Pitch angle dynamic compensation method based on multi-radar space-time synchronization

Abstract

The invention discloses a pitch angle dynamic compensation method based on multi-radar space-time synchronization, and relates to the technical field of wind power generation. In order to solve the problem that the pitch angle control precision is insufficient due to the fact that three-dimensional wind field space-time characteristics cannot be captured in traditional single-point wind speed measurement, a wind wheel plane three-dimensional space-time speed field is synchronously constructed through a horizontal-vertical radar array, and blade segmented aerodynamic load distribution is calculated in real time in combination with a flow field-structure coupling model; establishing a vorticity transport equation under a rotating coordinate system based on a vorticity-shearing force transfer function, mapping dynamic load deviation through a neural network to generate a reference variable pitch instruction, and fusing the tower shadow interference feature vector to perform pre-compensation; a quadratic programming algorithm is adopted to resist saturation optimization, so that the variable pitch rate strictly follows a dynamic constraint curve of a hydraulic system. According to the method, the problem of aerodynamic load phase delay is effectively solved, periodic load fluctuation caused by the tower shadow effect is reduced, and technical guarantee is provided for stable operation of a large wind turbine generator under the complex wind condition.

Description

Technical Field

[0001] The invention relates to the technical field of wind power generation, and in particular to a pitch angle dynamic compensation method based on multi-radar time-space synchronization. Background Art

[0002] As an important component of clean energy, wind power generation plays an increasingly important role in the transformation of the global energy structure. However, wind turbines face complex and changeable wind conditions, which poses a huge challenge to pitch angle control. Traditional wind turbine control systems mainly rely on single-point wind speed measurement. This method is difficult to accurately capture the spatiotemporal variation characteristics of the wind field under complex wind conditions, resulting in limited accuracy and response speed of pitch angle control.

[0003] At present, the pitch angle control systems of wind turbines generally have the following problems: first, single-point wind speed measurement cannot fully reflect the wind speed distribution on the rotor plane, resulting in uneven force on the blades; second, traditional control algorithms are difficult to cope with rapidly changing wind conditions, causing phase delays in the dynamic response of aerodynamic loads, affecting power generation efficiency and equipment life. Summary of the invention

[0004] The present invention provides a pitch angle dynamic compensation method based on multi-radar time-space synchronization to solve the problems in the background technology.

[0005] The present application provides a method for dynamic compensation of pitch angle based on multi-radar time-space synchronization, and the technical scheme is as follows: through the synchronous measurement data of the horizontal scanning radar array and the vertical profile radar array, a three-dimensional time-space velocity field of the wind rotor plane is constructed to comprehensively capture the wind speed changes around the wind rotor; the velocity field data is input into the flow field-structure coupling model, and the dynamic aerodynamic load distribution of each blade segment is calculated in real time to reflect the force conditions of the blades under different wind conditions; based on the deviation between the aerodynamic load distribution data and the target load, an initial variable pitch compensation instruction set is generated; the initial instruction set is input into the dynamic constraint model of the actuator for anti-saturation optimization, and the final pitch angle control instruction is output and synchronized to the variable pitch actuator.

[0006] Furthermore, the present application also proposes that the process of constructing the three-dimensional space-time velocity field includes: receiving the original point cloud data stream from multiple radar nodes, eliminating the coordinate offset caused by the tower swing through a space-time registration algorithm; mapping the registered point cloud data to a rotating coordinate system to generate a three-dimensional gridded velocity field with the impeller center as the origin; and calculating the spatial distribution of the flow field acceleration field based on the difference in velocity field data between adjacent time steps.

[0007] Furthermore, the present application also proposes that the data processing of the flow field-structure coupling model includes: converting the gridded velocity field into a vorticity tensor field: ;

[0008] In the impeller rotating coordinate system, considering the Coriolis force and centrifugal force effects, the modified momentum equation is:

[0009]

[0010] in, is the impeller angular velocity vector, r is the position vector;

[0011] The vorticity , take the curl and simplify it, and get the vorticity transport equation in the rotating coordinate system:

[0012]

[0013] in, Additional terms introduced for rotation effects;

[0014] Shear force on the blade surface The shear force transfer function is constructed based on the relationship with vorticity. The transfer function form is:

[0015]

[0016] in, is the rotation correction coefficient; the integration domain A covers the blade surface microelement;

[0017] The flow separation effect was characterized by fitting an asymmetric bimodal function using the least squares method:

[0018] Mapping flow field to shear force: 3D velocity field observed by radar → Calculation of vorticity field ; By transferring the function Map the vorticity field to the shear force distribution on the blade surface; integrate the shear force to obtain the aerodynamic bending moment ;

[0019] Real-time bending moment With target value Compare and calculate dynamic load deviation; Sliding window statistical deviation vector .

[0020] Furthermore, the present application also proposes that the generation of the initial pitch compensation instruction set includes: inputting the dynamic load deviation vector into a pre-trained neural network model, mapping and generating a baseline pitch angle increment for each blade; superimposing a tower shadow interference feature vector constructed based on historical operating data to generate an initial pitch instruction containing pre-compensation; and fusing the initial instruction set with the current pitch angle state data to generate a time-continuous pitch instruction sequence.

[0021] Furthermore, the present application also proposes that the generation of the tower shadow interference feature vector includes: obtaining the historical wind direction data stream from the SCADA system, and constructing a time-space distribution map of the tower wake interference pattern; extracting the near-tower velocity profile in the radar data in real time, and performing convolution similarity matching with the distribution map; extracting the wake vortex intensity coefficient and phase delay parameter based on the matching results, and constructing a time-varying interference feature vector.

[0022] Furthermore, the present application also proposes that the data processing of the anti-saturation optimization includes: receiving the real-time status data stream of the pitch actuator, including hydraulic oil pressure, actuator displacement and temperature monitoring values; inputting the status data into the actuator dynamic model to predict the pitch rate limit curve in the future time window; and using a constrained optimization algorithm to re-plan the initial instruction set so that the pitch trajectory falls within the range of the rate limit curve.

[0023] Furthermore, the present application also proposes that the updating of the dynamic model of the actuator includes: injecting a sweep frequency test signal during the idle period of the variable pitch system to collect the frequency response data of the actuator; identifying the damping coefficient and stiffness parameters in the transfer function based on the response data; and updating the weight matrix of the dynamic model when the parameter change exceeds the adaptive threshold.

[0024] Furthermore, the present application also proposes that the method also includes forward-looking flow field compensation: performing intrinsic orthogonal decomposition on the radar time-series velocity field data to extract the spatiotemporal evolution coefficients of the dominant flow modes; inputting the evolution coefficients into a pre-trained LSTM network to predict the flow field evolution trend of multiple rotation cycles in the future; and injecting the predicted flow field data into the flow field-structure coupling model in advance to generate an advance compensation instruction component.

[0025] Furthermore, the present application also proposes that the fusion of the advance compensation instructions includes: calculating the real-time error between the actual flow field data and the predicted data, and constructing a prediction credibility weight factor; using a weighted fusion algorithm to mix the advance compensation instructions with the real-time compensation instructions; when the prediction error continues to increase, gradually reducing the weight ratio of the advance compensation instructions until it is turned off.

[0026] From the above, it can be seen that the present application provides a method for dynamic compensation of pitch angle based on multi-radar time-space synchronization, which constructs a three-dimensional time-space velocity field of the wind rotor plane through the synchronous measurement data of the horizontal scanning radar array and the vertical profile radar array; the velocity field data is input into the flow field-structure coupling model, and the dynamic aerodynamic load distribution of each blade segment is calculated in real time; based on the deviation between the aerodynamic load distribution data and the target load, an initial variable pitch compensation instruction set is generated; the initial instruction set is input into the dynamic constraint model of the actuator for anti-saturation optimization, and the final pitch angle control instruction is output and synchronized to the variable pitch actuator, thereby realizing accurate capture and compensation of complex wind conditions, which has the advantages of improving the pitch angle control accuracy and response speed, improving the uniformity of blade force, effectively coping with rapidly changing wind conditions, reducing the phase delay of the dynamic response of the aerodynamic load, and effectively compensating for the tower shadow effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 This is a flow chart of a pitch angle dynamic compensation method based on multi-radar time-space synchronization of the present invention;

[0028] Figure 2 It is a sensitivity analysis diagram of the transfer function characteristics and the rotation correction coefficient in the present invention. DETAILED DESCRIPTION

[0029] The purpose of this application is to provide a pitch angle dynamic compensation method based on multi-radar time-space synchronization, which has the advantages of improving the pitch angle control accuracy and response speed, realizing accurate capture and compensation of complex wind conditions, improving the uniformity of blade force, effectively responding to rapidly changing wind conditions, reducing the phase delay of the dynamic response of the aerodynamic load, and effectively compensating for the tower shadow effect.

[0030] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention are shown in the drawings, rather than all of them.

[0031] Embodiment 1

[0032] like Figure 1 A method for dynamic compensation of pitch angle based on multi-radar time-space synchronization is shown in the flow chart:

[0033] The present invention belongs to the field of intelligent control of wind turbines, and specifically relates to a method for dynamic compensation of pitch angle based on multi-radar spatiotemporal synchronization. The core invention concept breaks through the limitations of traditional single-point wind speed measurement, constructs a three-dimensional spatiotemporal synchronous observation network through a multi-radar array, establishes a dynamic coupling relationship between the blade rotation coordinate system and the flow field evolution model, and proposes an adaptive compensation algorithm based on vortex-shear force transfer characteristics to solve the phase delay problem of the dynamic response of aerodynamic loads under complex wind conditions.

[0034] The dynamic response of aerodynamic loads on blades of wind turbines under complex wind conditions has always been a technical problem. The traditional single-point wind speed measurement method is difficult to fully capture the wind speed changes around the wind rotor, resulting in the inability to accurately reflect the force conditions of the blades, which in turn affects the accuracy and reliability of variable pitch control. In order to solve this problem, the present invention proposes a dynamic compensation method for pitch angle based on multi-radar spatiotemporal synchronization.

[0035] This method constructs a three-dimensional space-time velocity field of the wind rotor plane through the synchronous measurement data of the horizontal scanning radar array and the vertical profile radar array. These velocity field data are input into the flow field-structure coupling model to calculate the dynamic aerodynamic load distribution of each blade segment in real time. Based on the deviation between the aerodynamic load distribution data and the target load, an initial pitch compensation instruction set is generated. Then, the initial instruction set is input into the dynamic constraint model of the actuator for anti-saturation optimization, and the final pitch angle control instruction is output and synchronized to the pitch actuator.

[0036] By synchronously measuring data through the horizontal scanning radar array and the vertical profile radar array, a three-dimensional space-time velocity field of the wind rotor plane is constructed, which helps to fully capture the wind speed changes around the wind rotor. These velocity field data are input into the flow field-structure coupling model, and the dynamic aerodynamic load distribution of each blade segment is calculated in real time, which can accurately reflect the force conditions of the blades under different wind conditions. Based on the deviation between the aerodynamic load distribution data and the target load, an initial pitch compensation instruction set is generated to make the pitch control more accurate. The initial instruction set is input into the dynamic constraint model of the actuator for anti-saturation optimization to ensure that the pitch control instruction does not exceed the capacity of the actuator during execution. The final pitch angle control instruction is output and synchronized to the pitch actuator to ensure the reliability and effectiveness of the pitch control.

[0037] In practical applications, it is first necessary to synchronously obtain the wind speed data around the wind rotor through the horizontal scanning radar array and the vertical profile radar array. After these data are processed, the three-dimensional space-time velocity field of the wind rotor plane is constructed. Then, these velocity field data are input into the flow field-structure coupling model, which can calculate the dynamic aerodynamic load distribution of each blade segment in real time. Based on these calculation results, they are compared with the target load to generate an initial pitch compensation instruction set. In order to ensure that the pitch control instruction does not exceed the capacity of the actuator during execution, the initial instruction set needs to be input into the dynamic constraint model of the actuator for anti-saturation optimization, and the final output pitch angle control instruction will be synchronized to the pitch actuator.

[0038] Compared with the prior art, the present invention constructs a three-dimensional spatiotemporal synchronous observation network through a multi-radar array, which can fully capture the wind speed changes around the wind rotor and solve the limitations of traditional single-point wind speed measurement. Through the flow field-structure coupling model, the dynamic aerodynamic load distribution of each blade segment can be calculated in real time to ensure the accuracy of pitch control. Anti-saturation optimization is performed through the dynamic constraint model of the actuator to ensure that the pitch control command will not exceed the capacity of the actuator during execution, thereby ensuring the reliability and effectiveness of pitch control.

[0039] By synchronously measuring data through the horizontal scanning radar array and the vertical profile radar array, a three-dimensional space-time velocity field of the wind rotor plane is constructed, which helps to fully capture the wind speed changes around the wind rotor. These velocity field data are input into the flow field-structure coupling model, and the dynamic aerodynamic load distribution of each blade segment is calculated in real time, which can accurately reflect the force conditions of the blades under different wind conditions. Based on the deviation between the aerodynamic load distribution data and the target load, an initial pitch compensation instruction set is generated to make the pitch control more accurate. The initial instruction set is input into the dynamic constraint model of the actuator for anti-saturation optimization to ensure that the pitch control instruction does not exceed the capacity of the actuator during execution. The final pitch angle control instruction is output and synchronized to the pitch actuator to ensure the reliability and effectiveness of the pitch control.

[0040] Furthermore, the present application also proposes to receive the original point cloud data stream from multiple radar nodes, eliminate the coordinate offset caused by the tower swing through a spatiotemporal registration algorithm; map the registered point cloud data to a rotating coordinate system to generate a three-dimensional grid velocity field with the impeller center as the origin; and calculate the spatial distribution of the flow field acceleration field based on the difference in velocity field data between adjacent time steps.

[0041] By receiving the original point cloud data stream from multiple radar nodes and using the spatiotemporal registration algorithm to eliminate the coordinate offset caused by the tower swing, the accuracy of the data can be ensured. Mapping the registered point cloud data to the rotating coordinate system to generate a three-dimensional gridded velocity field with the center of the impeller as the origin helps to better reflect the actual aerodynamic environment of the wind rotor plane. Based on the difference in velocity field data between adjacent time steps, the spatial distribution of the flow field acceleration field is calculated, which can further improve the accurate prediction of the aerodynamic load changes on the wind rotor plane.

[0042] Receive the original point cloud data stream from multiple radar nodes, and use the spatiotemporal registration algorithm to eliminate the coordinate offset caused by the tower swing. The spatiotemporal registration algorithm can be implemented in a variety of ways to achieve the purpose of eliminating coordinate offset, such as using a least squares-based registration algorithm or an iterative closest point (ICP)-based registration method. Mapping the registered point cloud data to the rotating coordinate system can be achieved through rotation matrix transformation. Generating a three-dimensional gridded velocity field with the impeller center as the origin can be achieved through a grid division algorithm. Common grid division methods include uniform grid division and adaptive grid division. Based on the difference in velocity field data between adjacent time steps, the spatial distribution of the flow field acceleration field is calculated, which can be achieved through a numerical difference method, such as using backward difference, forward difference or center difference method.

[0043] Specifically, the synchronous measurement data of the horizontal scanning radar array and the vertical profile radar array are used to construct the three-dimensional spatiotemporal velocity field of the wind rotor plane to solve the problem that traditional single-point wind speed measurement can only obtain local wind speed information and cannot capture the spatiotemporal evolution characteristics of the three-dimensional wind field:

[0044] Heterogeneous radar selection criteria: X-band radar (horizontal scanning): using its high range resolution (0.5m) to capture horizontal wind shear effects;

[0045] Millimeter-wave radar (vertical profile scanning): vertical turbulence components are acquired through wide beam coverage (-30°~+45° elevation angle).

[0046] Dual-mode timing module (Beidou / GPS): ensures that the time synchronization error of multi-radar data is less than 1μs, eliminating the impact of time domain misalignment on three-dimensional flow field reconstruction.

[0047] Dynamic coordinate registration algorithm: 1. Tower shadow effect compensation: Based on the tower swing model, the radar coordinate offset is corrected in real time;

[0048] 2. Extended Kalman Filter (EKF): Integrate IMU attitude data and laser ranging data to establish a rotating coordinate system conversion model to ensure that the point cloud data is accurately mapped to the impeller center coordinate system.

[0049] Construction of three-dimensional gridded wind field: The registered point cloud is mapped to the impeller rotating coordinate system to generate a 0.5m×0.5m×0.5m gridded velocity field; based on the velocity field difference of adjacent time steps, the flow field acceleration distribution is calculated.

[0050] This application eliminates the coordinate offset caused by tower swing through a time-space registration algorithm to ensure data accuracy; maps the registered point cloud data to a rotating coordinate system to generate a three-dimensional grid velocity field with the center of the impeller as the origin, which can better reflect the actual aerodynamic environment of the wind rotor plane; based on the difference in velocity field data between adjacent time steps, the spatial distribution of the flow field acceleration field is calculated to further improve the accurate prediction of the aerodynamic load changes on the wind rotor plane. These technical features have significant advantages over the prior art, and can more accurately solve the coordinate offset problem of multi-radar node data and improve the prediction accuracy of the aerodynamic load changes on the wind rotor plane.

[0051] Furthermore, the present application also proposes to convert the gridded velocity field data into a vorticity tensor field, calculate the pressure gradient distribution on the blade surface through a preset vorticity-shear transfer function; integrate the pressure gradient data along the blade span to generate real-time aerodynamic bending moment data for each blade segment; perform a sliding window comparison on the aerodynamic bending moment data and the target load curve to output a dynamic load deviation vector.

[0052] By converting the gridded velocity field data into a vorticity tensor field, the rotation and shear characteristics of the fluid can be captured, and the pressure distribution on the blade surface can be accurately calculated. The role of this feature is to improve the accuracy of the pressure gradient distribution calculation. The pressure gradient data is integrated along the span of the blade to generate real-time aerodynamic bending moment data, which can realize real-time monitoring and control of the aerodynamic bending moment of each segment of the blade, which is a key step in the precise control of the aerodynamic load of the blade. By comparing the aerodynamic bending moment data with the target load curve through a sliding window, the deviation between the aerodynamic bending moment and the target load can be dynamically monitored, and the dynamic load deviation vector can be output. These technical features work together to achieve accurate calculation and real-time monitoring of the aerodynamic load on the blade, and solve the technical problem of how to use gridded velocity field data to calculate the pressure gradient distribution on the blade surface and generate real-time aerodynamic bending moment data.

[0053] The conversion of gridded velocity field data into vorticity tensor field can be achieved in the following ways: First, the gridded velocity field data can be obtained through numerical simulation or actual measurement. Then, the vorticity calculation formula in fluid mechanics is used to convert the velocity field data into a vorticity tensor field. The preset vorticity-shear transfer function is used to calculate the pressure gradient distribution on the blade surface. The pressure gradient data is integrated along the span of the blade to generate aerodynamic bending moment data, which can be achieved through numerical integration methods, such as trapezoidal integration method or Simpson integration method. The aerodynamic bending moment data is compared with the target load curve in a sliding window. By setting the size and step size of the sliding window, the deviation between the aerodynamic bending moment and the target load can be dynamically calculated, and the dynamic load deviation vector can be output.

[0054] Specifically, the flow field-structure coupling model is used to solve the phase delay problem of aerodynamic loads, and the vorticity-shear force transfer function is constructed:

[0055] Control equations in the rotating coordinate system:

[0056] Convert the gridded velocity field to a vorticity tensor field: ;

[0057] In the impeller rotating coordinate system, considering the Coriolis force and centrifugal force effects, the modified momentum equation is:

[0058]

[0059] in, is the impeller angular velocity vector, r is the position vector;

[0060] The vorticity , take the curl and simplify it, and get the vorticity transport equation in the rotating coordinate system:

[0061]

[0062] in, Additional terms introduced for rotation effects;

[0063] Shear force on the blade surface The shear force transfer function is constructed based on the relationship with vorticity. The transfer function form is:

[0064]

[0065] in, is the rotation correction coefficient; the integration domain A covers the blade surface microelement;

[0066] Further, transfer function fitting: the asymmetric bimodal function is fitted by the least squares method to characterize the flow separation effect:

[0067] Mapping flow field to shear force: 3D velocity field observed by radar → Calculation of vorticity field ; By transferring the function Map the vorticity field to the shear force distribution on the blade surface; integrate the shear force to obtain the aerodynamic bending moment ;

[0068] like Figure 2 The verification of the transfer function characteristics and the sensitivity analysis of the rotation correction coefficient C1 in the present invention are demonstrated.

[0069] (a) The transfer function characteristic curve shows the consistency among the theoretical model, experimental data and least squares fitting results, verifying the accuracy of the calculation of the pressure gradient distribution on the blade surface.

[0070] (b) Rotation correction factor Sensitivity analysis showed that Values ​​in the range of 0.70 to 1.00 have a significant effect on the transfer function. =0.85 can effectively improve the calculation accuracy. The flow field-structure coupling model is used to solve the aerodynamic load phase delay problem, and the vorticity-shear force transfer function is constructed:

[0071] Further, dynamic load deviation calculation: real-time bending moment With target value Compare and calculate dynamic load deviation; Sliding window statistical deviation vector .

[0072] This application proposes to convert the gridded velocity field data into a vorticity tensor field, and use the preset vorticity-shear transfer function to calculate the pressure gradient distribution on the blade surface, further generate real-time aerodynamic bending moment data through integration, and perform a sliding window comparison with the target load curve to output a dynamic load deviation vector. This technical solution solves the problem that it is difficult to accurately calculate the pressure gradient distribution on the blade surface in traditional methods, and improves the accuracy and real-time performance of aerodynamic load calculation. Compared with the prior art, the technical solution of the present application has higher accuracy and dynamic response capabilities, can more effectively monitor and control the aerodynamic load of the blade, and ensure the stable operation of the wind turbine under complex wind conditions.

[0073] Furthermore, the present application also proposes to input the dynamic load deviation vector into a pre-trained neural network model to map and generate a baseline pitch angle increment for each blade; superimpose the tower shadow interference feature vector constructed based on historical operating data to generate an initial pitch instruction with pre-compensation; and fuse the initial instruction set with the current pitch angle state data to generate a time-continuous pitch instruction sequence.

[0074] This application includes all the features of the preamble and the characteristic part. The preamble involves the acquisition of the dynamic load deviation vector, and the characteristic part includes inputting the vector into a pre-trained neural network model to generate a reference pitch angle increment, superimposing the tower shadow interference feature vector to generate an initial pitch instruction, and finally fusing the current pitch angle state data to generate a time-continuous pitch instruction sequence. These technical features work together to solve the problem of dynamic load compensation of blades under complex wind conditions. By inputting the dynamic load deviation vector into the pre-trained neural network model, the reference pitch angle increment of each blade can be accurately mapped and generated. Superimposing the tower shadow interference feature vector further optimizes the initial pitch instruction, making it more targeted and effective. Finally, through fusion with the current pitch angle state data, a time-continuous pitch instruction sequence is generated, thereby realizing dynamic load compensation for blades under complex wind conditions.

[0075] The acquisition of the dynamic load deviation vector can be achieved in many ways, such as by calculating the three-dimensional space-time velocity field data constructed by a multi-radar array combined with the flow field-structure coupling model. The pre-trained neural network model can be trained using deep learning technology, and the reference pitch angle increment is output after the dynamic load deviation vector is input. The generation of the tower shadow interference feature vector can be based on historical operating data and constructed through data mining and pattern recognition technology. The generation of the initial pitch instruction set requires the superposition of the reference pitch angle increment and the tower shadow interference feature vector. Finally, the fusion of the initial instruction set with the current pitch angle state data can be achieved through state estimation and control algorithms to generate a time-continuous pitch instruction sequence.

[0076] This application optimizes the control method of variable pitch angle by introducing a neural network model and historical data analysis, and can more accurately cope with dynamic load changes under complex wind conditions. Compared with traditional control methods, this application has higher accuracy and response speed when dealing with complex wind conditions, and can significantly improve the operating stability and efficiency of wind turbines.

[0077] Furthermore, the present application also proposes to obtain the historical wind direction data stream from the SCADA system to construct a time-space distribution map of the tower wake interference pattern; extract the near-tower velocity profile in the radar data in real time, and perform convolution similarity matching with the distribution map; extract the wake vortex intensity coefficient and phase delay parameter based on the matching results, and construct a time-varying interference feature vector.

[0078] The historical wind direction data stream is obtained from the SCADA system, and the time-space distribution map of the tower wake interference pattern is constructed to identify and predict the tower wake interference. The velocity profile of the near-tower area in the radar data is extracted in real time, and the convolution similarity matching is performed with the distribution map to ensure the matching accuracy. The tail vortex intensity coefficient and phase delay parameters are extracted according to the matching results, and the time-varying interference feature vector is constructed. Through these technical features, the influence of tower wake interference on pitch angle control can be effectively identified and compensated, and the accuracy and reliability of dynamic compensation of pitch angle can be improved.

[0079] Furthermore, the historical wind direction data stream of the SCADA system can be processed through data mining and analysis techniques to construct a more accurate tower wake interference pattern. The convolution similarity matching process can be implemented through an efficient algorithm to ensure real-time performance and accuracy. The extraction of the wake vortex intensity coefficient and the phase delay parameter can be performed through comparative analysis and optimization algorithms to improve the accuracy and reliability of the time-varying interference feature vector. For example, a machine learning-based algorithm can be used to optimize the matching process and parameter extraction process, thereby improving the performance and response speed of the overall system. Further, the wake database is constructed:

[0080] Extract 10 years of historical wind direction data from the SCADA system to construct a tower wake interference pattern map;

[0081] The convolutional neural network (CNN) matches the real-time radar data and outputs the tail vortex intensity coefficient and phase delay.

[0082] Add the tower shadow interference pre-compensation item to the pitch command:

[0083] This application constructs a time-space distribution map of the tower wake interference pattern, combines the convolution similarity matching of real-time radar data, extracts the wake vortex intensity coefficient and phase delay parameters, and can effectively solve the impact of tower wake interference on the dynamic compensation of pitch angle. Compared with the prior art, the method of this application can provide higher matching accuracy and real-time performance, improve the accuracy and reliability of dynamic compensation of pitch angle, and thus better adapt to complex wind conditions.

[0084] Furthermore, the present application also proposes receiving the real-time status data stream of the pitch actuator, including hydraulic oil pressure, actuator displacement and temperature monitoring values; inputting the status data into the actuator dynamic model to predict the pitch rate limit curve within the future time window; and using a constrained optimization algorithm to re-plan the initial instruction set so that the pitch trajectory falls within the rate limit curve.

[0085] By receiving the real-time status data stream of the pitch actuator, including hydraulic oil pressure, actuator displacement and temperature monitoring values, the working status of the actuator can be understood in real time; by inputting these status data into the actuator dynamic model, the pitch rate limit curve in the future time window can be predicted, thereby predicting the working capacity of the actuator in advance; the initial instruction set is re-planned using a constrained optimization algorithm to ensure that the pitch trajectory falls within the rate limit curve range and avoid actuator saturation problems.

[0086] In practical applications, receiving the real-time status data stream of the pitch actuator can be achieved through sensors installed on the actuator, which can monitor key parameters such as hydraulic oil pressure, actuator displacement and temperature in real time. These data will be transmitted to the control system in real time. Inputting these status data into the dynamic model of the actuator can be achieved using a physical model-based or data-driven modeling approach. Physical models usually involve detailed modeling of the mechanical and hydraulic characteristics of the actuator, while data-driven methods can use machine learning techniques to make predictions based on historical data. The initial instruction set is replanned using a constrained optimization algorithm. Specifically, a linear or nonlinear programming method can be used, combined with a pitch rate limit curve, to adjust the initial instruction set to ensure that the pitch trajectory does not exceed the capabilities of the actuator in the future time window.

[0087] This application can effectively avoid the actuator saturation problem by real-time monitoring of the state data of the pitch actuator and combining dynamic models and optimization algorithms. Compared with the existing technology, this method can more accurately predict and control the pitch rate, thereby improving the operating stability and efficiency of the wind turbine.

[0088] Furthermore, the present application also proposes to inject a frequency sweep test signal during the idle period of the variable pitch system to collect the frequency response data of the actuator. Based on the response data, the damping coefficient and stiffness parameters in the transfer function are identified. When the parameter change exceeds the adaptive threshold, the weight matrix of the dynamic model is updated.

[0089] By injecting a sweep test signal during the idle period of the pitch system, collecting the frequency response data of the actuator, and identifying the damping coefficient and stiffness parameters in the transfer function based on these data, the dynamic model of the actuator can be monitored and updated in real time. When the parameter change exceeds the adaptive threshold, the weight matrix of the dynamic model is updated. This process ensures that the pitch system can maintain high response accuracy and stability under different environments and operating conditions, thereby effectively solving the response accuracy and stability problems of the pitch system in a dynamic environment.

[0090] The sweep test signal can be generated by a conventional signal generator, and the signal frequency range can be adjusted according to the operating frequency of the variable pitch system. The acquisition of frequency response data can be achieved by sensors installed on the actuator, which need to have high sensitivity and fast response capabilities. The identification of the transfer function can adopt existing system identification algorithms, such as the least squares method or the recursive least squares method, through which the damping coefficient and stiffness parameters can be accurately extracted. When it is detected that the parameter change exceeds the preset adaptive threshold, the system automatically triggers the update process of the weight matrix. The update of the weight matrix can be achieved through an optimization algorithm to ensure that the dynamic model can accurately reflect the current system state.

[0091] This technical solution ensures the high response accuracy and stability of the system under different environments and operating conditions by real-time monitoring and updating the dynamic model of the variable pitch system. Compared with the existing technology, this solution introduces a sweep frequency test signal and an adaptive threshold mechanism, which can timely detect and correct changes in system parameters, avoiding system performance degradation caused by parameter drift. As a result, the reliability and control accuracy of the variable pitch system have been significantly improved.

[0092] Furthermore, the present application also proposes to perform intrinsic orthogonal decomposition on the radar time-series velocity field data to extract the spatiotemporal evolution coefficients of the dominant flow modes; input the evolution coefficients into a pre-trained LSTM network to predict the flow field evolution trend for multiple rotation cycles in the future; and inject the predicted flow field data into the flow field-structure coupling model in advance to generate an advance compensation instruction component.

[0093] The present application includes the technical features of forward-looking flow field compensation. By performing intrinsic orthogonal decomposition on the radar time-series velocity field data, the spatiotemporal evolution coefficients of the dominant flow modes are extracted, and these evolution coefficients are input into the pre-trained LSTM network to predict the flow field evolution trend of multiple rotation cycles in the future. Then, the predicted flow field data is injected into the flow field-structure coupling model in advance to generate the advance compensation instruction component. These technical features can cooperate with each other to predict and compensate in advance the impact of future wind field changes on pitch angle control, thereby improving the dynamic response capability and control accuracy of wind turbines under complex wind conditions. Through this solution, the problem of the impact of future wind field changes on pitch angle control can be effectively solved, ensuring that wind turbines can maintain a stable operating state under different wind conditions and improve power generation efficiency.

[0094] Furthermore, the radar time series velocity field data is subjected to intrinsic orthogonal decomposition to extract the spatiotemporal evolution coefficients of the dominant flow modes. Through the intrinsic orthogonal decomposition method, the main flow modes can be extracted from the complex wind field data, and these modes represent the main change trends of the wind field. These evolution coefficients are input into the pre-trained LSTM network, and the time series prediction ability of the LSTM network is used to predict the flow field evolution trend of multiple rotation cycles in the future. The predicted flow field data is injected into the flow field-structure coupling model in advance, and the future aerodynamic load changes are calculated through the coupling model, and the advance compensation instruction component is generated. As a preferred embodiment, a variety of different intrinsic orthogonal decomposition algorithms, such as the standard POD method or the improved POD method, can be used to improve the accuracy of flow mode extraction. The training of the LSTM network can be based on a large amount of historical wind field data to ensure the accuracy and reliability of the prediction.

[0095] Specifically, adaptive compensation control

[0096] 1. Generate a three-dimensional velocity field in real time through heterogeneous radar, and input it into the flow field-structure coupling model after EKF registration; the model outputs the dynamic aerodynamic load distribution of each blade segment; the dynamic aerodynamic load distribution is input into the pre-trained radial basis function network, and the reference pitch angle is output; its network structure: 7 nodes in the input layer, 32 nodes in the hidden layer, and 3 nodes in the output layer (independent compensation of three blades);

[0097] Input layer input features: [,bending moment standard deviation, hub height wind speed, impeller speed, current pitch angle, hydraulic system oil temperature,];

[0098] 2. Anti-saturation optimization: Actuator status monitoring: real-time collection of hydraulic oil pressure range, displacement, and temperature;

[0099] Dynamic constraint model: predicts the pitch rate limit curve for the next 5 seconds and replans the instructions using a quadratic programming algorithm;

[0100] 3. By extracting the dominant modes in the flow field's spatiotemporal data, a linear approximate dynamic model is established.

[0101] 1. Data matrix construction: Arrange the velocity field sequence measured by radar into a time-space matrix:

[0102] , where n: number of spatial points, m: number of time steps;

[0103] 2. Singular Value Decomposition (SVD): Truncate to the first r main modes;

[0104] 3. Construct an approximate Kooman matrix: ,in is the time offset matrix;

[0105] 4. Future flow field prediction: Among them, the prediction step length k=3 is about 1.2 seconds, which compensates for the mechanical delay of the actuator in advance.

[0106] 5. Weight fusion: When the prediction error is >15%, the advance compensation weight is linearly reduced.

[0107] Therefore, this application can predict and compensate the impact of future wind field changes on pitch angle control in advance through forward-looking flow field compensation technology. Compared with the prior art, this application significantly improves the dynamic response capability and control accuracy of wind turbines under complex wind conditions, ensuring that wind turbines can maintain a stable operating state under different wind conditions and improve power generation efficiency. This technical solution not only solves the problem of the impact of future wind field changes on pitch angle control, but also improves the intelligent control level of wind turbines through forward-looking prediction and compensation mechanisms.

[0108] Furthermore, the present application also proposes to calculate the real-time error between the actual flow field data and the predicted data, construct a prediction credibility weight factor, and use a weighted fusion algorithm to mix the advance compensation instructions with the real-time compensation instructions. When the prediction error continues to increase, the weight ratio of the advance compensation instructions is gradually reduced until it is turned off.

[0109] By calculating the real-time error between the actual flow field data and the predicted data and constructing the prediction credibility weight factor, the reliability of the predicted data can be evaluated. The weighted fusion algorithm is used to mix the advance compensation instruction with the real-time compensation instruction, and the weight of the compensation instruction can be flexibly adjusted under different circumstances. When the prediction error continues to increase, the weight of the advance compensation instruction is gradually reduced until it is turned off, which can prevent unreliable prediction data from having a negative impact on the compensation effect, thereby improving the accuracy and reliability of the compensation instruction.

[0110] The real-time error between the actual flow field data and the predicted data is calculated by comparing the current actual measured flow field data with the flow field data predicted based on historical data. The construction of the prediction credibility weight factor is based on the size of the error. The smaller the error, the higher the credibility of the predicted data and the larger the weight factor. The weighted fusion algorithm generates the final compensation instruction by weighted averaging the advance compensation instruction and the real-time compensation instruction according to the weight factor. When the prediction error continues to increase, the system will automatically reduce the weight of the advance compensation instruction until it completely relies on the real-time compensation instruction to ensure the reliability of the compensation instruction.

[0111] The present application can effectively improve the accuracy and reliability of compensation instructions through the above method, especially in complex and changeable wind conditions, and can flexibly adjust the compensation strategy to avoid compensation failure caused by prediction errors. Compared with the prior art, the present application provides a more intelligent and adaptive compensation method, which can significantly improve the operating efficiency and stability of wind turbines.

[0112] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the above-mentioned embodiments. It should be understood that those skilled in the art can still modify the technical solutions recorded in the above-mentioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A pitch angle dynamic compensation method based on multi-radar spatiotemporal synchronization, characterized in that: include: The three-dimensional space-time velocity field of the wind rotor plane is constructed by synchronously measuring data from the horizontal scanning radar array and the vertical profile radar array to fully capture the wind speed changes around the wind rotor. The velocity field data is input into the flow field-structure coupling model to calculate the dynamic aerodynamic load distribution of each blade segment in real time to reflect the force conditions of the blade under different wind conditions; Based on the deviation between the aerodynamic load distribution data and the target load, an initial pitch compensation instruction set is generated; The initial instruction set is input into the dynamic constraint model of the actuator for anti-saturation optimization, and the final pitch angle control instruction is output and synchronized to the pitch actuator, wherein the instruction is replanned through a quadratic programming algorithm so that the pitch rate is limited within the dynamic constraint curve corresponding to the oil pressure threshold of the hydraulic system.

2. A pitch angle dynamic compensation method based on multi-radar spatiotemporal synchronization as claimed in claim 1, characterized in that: The construction process of the three-dimensional space-time velocity field includes: Receive the original point cloud data stream from multiple radar nodes and eliminate the coordinate offset caused by tower swing through the spatiotemporal registration algorithm; The registered point cloud data is mapped to the rotating coordinate system to generate a three-dimensional grid velocity field with the impeller center as the origin; Based on the difference of velocity field data in adjacent time steps, the spatial distribution of flow acceleration field is calculated.

3. A pitch angle dynamic compensation method based on multi-radar spatiotemporal synchronization as claimed in claim 2, characterized in that: The data processing of the flow field-structure coupling model includes: Convert the gridded velocity field to a vorticity tensor field: ; In the impeller rotating coordinate system, considering the Coriolis force and centrifugal force effects, the modified momentum equation is: in, is the impeller angular velocity vector, r is the position vector; The vorticity , take the curl and simplify it, and get the vorticity transport equation in the rotating coordinate system: in, Additional terms introduced for rotation effects; Shear force on the blade surface The shear force transfer function is constructed based on the relationship with vorticity. The transfer function form is: in, is the rotation correction coefficient; the integration domain A covers the blade surface microelement; The flow separation effect was characterized by fitting an asymmetric bimodal function using the least squares method: Mapping flow field to shear force: 3D velocity field observed by radar → Calculation of vorticity field ; By transferring the function Map the vorticity field to the shear force distribution on the blade surface; integrate the shear force to obtain the aerodynamic bending moment ; Real-time bending moment With target value Compare and calculate dynamic load deviation; Sliding window statistical deviation vector .

4. A pitch angle dynamic compensation method based on multi-radar spatiotemporal synchronization as claimed in claim 3, characterized in that: Generating an initial pitch compensation instruction set comprises: The dynamic load deviation vector is input into the pre-trained neural network model to map and generate the reference pitch angle increment of each blade; Superimpose the tower shadow interference feature vector constructed based on historical operation data to generate the initial pitch change command with pre-compensation; The initial instruction set is fused with the current pitch angle state data to generate a time-continuous pitch change instruction sequence.

5. A pitch angle dynamic compensation method based on multi-radar spatiotemporal synchronization as claimed in claim 4, characterized in that: The generation of the tower shadow interference feature vector includes: Obtain historical wind direction data streams from the SCADA system to construct a time-space distribution map of tower wake interference patterns; Extracting the velocity profile of the near-tower area in the radar data in real time, and performing convolution similarity matching with the distribution map; The tail vortex intensity coefficient and phase delay parameter are extracted according to the matching results, and the time-varying interference feature vector is constructed.

6. The pitch angle dynamic compensation method based on multi-radar spatiotemporal synchronization as claimed in claim 1, characterized in that: The data processing of the anti-saturation optimization includes: Receive real-time status data streams from pitch actuators, including hydraulic oil pressure, actuator displacement, and temperature monitoring values; The state data is input into the actuator dynamic model to predict the pitch rate limit curve in the future time window; The initial instruction set is replanned using a constrained optimization algorithm so that the pitch trajectory falls within the range of the rate limit curve.

7. A pitch angle dynamic compensation method based on multi-radar time-space synchronization as claimed in claim 6, characterized in that: The updating of the actuator dynamic model includes: Inject a frequency sweep test signal during the idle period of the pitch system to collect the frequency response data of the actuator; Identify the damping coefficient and stiffness parameters in the transfer function based on the response data; When the parameter change exceeds the adaptive threshold, the weight matrix of the dynamic model is updated.

8. The pitch angle dynamic compensation method based on multi-radar time-space synchronization as claimed in claim 1, characterized in that: The method also includes forward-looking flow field compensation: The intrinsic orthogonal decomposition of the radar time series velocity field data is performed to extract the spatiotemporal evolution coefficients of the dominant flow mode; The evolution coefficients are input into the pre-trained LSTM network to predict the flow field evolution trend for multiple future rotation cycles; The predicted flow field data is injected into the flow field-structure coupling model in advance to generate an advance compensation instruction component.

9. A pitch angle dynamic compensation method based on multi-radar time-space synchronization as claimed in claim 8, characterized in that: The integration of the advance compensation instruction includes: Calculate the real-time error between the actual flow field data and the predicted data, and construct the prediction credibility weight factor; A weighted fusion algorithm is used to mix the advance compensation instructions with the real-time compensation instructions; When the prediction error continues to increase, the weight of the advance compensation instruction is gradually reduced until it is turned off.

Images