit pitch angle reference value and the upper limit pitch angle reference value.

[0084] Similarly, the main control calculation module compares the current tip speed ratio after limiting with the vertical axis sequence to determine the two adjacent tip speed ratio reference values ​​surrounding the current tip speed ratio, which are divided into the lower limit tip speed ratio reference value and the upper limit tip speed ratio reference value.

[0085] The main control computing module locates four adjacent known matrix nodes formed by the intersection of the above-mentioned reference values ​​in the two-dimensional interpolation table of aerodynamic thrust coefficient. These four nodes constitute a rectangular data grid for local interpolation.

[0086] S105, the main control calculation module calculates the distance weight parameters of the actual input coordinates relative to the boundary of the rectangular data grid. The main control calculation module calculates the horizontal distance weight and the vertical distance weight in the manner described above;

[0087] When the upper and lower limit reference values ​​in any direction coincide, the distance weight in that direction is set to 0, and one-dimensional interpolation is performed along the other direction. The horizontal and vertical distance weights are used to characterize the positional offset ratio of the actual input parameters within the rectangular data grid.

[0088] S106, the main control calculation module performs a bilinear weighted summation operation on the preset thrust coefficient values ​​corresponding to the four known matrix nodes based on the horizontal distance weight and the vertical distance weight.

[0089] The main control computing module extracts the first thrust coefficient of the node corresponding to the lower limit pitch angle and the lower limit tip speed ratio, the second thrust coefficient of the node corresponding to the upper limit pitch angle and the lower limit tip speed ratio, the third thrust coefficient of the node corresponding to the lower limit pitch angle and the upper limit tip speed ratio, and the fourth thrust coefficient of the node corresponding to the upper limit pitch angle and the upper limit tip speed ratio.

[0090] The main control calculation module applies lateral and longitudinal distance weights to perform bilinear weighted calculations on the first, second, third, and fourth thrust coefficients. The summation result is the initial thrust coefficient. The main control calculation module stores the initial thrust coefficient in its internal cache, providing it as a reference input parameter for subsequent aerodynamic induction zone forward extension depth estimation and induction factor calculation logic.

[0091] After calculating the initial thrust coefficient for the current control cycle, the main control computation module constructs a forward boundary mapping function for the aerodynamic induction zone, which is used to quantify the spatial distribution limit of the disturbed flow field in front of the impeller. This process includes the following sub-steps:

[0092] S201, the main control calculation module extracts the initial thrust coefficient and the rotor diameter of the wind turbine generator for the current control cycle. The initial thrust coefficient is used to characterize the degree of axial obstruction of the wind turbine to the airflow under the current operating conditions.

[0093] This axial hindrance causes the airflow in front of the rotor to decelerate and expand in the space before reaching the blade sweep surface. The flow field region where deceleration and expansion occur is defined as the aerodynamic induction zone. The spatial boundary of the aerodynamic induction zone extending along the rotor's central axis towards the windward direction varies with the initial thrust coefficient.

[0094] S202, the main control calculation module, based on the initial thrust coefficient and impeller diameter, calls the preset forward boundary mapping function of the aerodynamic induction zone to calculate the estimated forward extension depth of the aerodynamic induction zone. The forward boundary mapping function of the aerodynamic induction zone is a quadratic polynomial mathematical model with the initial thrust coefficient as the independent variable. The specific calculation formula is as follows:

[0095] ;

[0096] In the formula, This represents the estimated forward extension depth of the aerodynamic induction zone. Indicates the rotor diameter of the wind turbine generator set; This represents the initial thrust coefficient for the current control cycle; Indicates the coefficient of the constant term; Denotes the coefficient of the linear term; This represents the coefficient of the quadratic term.

[0097] The coefficients of the constant term, the first term, and the second term mentioned above are preset fixed constants. The determination method is as follows: obtain offline flow field simulation data of the same type of wind turbine generator under multiple discrete wind speed conditions, and extract the spatial longitudinal coordinates of the position where the wind speed attenuates to a preset proportion of the undisturbed incoming wind speed on the central axis of the flow field. The value range of the preset proportion is set to 95% to 99%.

[0098] After obtaining the above-mentioned spatial longitudinal coordinates, the least squares method is applied to perform nonlinear curve fitting between the spatial longitudinal coordinates and the corresponding initial thrust coefficients to obtain the coefficients of each polynomial and store them inside the main control calculation module.

[0099] S203, the main control computing module performs range limiting determination on the forward extension depth estimate and outputs the final forward boundary distance of the aerodynamic induction zone. The main control computing module reads the minimum and maximum hardware detection ranges of the radar measurement module.

[0100] When the calculated forward extension depth estimate is less than the minimum hardware detection range, the main control module sets it to the minimum hardware detection range; when the forward extension depth estimate is greater than the maximum hardware detection range minus the preset safety margin, the radar feedforward effective flag is set to invalid, and the maximum hardware detection range is no longer used as the anti-blocking range gate.

[0101] Only when the estimated forward depth is within the effective detection range, the main control calculation module outputs the forward boundary distance of the aerodynamic induction zone. This forward boundary distance of the aerodynamic induction zone serves as the spatial starting point for selecting the anti-blocking section from multiple radar range gates in subsequent control steps.

[0102] After obtaining the final forward boundary distance of the aerodynamic induction zone, the main control computing module executes filtering logic in the discrete range gate array output by the radar measurement module to determine the optimal reference range gate for subsequent calls. This process includes the following sub-steps:

[0103] S204, the main control computing module reads the discrete distance gate spatial sequence and simultaneously obtains the carrier-to-noise ratio, device valid code and timestamp valid code of each distance gate to form the distance gate data quality status.

[0104] S205, the main control calculation module performs anti-blocking condition determination on the distance gate sequence. The main control calculation module includes distance gates whose spatial distance value is greater than or equal to the sum of the forward boundary distance of the aerodynamic induction zone and the preset safety margin into the candidate distance gate set.

[0105] S206, the main control processing module determines the optimal reference distance gate from the candidate distance gate set. The main control processing module traverses the candidate distance gate set and removes invalid distance gates whose real-time data quality status word is lower than the preset confidence threshold;

[0106] The main control computing module removes distance gates with a carrier-to-noise ratio below the threshold, abnormal device valid codes, or invalid timestamps, and selects the distance gate with the smallest distance difference from the forward boundary of the aerodynamic induction zone as the optimal reference distance gate from the remaining distance gates.

[0107] The main control module then selects a preset distance gate as the far-end reference distance gate from the effective distance gates whose spatial distance is greater than the optimal reference distance gate and whose distance between the two is not less than the minimum distance span threshold; if the preset distance gate is invalid, the farthest effective distance gate that meets the conditions is selected.

[0108] S207, the main control computing module extracts the radar-measured wind speed corresponding to the optimal reference distance gate and stores it in the internal cache. When performing anti-blocking condition judgment and data quality verification, if there are insufficient candidate distance gates, they are only allowed to continue to be used if the optimal reference distance gate and the far-end reference distance gate of the previous cycle still meet the current anti-blocking boundary, data quality and data timeliness requirements, and the continuous use shall not exceed the preset maximum number of cycles.

[0109] If the above conditions are not met, the radar feedforward valid flag is set to invalid, the online thrust coefficient update is frozen, and the feedforward additional pitch angle is reduced to 0° according to the rate limit. The main control calculation module obtains the measured wind speed through the above filtering logic to provide a data reference for subsequent control steps.

[0110] After determining the optimal reference distance gate, the main control computing module performs the extraction and matching operation of multi-section measurement wind speed data, mapping the discrete wind speed data collected from different spatial locations to a unified spatiotemporal reference. This process includes the following sub-steps:

[0111] S301, the main control computing module converts the radar radial wind speed into the wind turbine axial wind speed based on the radar beam direction, installation angle and nacelle yaw angle;

[0112] Extract the first spatial distance and first measured wind speed of the optimal reference distance gate, and the second spatial distance and second measured wind speed of the far reference distance gate, ensuring that the second spatial distance is greater than the first spatial distance. Initial cross-sectional data extraction and basic spatial parameter verification are complete.

[0113] S302, based on the Taylor frozen turbulence assumption, the main control calculation module performs a lower limit boundary determination on the extracted measured wind speed. When the measured wind speed is lower than the set minimum cut-in wind speed, it is set as the minimum cut-in wind speed.

[0114] The main control and calculation module divides the spatial distance of each distance gate by the corresponding limit and then measures the wind speed to obtain the estimated transport time from each distance gate to the wind turbine plane.

[0115] S303, the main control computing module uses the current control time plus the total response delay of communication, calculation and pitch actuator as the target arrival time, and limits the response delay to no more than the smaller of the two distance gate's expected transport times.

[0116] For each range gate, the target measurement timestamp is obtained by subtracting the corresponding estimated transport time from the target arrival time, and addressing is performed in the radar wind speed history cache queue. The history cache duration is not less than the sum of the maximum estimated transport time and the correlation evaluation window. The main control computing module locks the adjacent historical boundary data points surrounding the mapped timestamp, and the standard time-series envelope positioning logic is completed.

[0117] S304, the main control computing module performs linear interpolation between adjacent historical boundary data points on both sides of the target based on the target measurement timestamp, to obtain the first and second matching wind speeds under the same target arrival time.

[0118] When the target measurement timestamp exceeds the buffer range, or the time interval between adjacent data exceeds the maximum allowable interval, the radar spatiotemporal matching result for this cycle is marked as invalid, and the buffer boundary value is not used as a substitute. The aligned multi-section matched wind speed provides a delay-free data benchmark for the main control computing module to subsequently evaluate the spatial evolution gradient of wind speed.

[0119] See attached document Figure 3 After obtaining the aligned multi-section matched wind speeds, the main control calculation module constructs a first-order difference model of wind speed spatial evolution and extrapolates the equivalent incoming wind speed reaching the wind turbine rotor plane based on this model. This process includes the following sub-steps:

[0120] S305, the main control calculation module calculates the evolution gradient of the airflow field in axial space. The main control calculation module extracts the first spatial distance, the second spatial distance, and the corresponding first and second matched wind velocities. The main control calculation module calculates the difference between the second matched wind speed and the first matched wind speed as the wind speed difference value;

[0121] The difference between the second spatial distance and the first spatial distance is calculated and taken as the distance difference value. The main control calculation module divides the wind speed difference value by the distance difference value to obtain the spatial evolution gradient of the wind speed.

[0122] To avoid division by zero errors caused by excessively close spatial distances, the main control module checks the distance difference before performing division operations. Before performing gradient calculations, the main control module determines whether the second spatial distance is greater than the first spatial distance and whether the distance difference is not less than the minimum distance span threshold. If these conditions are not met, a new far-end reference distance gate is selected. If the conditions are still not met, the spatial gradient calculation for this cycle is marked as invalid.

[0123] The minimum distance span threshold is determined based on the physical spatial resolution of the adjacent distance gate of the radar measurement module, and its value ranges from 10m to 20m.

[0124] S306, the main control computing module calls the first-order difference model to calculate the equivalent incoming wind speed. The first-order difference model approximates the variation of wind speed along the spatial axis as a linear process. This model establishes a one-dimensional spatial distance coordinate system, setting the rotor sweep plane of the wind turbine generator as the origin.

[0125] Within the extrapolation interval, the main control calculation module constructs a spatial linear equation based on the first spatial distance, the second spatial distance, and the matching wind speed between the two. Using the gradient of wind speed spatial evolution as the slope of this linear equation, and solving for the intercept of the linear equation at the origin, linear extrapolation to the impeller plane is achieved. The specific calculation formula is as follows:

[0126] ;

[0127] In the formula, This represents the equivalent incoming air velocity calculated to the impeller plane, in m / s. This indicates the first matching wind speed, in m / s; This indicates the second matching wind speed, in m / s; This represents the first spatial distance, in meters (m). This represents the second spatial distance, in meters (m).

[0128] S307, the main control computing module performs range verification and amplitude limiting on the calculated equivalent incoming wind speed. To prevent out-of-bounds data, the main control computing module reads the preset physical minimum operating wind speed and maximum wind speed threshold from the wind turbine generator controller.

[0129] The aforementioned physical minimum operating wind speed is determined based on the cut-in wind speed set for the wind turbine generator set. The upper limit of the equivalent incoming wind speed is determined based on the smaller of the radar effective measurement upper limit and the maximum operating wind speed of the generator set control system. When the equivalent incoming wind speed is less than the physical minimum operating wind speed, the main control calculation module sets it to the physical minimum operating wind speed.

[0130] When the equivalent incoming wind speed exceeds the maximum wind speed threshold, the main control calculation module sets it to the maximum wind speed threshold. The main control calculation module stores the equivalent incoming wind speed after amplitude limiting verification into an internal buffer queue to replace the directly measured wind speed affected by aerodynamic blocking effects, providing data support reflecting the free flow field wind energy state for the subsequent calculation of pitch feedforward control commands.

[0131] After obtaining the initial thrust coefficient for the current control cycle of the wind turbine generator, the main control calculation module solves for the axial induction factor of the rotor plane based on one-dimensional air momentum theory. This process includes the following sub-steps:

[0132] S401, the main control calculation module reads the initial thrust coefficient of the current control cycle and performs calculation domain boundary determination and amplitude limit verification. According to aerodynamic principles, when the wind turbine thrust coefficient is close to or greater than 1, the wake velocity drops to zero or a negative value, the basic assumption of one-dimensional momentum theory fails, and the subsequent square root calculation produces a non-real number result.

[0133] The main control calculation module presets an upper limit threshold for the thrust coefficient, with a value range of 0.85 to 0.99. When the initial thrust coefficient is greater than the upper limit threshold, the main control calculation module forcibly sets the thrust coefficient used in the calculation to this upper limit threshold; when the initial thrust coefficient is less than 0, the main control calculation module sets it to 0. The main control calculation module verifies the thrust coefficient through the above-mentioned limiting logic output to avoid abnormal interruptions in the controller's underlying algorithm.

[0134] S402, the main control calculation module calls the axial induced factor solution model to calculate the estimated value of the axial induced factor. The main control calculation module performs the following arithmetic logic operations:

[0135] The internal difference is obtained by subtracting the verification thrust coefficient from the numerical value 1. The arithmetic square root of this internal difference is then calculated, and the numerator is obtained by subtracting the arithmetic square root from the numerical value 1. Finally, the numerator is divided by the numerical value 2 to directly calculate the predicted axial induction factor under the current control cycle. The above calculation process is equivalent to finding the roots of the fundamental equation of one-dimensional momentum theory in the real number domain.

[0136] S403, the main control computing module performs aerodynamic boundary limiting on the axial induction factor prediction and outputs the final axial induction factor. When the axial induction factor exceeds a certain critical value, the flow field behind the wind turbine enters a turbulent wake state.

[0137] The main control calculation module presets a physical upper limit threshold for the induction factor. This threshold is determined based on the Grouert empirical correction criterion for wind turbine aerodynamics, and its value range is set to 0.33 to 0.40. When the calculated axial induction factor estimate is greater than the physical upper limit threshold, the main control calculation module sets it to the physical upper limit threshold.

[0138] When the estimated axial induction factor is less than or equal to the physical upper limit threshold of the induction factor, the original value is maintained and output. The main control module stores the final axial induction factor with limited output into its internal memory. This axial induction factor is used to quantify the actual aerodynamic drag ratio of the wind turbine to the free flow, providing a parameter benchmark for the main control module to subsequently reconstruct the effective wind speed in the wind turbine plane.

[0139] See attached document Figure 4 After obtaining the equivalent incoming wind speed and axial induction factor, the main control calculation module performs an aerodynamic attenuation correction operation to reconstruct the effective wind speed actually felt by the wind turbine rotor plane. This process includes the following sub-steps:

[0140] S404, the main control calculation module reads the equivalent incoming wind speed and the axial induction factor output from the previous step to establish an aerodynamic attenuation correction model. When the free flow passes through the swept plane of the wind turbine rotor, the wind speed undergoes aerodynamic attenuation, and the axial induction factor quantifies the proportion of this velocity attenuation. The main control calculation module uses the axial induction factor to construct a velocity attenuation coefficient, which is equal to the value 1 minus the axial induction factor.

[0141] In S405, the main control arithmetic module internally executes multiplication arithmetic logic, multiplying the equivalent incoming wind speed by the aforementioned velocity attenuation coefficient to directly calculate the effective wind speed at the rotor plane. This arithmetic operation transforms the physical attenuation effect into discrete numerical processing steps for the controller, generating an initial wind speed reference for subsequent control.

[0142] S406, the main control module performs dynamic filtering and smoothing on the calculated effective wind speed in the rotor plane. Due to the influence of radar measurement mechanisms and spatiotemporal mapping calculations, the calculated effective wind speed in the rotor plane contains high-frequency components. If these are directly input into subsequent control circuits, it will increase the frequency of the pitch actuator's operation.

[0143] The main control computing module is equipped with a first-order low-pass filter to smooth the effective wind speed in the rotor plane during the current control cycle. The main control computing module presets a filtering time constant, which is determined based on the mechanical response inertia time of the wind turbine pitch system, and its value range is set from 0.5s to 2.0s.

[0144] The main control calculation module obtains the current discrete sampling period of the control system, divides the discrete sampling period by the sum of the filter time constant and the discrete sampling period, and calculates the filter weight coefficient. Subsequently, the main control calculation module multiplies the effective wind speed at the wind turbine plane calculated in the current control period by the filter weight coefficient, and adds it to the result of multiplying the smoothed wind speed output in the previous control period by 1 and subtracting the filter weight coefficient, thus completing the discrete difference iterative calculation.

[0145] The main control computing module stores the smoothed effective wind speed of the wind turbine plane, which is calculated iteratively, into the internal variable area, and uses it as the final control input variable output to trigger the subsequent feedforward pitch action calculation logic.

[0146] After obtaining the effective wind speed in the smoothed rotor plane, the main control module performs variance dead zone limiting and abnormal jump amplitude limiting operations based on a rolling data window to suppress measurement noise and high-frequency actions of the pitch mechanism caused by abnormal data. This process includes the following sub-steps:

[0147] S501, the main control computing module establishes a discrete-time rolling data window of the effective wind speed in the smoothed wind turbine plane and calculates the statistical variance of the window data. The main control computing module allocates a fixed-length first-in-first-out circular queue in its internal memory to store the effective wind speed in the smoothed wind turbine plane for multiple consecutive control cycles.

[0148] The main control calculation module presets a window time span, which is set based on the fatigue load cycle of the wind turbine generator's pitch bearing, with a value range of 5s to 15s. The main control calculation module divides the window time span by the system's discrete sampling period and rounds it to determine the total number of data points in the rolling data window. When the window is not full, the variance is calculated based on the actual number of valid data points. Feedforward control is disabled when the number of valid data points is less than the preset minimum.

[0149] In each control cycle, the main control operation module calculates the arithmetic mean of all discrete data points within the rolling data window, calculates the difference between each discrete data point and the arithmetic mean and squares it, sums all the squared terms and divides them by the total number of discrete data points, thus completing the calculation logic of discrete statistical variance.

[0150] S502, the main control arithmetic module performs dead-zone determination on the calculated statistical variance and outputs the feedforward control enable status. The main control arithmetic module reads the preset variance dead-zone threshold, which is determined based on the minimum start-stop torque set by the pitch mechanism, and its value range is set to 0.1 (m / s). 2 Up to 0.5 (m / s) 2 The main control module compares the statistical variance of the current control cycle with the variance dead zone threshold.

[0151] When the statistical variance is less than or equal to the variance dead zone threshold, the main control module sets the feedforward control enable flag to logic 0, forcing the subsequent feedforward pitch additional command output to 0 degrees; when the statistical variance is greater than the variance dead zone threshold, the main control module sets the feedforward control enable flag to logic 1, allowing the system to execute the corresponding feedforward pitch calculation.

[0152] S503, with the feedforward control enable flag set to logic 1, the main control arithmetic module performs abnormal jump limiting on the latest input smoothed wind turbine plane effective wind speed. To eliminate abnormal numerical jumps in radar measurement data, the main control arithmetic module extracts the smoothed wind turbine plane effective wind speed of the current control cycle, subtracts the smoothed wind turbine plane effective wind speed of the previous control cycle, and calculates the wind speed difference between adjacent cycles.

[0153] The main control module reads the preset maximum wind speed change rate threshold and multiplies the threshold by the system discrete sampling period to obtain the maximum wind speed increment allowed in a single control cycle.

[0154] The main control calculation module executes branch limit replacement logic based on the positive and negative signs and the magnitude of the wind speed difference: when the wind speed difference is greater than the maximum wind speed increment threshold, the main control calculation module determines that the data has undergone a positive abnormal jump, and adds the maximum wind speed increment threshold to the effective wind speed of the smoothed wind turbine plane of the previous control cycle as the current controlled input wind speed.

[0155] When the wind speed difference is less than the negative of the maximum wind speed increment threshold, the main control calculation module determines that the data has undergone a negative abnormal jump and subtracts the maximum wind speed increment threshold from the effective wind speed of the smoothed wind turbine plane in the previous control cycle as the current controlled input wind speed.

[0156] When the absolute value of the wind speed difference is less than or equal to the maximum wind speed increment threshold, the main control calculation module uses the smoothed effective wind speed of the rotor plane in the current control cycle as the current controlled input wind speed. The main control calculation module uses the controlled input wind speed after the abnormal jump limit as the reference control variable and inputs it to the pitch feedforward calculation unit.

[0157] After completing the abnormal jump limiting, the main control computing module performs bias correlation calculation and interlock flag generation to verify the consistency between the radar feedforward wind speed and the actual wind speed sensed by the wind turbine generator. In a specific embodiment of the present invention, this process includes the following sub-steps:

[0158] S504, the main control computing module acquires the wind speed data sequence and performs time offset alignment. The main control computing module reads the nacelle-estimated wind speed output by the wind speed state observer, which estimates the wind turbine plane wind speed based on the generator speed, generator torque, actual pitch angle, and preset aerodynamic model.

[0159] When the controlled input wind speed is written to the buffer, the arrival timestamp of the target determined by S303 is recorded simultaneously. Based on this target arrival timestamp, the main control computing module interpolates and aligns the controlled input wind speed sequence with the estimated wind speed sequence in the cabin, without repeatedly calculating the delay based on the radar detection range.

[0160] The main control computing module extracts the estimated wind speed sequence of the cabin within a predetermined time period before the current control cycle, and calculates the arithmetic mean of the sequence as the current average wind speed.

[0161] When the current average wind speed is less than or equal to a preset minimum wind speed threshold, the main control calculation module sets the average wind speed participating in subsequent calculations to this minimum wind speed threshold, which is set to 0.1 m / s. The main control calculation module divides the radar detection distance by the current average wind speed to calculate the time offset.

[0162] The main control module divides the time offset by the system's discrete sampling period and rounds it to obtain the corresponding discrete delay tick number. Based on this discrete delay tick number, the main control module performs an index-shift extraction of the controlled input wind speed sequence stored in the internal buffer queue, aligning it with the nacelle-estimated wind speed sequence on the time axis.

[0163] S505, the main control computing module calculates the bias correlation coefficient. Within a preset evaluation time window, the main control computing module extracts the aligned controlled input wind speed sequence and the estimated wind speed sequence of the nacelle. The time span of this evaluation time window is set to 5s to 15s.

[0164] The main control computing module calculates the arithmetic mean of the controlled input wind speed sequence and the estimated wind speed sequence of the cabin, respectively; it calculates the sum of the products of the differences between the data points at corresponding time points in the two sequences and their respective arithmetic means, and uses this as the numerator of the covariance;

[0165] Calculate the sum of squares of the differences between the data points in each of the two sequences and their respective arithmetic means. Multiply the sums of squares and calculate the square root, which is used as the denominator of the standard deviation. When the denominator of the standard deviation equals 0, the main control module assigns the bias correlation coefficient a value of 0.

[0166] When the denominator of the standard deviation is not zero, the main control calculation module divides the numerator of the covariance by the denominator of the standard deviation to obtain the bias correlation coefficient. This arithmetic operation process transforms the mathematical statistics model into a direct data processing step for the controller.

[0167] S506, the main control module generates an interlock flag and executes control state intervention. Due to changes in the external wind field, the flow field detected by radar deviates spatially from the actual flow field reaching the wind turbine plane. The main control module sets a correlation locking threshold and a correlation recovery threshold, with the correlation recovery threshold being greater than the correlation locking threshold.

[0168] The main control module compares the calculated bias correlation coefficient with the correlation threshold. When the bias correlation coefficient is not greater than the correlation locking threshold for a consecutive preset period, the interlock flag is set to logic 1 and the correlation correction coefficient is set to 0; when the bias correlation coefficient is not less than the correlation recovery threshold for a consecutive preset period, the interlock flag is set to logic 0.

[0169] When the bias correlation coefficient is between two thresholds, the interlocking state of the previous cycle is maintained; after the interlock is released, a correlation correction coefficient between 0 and 1 is generated based on the position ratio of the bias correlation coefficient between the two thresholds.

[0170] See attached document Figure 5 After generating the interlock flag, the main control module performs multi-condition state machine diagnosis and online thrust coefficient update operations based on the real-time operating conditions of the wind turbine generator set, thereby achieving closed-loop iteration of aerodynamic parameters. In a specific embodiment of the present invention, this process includes the following sub-steps:

[0171] The S507 main control computing module collects the underlying operating status words of the wind turbine generator set and performs multi-condition state machine logic diagnosis. The main control computing module extracts the grid connection status feedback signal, generator speed, and actual pitch angle for the current control cycle.

[0172] The main control computing module presets a minimum operating speed threshold and a pitch shutdown protection threshold. The minimum operating speed threshold is determined based on the cut-in speed of the wind turbine generator set, and the value range is set to 6 r / min to 8 r / min.

[0173] The pitch stop protection threshold is determined based on the feathering position, with a range of 80 to 85 degrees. The main control module internally performs logical AND operations:

[0174] The main control module divides the generator speed by the gearbox transmission ratio to obtain the wind turbine speed.

[0175] When the grid connection status is closed, the wind turbine speed is not lower than the minimum operating speed threshold, the actual pitch angle is less than the pitch shutdown protection threshold, the grid voltage is within the allowable range, the converter is fault-free, the radar feedforward is effective and the interlock flag is logic 0, the state machine enters the monitoring state; otherwise, it enters the frozen state; when any of the above conditions are not met, the main control arithmetic module sets the state machine output flag to logic 0.

[0176] In the S508 main control module, when the state machine is in monitoring mode, the current tip speed ratio is calculated based on the rotor angular velocity, rotor radius, and equivalent incoming air velocity. The tip speed ratio represents the physical ratio of the rotor tip linear velocity to the free air velocity, and is a fundamental parameter for solving the aerodynamic thrust of the rotor.

[0177] The main control module reads the generator speed of the current control cycle and divides it by the preset gearbox ratio to obtain the actual wind turbine speed. Since the actual wind turbine speed is usually expressed in r / min, the main control module multiplies it by the constant pi and divides it by 30 to convert it into the wind turbine angular velocity in radians per second.

[0178] Subsequently, the main control calculation module multiplies the rotor angular velocity by the preset rotor radius to calculate the rotor tip linear velocity. Using the current tip speed ratio and actual blade pitch angle as indices, the main control calculation module extracts the aerodynamic torque coefficient from a preset two-dimensional table of aerodynamic torque coefficients, and calculates the model aerodynamic torque based on the equivalent incoming wind speed, real-time air density, and rotor parameters.

[0179] The main control calculation module converts the generator torque to the wind turbine side and combines it with the wind turbine speed derivative after low-pass filtering, the equivalent moment of inertia, and the preset mechanical loss torque to calculate the actual aerodynamic torque; the difference between the actual aerodynamic torque and the model aerodynamic torque is used as the generator torque residual. This arithmetic calculation process eliminates the case where the denominator of the division operation is zero by setting a lower limit for wind speed.

[0180] S509, the main control calculation module performs aerodynamic performance degradation judgment and online updates of the thrust coefficient compensation table based on the generator torque residual. The main control calculation module performs low-pass filtering on the generator torque residual and sets the residual dead zone, degradation judgment threshold, and recovery threshold.

[0181] When the absolute value of the generator torque residual is greater than the degradation judgment threshold for a consecutive preset period, and the residual sign remains consistent, the main control calculation module determines that aerodynamic performance degradation exists; when the absolute value of the generator torque residual is less than the recovery threshold for a consecutive preset period, the aerodynamic performance degradation judgment is lifted. The recovery threshold is less than the degradation judgment threshold.

[0182] When aerodynamic performance degradation is detected, the main control module uses the current tip speed ratio and actual pitch angle as input coordinates to extract the compensation sensitivity coefficient from the torque residual-thrust coefficient compensation sensitivity table obtained in advance through offline aerodynamic simulation. The main control module multiplies the generator torque residual by the compensation sensitivity coefficient to obtain the thrust coefficient compensation amount for the current control cycle.

[0183] The main control computing module determines the four adjacent nodes in the thrust coefficient compensation table that enclose the current tip speed ratio and the actual pitch angle. Based on the bilinear interpolation weights corresponding to the current input coordinates, the thrust coefficient compensation amount is multiplied by the preset learning rate and then accumulated to the four adjacent nodes respectively.

[0184] The main control module applies limits to the node update amount, cumulative node compensation amount, and the sum of the basic thrust coefficient and the compensated thrust coefficient for a single control cycle. The online updated thrust coefficient compensation table and the basic thrust coefficient two-dimensional interpolation table are used together to calculate the initial thrust coefficient for the next control cycle.

[0185] When the state machine is frozen, the radar data is invalid, the interlock flag is logic 1, or no aerodynamic performance degradation is determined, the main control computing module keeps the thrust coefficient compensation table unchanged.

[0186] After completing the online update of the thrust coefficient, the main control computing module performs a dynamic scheduling mapping planning operation for the feedforward gain coefficient to adapt to the nonlinear changes in the aerodynamic sensitivity of the wind turbine generator under different wind speed conditions. In a specific embodiment of the present invention, this process includes the following sub-steps:

[0187] S601, the main control calculation module determines the scheduling input variables and performs smoothing processing. The main control calculation module reads the smoothed effective wind speed of the wind turbine plane output from the previous step and uses it as the basic independent variable for gain scheduling. To suppress high-frequency jumps in the gain coefficient caused by transient wind speed fluctuations, the main control calculation module configures a moving average filter to perform secondary smoothing on this basic independent variable.

[0188] The main control module establishes a fixed-length discrete data buffer queue in its internal memory, with the queue length ranging from 10 to 50. The main control module stores the smoothed effective wind speed of the wind turbine plane in the current control cycle into the buffer queue, calculates the arithmetic mean of all data in the queue, and outputs it as the scheduling reference wind speed.

[0189] During the initialization phase of the control system or when the buffer queue is not full, the main control calculation module obtains the number of actual valid data that has been stored and uses this number as the denominator when calculating the arithmetic mean, so as to avoid zero values ​​from participating in the calculation and causing the mean to be distorted.

[0190] S602, the main control computing module calculates the basic feedforward gain coefficient. The aerodynamic torque sensitivity of wind turbine generators varies across different wind speed ranges. The main control computing module has a pre-built one-dimensional gain mapping table, which uses wind speed as the index node and stores the offline calculated baseline feedforward gain value.

[0191] The main control module reads the wind speed index range from the one-dimensional gain mapping table and determines whether the dispatch benchmark wind speed exceeds this range. If it does, the main control module limits it to the corresponding boundary wind speed value. The main control module uses the limited dispatch benchmark wind speed as input coordinates and employs a linear interpolation algorithm to extract the corresponding basic feedforward gain coefficient from the one-dimensional gain mapping table.

[0192] The main control computing module performs the following arithmetic calculation steps internally: calculate the difference between the scheduling benchmark wind speed after the amplitude limit and the adjacent lower boundary wind speed node, divide it by the difference between the upper boundary wind speed node and the lower boundary wind speed node to obtain the wind speed offset ratio;

[0193] Subtract the reference feedforward gain value corresponding to the lower limit wind speed node from the reference feedforward gain value corresponding to the upper limit wind speed node to obtain the gain difference; multiply the gain difference by the wind speed offset ratio and add it to the reference feedforward gain value corresponding to the lower limit wind speed node to calculate the basic feedforward gain coefficient.

[0194] To prevent division by zero anomalies, the numerical difference between adjacent wind speed nodes in the one-dimensional gain mapping table is set to a positive number. This arithmetic operation transforms the mathematical interpolation model into a direct data processing step for the controller.

[0195] S603, the main control computing module performs comprehensive corrections based on environmental density, thrust coefficient, and offset correlation coefficient. The basic feedforward gain coefficient is calibrated according to the operating conditions related to standard air density, rated thrust coefficient, radar wind speed, and estimated wind speed in the nacelle.

[0196] The main control computing module calculates the real-time air density based on the cabin ambient temperature, air pressure and humidity output by the status acquisition module, and reads the initial thrust coefficient of the current control cycle and the correlation correction coefficient generated in the previous steps.

[0197] To prevent overflow errors during division, the main control module performs a limit check on the denominator parameter before performing the division. When the real-time air density is lower than a preset safe lower limit, the main control module sets it to that safe lower limit, which is set to 0.9 kg / m³. 3 .

[0198] When the initial thrust coefficient is less than the preset thrust safety lower limit, the main control calculation module sets it to this thrust safety lower limit, which is set to 0.1. Subsequently, the main control calculation module extracts the internally preset standard air density and rated thrust coefficient, with the standard air density set to 1.225 kg / m³. 3 The rated thrust coefficient is determined based on the aerodynamic design parameters of the wind turbine generator set, and the value range is set to 0.6 to 0.8.

[0199] The main control computing module divides the standard air density by the real-time air density after the limit to obtain the density correction factor; it divides the rated thrust coefficient by the initial thrust coefficient after the limit to obtain the thrust correction factor.

[0200] The main control module multiplies the basic feedforward gain coefficient by the density correction factor, thrust correction factor, and correlation correction factor in sequence to obtain the dynamic feedforward gain coefficient. The correlation correction factor ranges from 0 to 1.

[0201] The main control module applies upper and lower limits to the dynamic feedforward gain coefficient and limits the rate of change of adjacent control cycles, and stores the limited dynamic feedforward gain coefficient in the internal variable area.

[0202] See attached document Figure 6 After completing the dynamic scheduling mapping planning of the feedforward gain coefficient, the main control computing module performs aerodynamic inverse model solving and multi-command loop linear fusion operations to generate the total pitch command for controlling the pitch actuator. In a specific embodiment of the present invention, this process includes the following sub-steps:

[0203] S604, the main control calculation module performs feedforward calculations based on a locally linearized aerodynamic inverse model to generate a feedforward additional pitch angle. When the wind turbine is running above the rated wind speed, the main control calculation module extracts the controlled input wind speed after the abnormal jump limit and reads the preset rated wind speed.

[0204] The main control module subtracts the rated wind speed from the controlled input wind speed to obtain the feedforward wind speed deviation. When the feedforward wind speed deviation is less than 0, it is set to 0.

[0205] The main control module multiplies the limited feedforward wind speed deviation by the dynamic feedforward gain coefficient to obtain the feedforward additional pitch angle. The feedforward additional pitch angle is used to superimpose the basic absolute pitch angle command output by the proportional-integral feedback controller, and does not include the basic pitch angle under rated operating conditions.

[0206] The S605 main control module, combining control flags and pitch rate constraints, enables and limits the feedforward additional pitch angle to generate an effective feedforward additional pitch angle. The main control module reads the feedforward control enable flag, interlock flag, state machine status, and radar feedforward effective flag.

[0207] When the feedforward control enable flag is logic 1, the interlock flag is logic 0, the state machine is not in a frozen state, and the radar feedforward valid flag is logic 1, the main control calculation module sets the feedforward additional pitch angle output by S604 to the target feedforward additional pitch angle.

[0208] If any of the above conditions are not met, the main control calculation module will set the target feedforward additional pitch angle to 0°.

[0209] Subsequently, the main control module reads the preset maximum pitch rate limit value of the pitch motor, which is set to a range of 8° / s to 10° / s. The main control module multiplies the maximum pitch rate limit value by the system's discrete sampling period to calculate the maximum angle increment per step.

[0210] The main control calculation module calculates the difference between the target feedforward additional pitch angle and the effective feedforward additional pitch angle of the previous control cycle.

[0211] When the difference is greater than the maximum angle increment in a single step, the effective feedforward additional pitch angle of the previous control cycle is added to the maximum angle increment in a single step, and this is used as the effective feedforward additional pitch angle of the current control cycle.

[0212] When the difference is less than the negative single-step maximum angle increment, the effective feedforward additional pitch angle of the previous control cycle is subtracted from the single-step maximum angle increment, and this is taken as the effective feedforward additional pitch angle of the current control cycle.

[0213] When the absolute value of the difference is less than or equal to the maximum angle increment in a single step, the target feedforward additional pitch angle is taken as the effective feedforward additional pitch angle for the current control cycle.

[0214] The S606 main control module fuses feedforward and feedback control commands to generate the overall pitch control command, which is then sent to the pitch drive module. The pitch control system of a wind turbine generator includes a proportional-integral feedback loop and a radar feedforward loop.

[0215] The main control module reads the basic absolute pitch angle command output by the proportional-integral feedback controller based on the generator speed deviation, and adds the basic absolute pitch angle command to the effective feedforward additional pitch angle output by S605 to obtain the blended pitch command.

[0216] The main control module reads the mechanical limit parameters of the pitch mechanism, including the minimum pitch angle and the maximum pitch angle. The minimum pitch angle is set to 0° and the maximum pitch angle is set to 90°.

[0217] The main control module limits the fused pitch command to between the minimum and maximum pitch angles and calculates the difference between the limited fused pitch command and the total pitch command of the previous control cycle.

[0218] The main control and calculation module determines the maximum single-step angle increment of the total command based on the maximum allowable pitch rate of the pitch mechanism and the discrete sampling period of the system, and performs a rate-of-change limit on the fused pitch command according to the maximum single-step angle increment to obtain the total pitch command.

[0219] When the fused pitch command causes angle limiting or rate of change limitation, the main control calculation module feeds back the limiting status to the proportional-integral feedback controller, stops the accumulation of integral terms or performs back calculation on integral terms to prevent integral saturation.

[0220] The main control module converts the overall pitch command into a drive signal via the industrial field communication bus and sends it to the pitch driver of each blade, driving the servo motor to perform pitch angle deflection action, thus completing the closed-loop control of the pitch action of the wind turbine generator.

[0221] Specific application examples:

[0222] The adaptive pitch control method and system based on laser wind radar of the present invention obtains the spatial depth flow field distribution in front of the wind turbine in advance through the radar measurement module, constructs a first-order difference model to deduce the equivalent incoming wind speed, and reconstructs the effective wind speed actually acting on the wind turbine plane by combining momentum theory.

[0223] The main control module dynamically adjusts the feedforward gain coefficient based on the effective wind speed, calculates the additional feedforward pitch angle using a locally linearized aerodynamic inverse model, and finally merges and issues this feedforward command with the proportional-integral feedback command. This process enables the pitch control logic to anticipate and compensate for external wind disturbances, eliminating the pitch lag caused by the integral delay of generator speed in traditional single feedback control. It limits the speed fluctuation of the wind turbine generator to a safe operating range and reduces the fatigue load on the system's mechanical transmission chain and tower.

[0224] See attached document Figure 7 , Figure 7 The horizontal axis is set to the control operation time, and the vertical axis corresponds to the wind speed amplitude and the pitch angle amplitude, respectively. In the legend, the dotted line represents the radar measured wind speed trajectory, the dashed line represents the pitch angle generated by the feedback control, and the solid line represents the total pitch angle after fusion feedforward compensation.

[0225] Set the rotor diameter of a certain type of wind turbine generator The distance is 120m, the system discrete sampling period is 0.1s, and the rated wind speed is 11m / s. Under initial operating conditions, the main control module determines the first spatial distance of the optimal reference distance gate. The distance is 120m, and the second spatial distance from the far end to the door is a reference distance. The distance is 150m. The radar measurement module extracts the first matching wind speed on these two spatial sections. The second matching wind speed is 13.2 m / s. It is 13.5 m / s.

[0226] The main control module calls the first-order difference model to extrapolate and calculate the equivalent incoming air velocity. The calculation formula is: Substituting the above values ​​into the formula, the spatial evolution gradient is calculated to be 0.01 (m / s) / m, from which the equivalent incoming air velocity is deduced to the impeller plane. It is 12.0 m / s.

[0227] Subsequently, the main control calculation module calculates the axial induction factor based on the initial thrust coefficient, performs attenuation correction on the equivalent incoming wind speed, and obtains a feedforward wind speed deviation of 2.2 m / s. The main control calculation module queries the scheduling mapping table to extract the dynamic feedforward gain coefficient of 1.5° / (m / s), and sets the base pitch angle corresponding to the rated wind speed to 0°.

[0228] The main control module calculates the initial feedforward pitch command using the aerodynamic inverse model formula, outputting an initial feedforward pitch command of 3.3°. Combined with the maximum pitch rate limit of 8° / s, the main control module outputs the effective feedforward additional pitch angle in a slope-constrained manner during subsequent consecutive control cycles. (From the attached...) Figure 7 The timing curves show that the total pitch angle was raised in advance before the wind speed step reached the rotor plane, completing the preventive load unloading action.

[0229] See attached document Figure 8 , Figure 8 The horizontal axis represents the simulation running time, and the three vertical axes correspond to the generator speed, actual pitch angle, and longitudinal bending moment at the tower base, respectively. In the legend, the dashed line corresponds to the data reference of the traditional feedback control system, and the solid line corresponds to the operating data of the adaptive pitch control system of this invention.

[0230] In the time-domain simulation of the multibody dynamics of the entire unit, a standard extreme operating gust model was introduced, setting the base incoming wind speed to 12 m / s. The gust increased to 18 m / s within 5 seconds and then returned to the base state. Test data showed that under the traditional feedback control mode, the peak generator speed reached 1850 r / min, exceeding the rated full-load speed by 12%, and the actual pitch angle curve's response ramp-up point lagged behind the wind speed ramp-up point by approximately 1.5 seconds.

[0231] After applying the adaptive pitch control system of this invention, the peak generator speed was suppressed to 1710 r / min, and the speed overshoot was controlled within 3%. Simultaneously, due to the introduction of slope limiting logic in the feedforward control loop, the actual pitch angle curve transitioned smoothly during the acceleration climb, without any mechanical oscillations caused by sudden command changes.

[0232] Longitudinal bending moment data at the tower base was extracted throughout the simulation cycle. The stress amplitude and cycle number were statistically analyzed using a rainflow counting algorithm. The results showed that the equivalent fatigue load at the tower base after applying this invention was 8.5% lower than that of traditional feedback control. Simulation data verified the effectiveness and engineering applicability of this control scheme in suppressing generator overspeed and reducing structural fatigue load.

Claims

1. An adaptive pitch control method based on laser wind radar, characterized in that, include: The initial thrust coefficient is calculated using a two-dimensional interpolation table of aerodynamic thrust coefficients. The optimal reference range gate and the far-end reference range gate are determined in the range gate sequence based on the initial thrust coefficient. Extract the synchronously measured wind speeds on the optimal reference distance gate and the far-end reference distance gate, and construct a first-order difference model to calculate the equivalent incoming wind speed; The equivalent incoming wind speed is corrected by using the axial induction factor corresponding to the initial thrust coefficient, and the effective wind speed in the smoothed impeller plane is calculated. A controlled input wind speed is generated by limiting the effective wind speed of the smooth wind turbine plane, and aerodynamic performance degradation diagnosis and compensation update are performed in combination with electromechanical operating status parameters; The total pitch command is generated based on the effective wind speed of the smooth wind turbine plane and the controlled input wind speed.

2. The adaptive pitch control method based on laser wind radar according to claim 1, characterized in that, The calculation of the initial thrust coefficient using a two-dimensional interpolation table of aerodynamic thrust coefficients includes: Obtain the actual blade pitch angle feedback value and the current tip speed ratio; the current tip speed ratio is calculated based on the generator speed feedback value, gearbox transmission ratio, impeller radius and the equivalent incoming wind speed of the previous control cycle; Amplitude limiting is applied to the actual pitch angle feedback value and the current tip speed ratio, and four adjacent known nodes surrounding the two are determined in the two-dimensional interpolation table of aerodynamic thrust coefficient; Calculate the lateral and longitudinal distance weights of the actual blade pitch angle feedback value and the current tip speed ratio relative to the boundary of the rectangular grid formed by the four nodes; Using the lateral and longitudinal distance weights, the preset thrust coefficient values ​​corresponding to the four nodes are weighted and summed to obtain the initial thrust coefficient.

3. The adaptive pitch control method based on laser wind radar according to claim 1, characterized in that, Determining the optimal reference range gate and the far-end reference range gate in the range gate sequence based on the initial thrust coefficient includes: Based on the initial thrust coefficient and impeller diameter, the estimated forward extension depth of the aerodynamic induced zone is calculated, and the forward boundary distance of the aerodynamic induced zone is output after limiting the amplitude. Read the distance gate sequence and the corresponding data quality status; Distance gates whose spatial distance is not less than the sum of the forward boundary distance of the aerodynamic induction zone and the preset safety margin are included in the candidate set; After removing invalid distance gates from the candidate set that are below a preset confidence threshold, the distance gate with the smallest difference from the forward boundary of the aerodynamic induction zone is selected as the optimal reference distance gate; among the valid distance gates whose spatial distance is greater than the spatial distance corresponding to the optimal reference distance gate and whose spacing is not less than the minimum span threshold, a preset distance gate is selected as the far-end reference distance gate.

4. The adaptive pitch control method based on laser wind radar according to claim 1, characterized in that, Extract the synchronously measured wind speeds on the optimal reference distance gate and the far-end reference distance gate, and construct a first-order difference model to calculate the equivalent incoming wind speed, including: Obtain the first spatial distance and the second spatial distance corresponding to the optimal reference distance gate and the far-end reference distance gate; The radar radial wind speeds on the optimal reference range gate and the far-end reference range gate are converted into axial wind speeds. Adjacent historical data points are locked in the historical cache queue in combination with the transport time, and interpolation is used to obtain the first matching wind speed and the second matching wind speed. The difference between the second spatial distance and the first spatial distance is calculated as the distance difference, and the difference between the second matching wind speed and the first matching wind speed is calculated as the wind speed difference. The wind speed difference is divided by the distance difference to obtain the wind speed evolution gradient. Based on the first spatial distance, the second spatial distance, and the corresponding first and second matching wind speeds, a spatial straight line equation is constructed. The wind speed evolution gradient is used as the slope, and the intercept at the origin of the coordinate system is solved to obtain the equivalent incoming wind speed and limit its amplitude.

5. The adaptive pitch control method based on laser wind radar according to claim 1, characterized in that, The equivalent incoming wind speed is corrected using an axial induction factor corresponding to the initial thrust coefficient, and the effective wind speed in the smooth rotor plane is calculated, including: Perform a limit check on the initial thrust coefficient and output a verified thrust coefficient; Calculate the difference between value 1 and the verification thrust coefficient and take the arithmetic square root. Subtract the arithmetic square root from value 1 to obtain the numerator. Divide the numerator by value 2 to obtain the estimated value. Limit the estimated value and output the axial induction factor. The attenuation coefficient is constructed by subtracting the axial induction factor from the value 1, and the effective wind speed in the wind turbine plane is obtained by multiplying the equivalent incoming wind speed by the attenuation coefficient. The filtering weight coefficients are calculated based on the mechanical response inertia time. Discrete difference iterative calculation is then performed on the effective wind speed of the wind turbine plane using the filtering weight coefficients to output the smoothed effective wind speed of the wind turbine plane.

6. The adaptive pitch control method based on laser wind radar according to claim 1, characterized in that, The effective wind speed of the smoothed wind turbine plane is limited to generate a controlled input wind speed, and aerodynamic performance degradation diagnosis and compensation updates are performed in conjunction with electromechanical operating status parameters, including: Establish a data window for the effective wind speed of the smoothed wind turbine plane and calculate the statistical variance; when the variance is greater than the preset dead zone threshold, generate a feedforward control enable flag, and perform abnormal jump amplitude limiting on the latest input effective wind speed of the smoothed wind turbine plane to generate the controlled input wind speed. The controlled input wind speed sequence, which is composed of the historical cached controlled input wind speed, is interpolated and aligned with the cabin estimated wind speed sequence. The time offset is calculated based on the current average wind speed of the cabin estimated wind speed sequence, and the controlled input wind speed sequence is translated and extracted. Calculate the bias correlation coefficient between the aligned controlled input wind speed sequence and the estimated wind speed sequence in the cabin; when the bias correlation coefficient is not greater than the correlation locking threshold within a continuous preset period, set the interlock flag to logic 1 and set the set correlation correction coefficient to 0.

7. The adaptive pitch control method based on laser wind radar according to claim 6, characterized in that, The process of combining electromechanical operating status parameters to perform aerodynamic performance degradation diagnosis and compensation updates also includes: Extract the electromechanical operating status parameters for logical diagnosis, and combine the interlock flag to determine whether the state machine enters a monitoring or freezing state. When in monitoring mode, the current tip speed ratio is calculated based on the impeller angular velocity, impeller radius and the equivalent incoming wind speed, and the torque residual is calculated by combining the difference between the actual aerodynamic torque and the model aerodynamic torque. When the absolute value of the filtered residual is continuously greater than a preset judgment threshold and the signs are consistent, the aerodynamic performance degradation is determined to exist. When degradation exists, the actual pitch angle feedback value is obtained, and the current tip speed ratio and the actual pitch angle feedback value are used as coordinates to extract the compensation sensitivity coefficient and multiply it with the torque residual to obtain the thrust coefficient compensation amount. The thrust coefficient compensation amount is multiplied by the learning rate and accumulated into the four nodes surrounding the current coordinates in the preset thrust coefficient compensation table according to the bilinear interpolation weight to complete the update.

8. The adaptive pitch control method based on laser wind radar according to claim 7, characterized in that, Generate a total pitch command based on the effective wind speed in the smoothed rotor plane and the controlled input wind speed, including: The effective wind speed of the smooth wind turbine plane is used as the reference wind speed after sliding filtering, and the basic feedforward gain coefficient is extracted from the preset one-dimensional gain table. Calculate the real-time air density, and combine it with the initial thrust coefficient to calculate the density and thrust correction factor; multiply the basic feedforward gain coefficient by the density and thrust correction factor and the correlation correction coefficient in sequence to generate the dynamic gain coefficient; The wind speed deviation is obtained by subtracting the rated wind speed from the controlled input wind speed, and the additional pitch angle is obtained by multiplying the wind speed deviation by the dynamic gain coefficient. When the feedforward control enable flag is logic 1, the interlock flag is logic 0, and the state machine is not frozen, the pitch rate limit is applied to the additional pitch angle to generate a valid feedforward additional pitch angle. The effective feedforward additional pitch angle is added to the acquired basic absolute pitch angle command, and the total pitch command is generated after limiting.

9. An adaptive pitch control system based on laser wind radar, characterized in that, include: The status acquisition module is distributed and installed on the drive chain and electrical circuit of the wind turbine generator set to collect electromechanical operating status parameters. The radar measurement module is rigidly installed on the top of the nacelle of the wind turbine generator set and is used to output wind speed data on multiple spatial depth sections. The pitch drive module is used to drive the blades to perform pitch angle adjustment actions; The main control and computing module is communicatively connected to the above-mentioned modules and is used to execute the method as described in any one of claims 1 to 8 to generate a total pitch command and send it to the pitch drive module.

10. The adaptive pitch control system based on laser wind radar according to claim 9, characterized in that, The main control computing module opens multiple ring buffers in its internal memory and pushes the electromechanical operating status parameters with hardware clock timestamps into them. Using the current task trigger pulse as the reference time point, traverse the timestamps, extract the adjacent historical data points with the smallest time difference, perform linear interpolation, and obtain the electromechanical operating state parameters after synchronization and alignment.

Images