Safe operation method and system for subsynchronous oscillation wind power integration system

Abstract

The invention relates to the technical field of wind power integration control, and discloses a wind power integration system safe operation method aiming at subsynchronous oscillation and a system thereof. The method comprises the following steps: performing time-frequency feature extraction processing on an original data set to obtain an oscillation mode feature quantity; performing stability analysis processing on the oscillation modal characteristic quantity to obtain a subsynchronous oscillation risk early warning signal; performing multi-time-scale decoupling processing on the sub-synchronous oscillation risk early warning signal to obtain a mechanical-electrical coordination control instruction; performing task decomposition processing on the coordination control instruction to obtain a real-time control quantity of each execution mechanism; and performing evaluation processing and correction on the execution effect of the real-time control quantity of each execution mechanism to obtain a correction control parameter set, and performing real-time control on the wind power integration system through the correction control parameter set. According to the invention, coordinated control of the electrical subsystem and the mechanical subsystem of the wind power grid-connected system is realized, subsynchronous oscillation is effectively suppressed, and the operation stability of the system is improved.

Description

Technical Field

[0001] The present application relates to the field of wind power grid-connected control, and in particular to a method and system for safe operation of a wind power grid-connected system targeting subsynchronous oscillation. Background Art

[0002] With the continuous expansion of wind power grid-connected scale, the subsynchronous oscillation problem of large wind farms has become increasingly prominent. The existing wind power grid-connected system mainly suppresses subsynchronous oscillations by setting up damping controllers, power oscillation suppressors and other devices, and adopts a fast detection method based on state observation and a control strategy based on multiple feedback to suppress oscillations. At the same time, the industry has also developed a variety of advanced control methods including adaptive control and model predictive control to improve the system's anti-oscillation ability. These methods have alleviated the adverse effects of subsynchronous oscillations to a certain extent.

[0003] However, the existing technology still has the following shortcomings: First, traditional oscillation suppression methods often separate the electrical system and the mechanical system for independent control, ignoring the coupling effect between the two subsystems, resulting in unsatisfactory control effect; second, the existing oscillation detection method is mainly based on a single time scale for analysis, which cannot accurately capture the oscillation characteristics of the system in different frequency bands; finally, the design of the control strategy lacks consideration of the overall performance of the system, making it difficult to maintain a good oscillation suppression effect under complex working conditions. Summary of the invention

[0004] The present application provides a method and system for safe operation of a wind power grid-connected system for subsynchronous oscillations, which are used to achieve coordinated control of the electrical subsystem and mechanical subsystem of the wind power grid-connected system, effectively suppress subsynchronous oscillations, and improve the operating stability of the system.

[0005] In a first aspect, the present application provides a method for the safe operation of a wind power grid-connected system against subsynchronous oscillations, and the method for the safe operation of a wind power grid-connected system against subsynchronous oscillations comprises: performing multi-source data acquisition and processing on electrical parameters and mechanical parameters of the wind power grid-connected system to obtain an original data set; performing time-frequency feature extraction and processing on the original data set to obtain oscillation modal feature quantities; performing stability analysis and processing on the oscillation modal feature quantities to obtain a subsynchronous oscillation risk warning signal; performing multi-time scale decoupling processing on the subsynchronous oscillation risk warning signal to obtain a mechanical-electrical coordinated control instruction; performing task decomposition processing on the coordinated control instruction to obtain a real-time control quantity of each actuator; evaluating and processing the execution effect of the real-time control quantity of each actuator and correcting it to obtain a corrected control parameter set, and performing real-time control of the wind power grid-connected system through the corrected control parameter set.

[0006] In a second aspect, the present application provides a wind power grid-connected system safety operation system for subsynchronous oscillations, the wind power grid-connected system safety operation system for subsynchronous oscillations comprising: The acquisition module is used to perform multi-source data acquisition and processing on the electrical parameters and mechanical parameters of the wind power grid-connected system to obtain the original data set; An extraction module is used to perform time-frequency feature extraction processing on the original data set to obtain oscillation mode feature quantities; An analysis module is used to perform stability analysis on the oscillation modal characteristic quantity to obtain a subsynchronous oscillation risk warning signal; A decoupling module, used for performing multi-time scale decoupling processing on the sub-synchronous oscillation risk warning signal to obtain a mechanical-electrical coordinated control instruction; The decomposition module is used to perform task decomposition processing on the coordinated control instructions to obtain the real-time control quantity of each actuator; The correction module is used to evaluate and correct the execution effect of the real-time control quantity of each actuator to obtain a corrected control parameter set, and to control the wind power grid-connected system in real time through the corrected control parameter set.

[0007] In the technical solution provided by the present application, by performing multi-source data collection and processing on the electrical parameters and mechanical parameters of the wind power grid-connected system, the system operation status information can be fully obtained, laying a data foundation for the subsequent oscillation feature extraction and analysis; the time-frequency feature extraction and processing of the original data set can accurately capture the oscillation characteristics of the system in both the time domain and the frequency domain, and improve the accuracy of oscillation mode recognition; the stability analysis and processing of the oscillation mode feature quantity can timely discover potential unstable factors in the system and realize early warning of subsynchronous oscillations; by performing multi-time scale decoupling processing on the subsynchronous oscillation risk warning signal, the oscillation components of different frequency bands in the system can be separated. The system can decompose the coordinated control instructions into tasks, reasonably allocate the control tasks to each actuator, and ensure the executability of the control instructions; evaluate and correct the execution effect of the real-time control quantity of each actuator, and dynamically adjust the control parameters to continuously improve the control performance; the entire control process forms a closed-loop adaptive control system, which not only takes into account the coupling characteristics of the electrical system and the mechanical system, but also realizes coordinated control on multiple time scales. At the same time, the real-time evaluation and correction mechanism ensures the continuous optimization of the control effect, thereby effectively improving the operating reliability and stability of the wind power grid-connected system. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0009] Figure 1 A schematic diagram of an embodiment of a method for safe operation of a wind power grid-connected system for subsynchronous oscillation in an embodiment of the present application; Figure 2 It is a schematic diagram of an embodiment of a safe operation system of a wind power grid-connected system for subsynchronous oscillation in an embodiment of the present application. DETAILED DESCRIPTION

[0010] The embodiment of the present application provides a method and system for the safe operation of a wind power grid-connected system for subsynchronous oscillation. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0011] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiment of the present application, an embodiment of a safe operation method for a wind power grid-connected system for subsynchronous oscillation includes: Step S101, performing multi-source data collection and processing on electrical parameters and mechanical parameters of the wind power grid-connected system to obtain an original data set; Step S102, performing time-frequency feature extraction processing on the original data set to obtain oscillation mode feature quantities; Step S103, performing stability analysis on the oscillation modal characteristic quantity to obtain a subsynchronous oscillation risk warning signal; Step S104, performing multi-time scale decoupling processing on the subsynchronous oscillation risk warning signal to obtain a mechanical-electrical coordinated control instruction; Step S105, performing task decomposition processing on the coordinated control instruction to obtain the real-time control amount of each actuator; Step S106: Evaluate and correct the execution effect of the real-time control quantity of each actuator to obtain a corrected control parameter set, and use the corrected control parameter set to perform real-time control on the wind power grid-connected system.

[0012] It is understandable that the execution subject of the present application may be a wind power grid-connected system safety operation system for subsynchronous oscillation, or a terminal or a server, which is not limited here. The present application embodiment is described by taking a server as the execution subject as an example.

[0013] Specifically, multi-source data acquisition processing is performed. In this step, voltage and current signals are collected at a sampling frequency of 10kHz by sampling devices arranged at the stator end and the rotor end. The collected data include stator voltage, stator current, rotor voltage and rotor current. The data are processed by a digital low-pass filter to filter out high-frequency noise higher than 1kHz to obtain electrical parameter data. At the same time, the wind wheel speed and generator rotor speed are collected with a sampling frequency of 1kHz. The speed spectrum characteristics in the range of 0.1-50Hz are obtained through spectrum analysis. The vibration signal is collected on the transmission shaft with a sampling frequency of 5kHz. The measurement noise is eliminated by wavelet transform, and the vibration characteristics in the range of 0.1-100Hz are extracted. Finally, these data are time-aligned and fused to form a complete original data set. The time-frequency features of the original data set are extracted, the voltage and current signals are converted to the frequency domain through Fourier transform, the spectrum data in the range of 0.1-50Hz is extracted, and the main harmonic components in the electrical system are identified; the speed signal is decomposed by wavelet, and the db4 wavelet basis function is selected to decompose the signal into 4 layers, corresponding to the speed characteristics of different frequency bands; the vibration signal is subjected to empirical mode decomposition, and the complex vibration signal is decomposed into multiple intrinsic mode functions, each of which represents an inherent vibration mode; then the correlation between electrical characteristics and mechanical characteristics is determined through cross-correlation analysis, and finally the oscillation modal characteristic quantity of the system is obtained.

[0014] After obtaining the oscillation modal characteristic quantity, stability analysis is performed, the system characteristic value is calculated, the real and imaginary parts of the characteristic value are extracted, and the damping ratio of each oscillation mode is calculated. The mode with a damping ratio less than 5% is marked as a potential unstable mode; the energy distribution analysis of these modes is performed, the oscillation intensity of each frequency band is calculated, and the risk level is divided into three levels: low, medium, and high according to the magnitude of the oscillation intensity. The risk probability is updated through Bayesian reasoning to generate a sub-synchronous oscillation risk warning signal. Based on the warning signal, multi-time scale decoupling is performed, and the oscillation signal is divided into a fast oscillation component (greater than 5Hz) and a slow oscillation component (0.1-5Hz) according to the frequency characteristics, and the control requirements of the electrical subsystem and the mechanical subsystem are extracted respectively; the system coupling characteristics are analyzed, the control target constraints are established, the benchmark control sequence is generated and dynamic compensation is performed, and the mechanical-electrical coordinated control instructions are formed.

[0015] The coordinated control instructions are decomposed, and the control tasks are assigned to actuators such as converter control, pitch control, and excitation control. The control parameters required by each actuator are calculated, and timing coordination is performed to ensure the coordination of control actions. Finally, the control effect is evaluated in real time, the deviation of the control parameters is calculated, and the corrected control parameter set is obtained through limiting and parameter updating to achieve closed-loop control of the wind power grid-connected system.

[0016] For example, when the unit operating power is 1.5MW, the original data of stator voltage 220V (effective value), stator current 1.2kA, rotor speed 1500rpm, and transmission shaft vibration amplitude 0.5mm are obtained through sampling. After time-frequency feature extraction, an oscillation component with an amplitude of 0.3pu is identified near 15Hz. The damping ratio of this oscillation mode is 3% through damping ratio calculation, triggering a medium-level warning signal. The system decomposes the oscillation into an electrical oscillation component of 12Hz and a mechanical oscillation component of 3Hz, and generates the current control command (1.0pu) of the rotor-side converter and the angle control command (2 degrees) of the pitch system respectively. After the control is executed, the system oscillation amplitude is reduced to below 0.1pu, and the damping ratio is increased to 8%, achieving effective suppression of subsynchronous oscillations.

[0017] In the implementation of this application, by performing multi-source data collection and processing on the electrical parameters and mechanical parameters of the wind power grid-connected system, the system operation status information can be fully obtained, laying a data foundation for the subsequent oscillation feature extraction and analysis; the time-frequency feature extraction and processing of the original data set can accurately capture the oscillation characteristics of the system in both the time domain and the frequency domain, and improve the accuracy of oscillation mode identification; the stability analysis and processing of the oscillation mode feature quantity can timely discover potential unstable factors in the system and realize early warning of subsynchronous oscillations; by performing multi-time scale decoupling processing on the subsynchronous oscillation risk warning signal, the oscillation components of different frequency bands in the system can be separated. Develop control strategies in a targeted manner to improve control accuracy; decompose coordinated control instructions to reasonably allocate control tasks to various actuators to ensure the executability of control instructions; evaluate and correct the execution effect of the real-time control quantity of each actuator, and dynamically adjust control parameters to continuously improve control performance; the entire control process forms a closed-loop adaptive control system, which not only takes into account the coupling characteristics of the electrical system and the mechanical system, but also realizes coordinated control on multiple time scales. At the same time, the real-time evaluation and correction mechanism ensures the continuous optimization of the control effect, thereby effectively improving the operating reliability and stability of the wind power grid-connected system.

[0018] In a specific embodiment, the process of executing step S101 may specifically include the following steps: (1) synchronously sampling and processing the voltage signal and current signal at the stator end of the wind power grid-connected system to obtain stator electrical parameter data, and synchronously sampling and processing the voltage signal and current signal at the rotor end of the wind power grid-connected system to obtain rotor electrical parameter data; (2) Perform digital filtering on the stator electrical parameter data and the rotor electrical parameter data to obtain filtered electrical signal data, and perform power parameter calculation on the filtered electrical signal data to obtain real-time power data of the system; (3) Perform speed collection and processing on the wind turbine rotor speed and generator rotor speed of the wind power grid-connected system to obtain mechanical speed data, and perform spectrum analysis on the mechanical speed data to obtain speed spectrum characteristic data; (4) Collect and process the vibration characteristics of the transmission shaft vibration signal of the wind power grid-connected system to obtain the original vibration data, and eliminate the noise of the original vibration data through wavelet transform to obtain the vibration characteristic data after noise reduction; (5) Perform data fusion processing on the system real-time power data, speed spectrum feature data, and vibration feature data after noise reduction to obtain the original data set.

[0019] Specifically, high-precision voltage sensors and current sensors are installed at the stator end of the wind turbine, and synchronous sampling technology is used to collect the three-phase voltage and current signals of the stator. The sampling frequency is set to 10kHz to ensure the sampling accuracy while satisfying the Nyquist sampling theorem. The collected original voltage signal range is 0-1000V, and the current signal range is 0-5000A. The signal is normalized to the range of ±10V through the signal conditioning circuit, and digitized through a 16-bit AD converter to obtain the stator electrical parameter data. At the same time, the same sampling scheme is used at the rotor end to synchronously sample the three-phase voltage and current of the rotor. The rotor voltage signal range is 0-690V, and the current signal range is 0-2000A. After the same signal conditioning and AD conversion process, the rotor electrical parameter data is obtained. Then the collected stator and rotor electrical parameter data are digitally filtered, and a Butterworth low-pass filter is used with a cutoff frequency set to 1kHz to filter out high-frequency noise and interference components. The filtered data is used to calculate the real-time power parameters of the system, including active power and reactive power on the stator side, active power and reactive power on the rotor side, etc. The power calculation adopts the three-phase instantaneous power theory, calculates the corresponding power value for each sampling point, outputs a set of average power data every 100ms, and obtains the real-time power data of the system.

[0020] In terms of mechanical parameter acquisition, the speed signal is collected by the speed sensor installed on the wind rotor main shaft and the generator shaft, with a sampling frequency of 1kHz. The speed signal is digitally processed after photoelectric conversion and signal conditioning to obtain mechanical speed data. The mechanical speed data is subjected to fast Fourier transform to analyze the speed fluctuation characteristics in the frequency range of 0.1-50Hz, focusing on the sub-synchronous frequency components in the range of 1-5Hz to obtain the speed spectrum characteristic data. For vibration monitoring of the transmission system, acceleration sensors are installed at key positions of the transmission shaft, and the sampling frequency is set to 5kHz. The collected vibration acceleration signal is passed through a charge amplifier and signal conditioning to obtain the original vibration data. The db4 wavelet basis function is used to perform a 4-layer wavelet decomposition on the original vibration data to remove the noise component in the highest frequency band, retain the medium and low frequency signals containing the system vibration characteristics, and obtain the noise-reduced vibration characteristic data through wavelet reconstruction.

[0021] Finally, the real-time power data, speed spectrum characteristic data and vibration characteristic data after noise reduction are fused. The data fusion uses timestamp alignment to unify the data with different sampling frequencies into a 100ms time interval, establish a multidimensional data matrix containing electrical parameters and mechanical parameters, and form a complete original data set.

[0022] Taking a 1.5MW doubly-fed wind turbine as an example, when the wind speed is 10m / s, the stator end sampling obtains the effective value of the A-phase voltage of 220V and the current of 1200A, and the rotor end A-phase voltage of 150V and the current of 500A. After digital filtering and power calculation, the active power of the stator side is 1.2MW and the reactive power is 0.3MVar, and the active power of the rotor side is 0.3MW and the reactive power is 0.1MVar. At the same time, the wind wheel speed is measured to be 15rpm and the generator rotor speed is 1500rpm. Through FFT analysis, it is found that there is a speed fluctuation with an amplitude of 0.1rpm at 2.5Hz. The original amplitude of the transmission shaft vibration signal is 2g, which is reduced to 1.5g after wavelet noise reduction processing, and the main vibration frequency is concentrated around 3Hz. After these data are time-aligned, a set of data vectors containing 22 parameters is formed every 100ms, forming the original data set for subsequent analysis.

[0023] In a specific embodiment, the process of executing step S102 may specifically include the following steps: (1) Perform Fourier transform processing on the voltage signal and current signal in the original data set to obtain electrical signal spectrum data, and perform harmonic component extraction processing on the electrical signal spectrum data to obtain electrical harmonic characteristic data; (2) Perform wavelet decomposition on the speed signal in the original data set to obtain multi-scale speed characteristic data, and perform principal component analysis on the multi-scale speed characteristic data through singular value decomposition to obtain the main oscillation components of the speed; (3) Performing empirical mode decomposition on the vibration signal in the original data set to obtain an intrinsic mode function set, and performing envelope analysis on the intrinsic mode function set to obtain vibration amplitude characteristic data; (4) Perform cross-correlation analysis on the electrical harmonic characteristic data and the main oscillation components of the speed to obtain the electro-mechanical coupling characteristic data, and perform feature fusion processing on the electro-mechanical coupling characteristic data and the vibration amplitude characteristic data to obtain the multi-source oscillation characteristic data; (5) Perform modal separation processing on the multi-source oscillation characteristic data to obtain the oscillation modal characteristic quantities.

[0024] Specifically, the second step of the safe operation method of the wind power grid-connected system for subsynchronous oscillation is to extract time-frequency features. First, the voltage and current signals in the original data set are processed by Fourier transform to convert the time domain signals into the frequency domain. The Fourier transform uses a 1024-point FFT algorithm to transform the data once every 10ms to obtain spectrum data in the range of 0-500Hz. The fundamental component (50Hz) and the harmonic components in the subsynchronous frequency range (0.1-50Hz) are extracted from the spectrum data, and the amplitude and phase information of each frequency point are recorded. The extracted harmonic components are sorted by amplitude, and the components with an amplitude exceeding 1% of the fundamental wave are selected as the main harmonic features to form electrical harmonic feature data. The speed signal in the original data set is decomposed by wavelet, and the db4 wavelet basis function is used for 4-layer decomposition. Wavelet decomposition divides the speed signal into different frequency bands: the first layer corresponds to 25-50Hz, the second layer corresponds to 12.5-25Hz, the third layer corresponds to 6.25-12.5Hz, the fourth layer corresponds to 3.125-6.25Hz, and the approximate component corresponds to 0-3.125Hz. The Hankel matrix is ​​constructed for the obtained multi-scale speed feature data, and singular value decomposition is performed. Singular value decomposition decomposes the data matrix into a left singular matrix, a singular value diagonal matrix, and a right singular matrix. The eigenvectors corresponding to the largest three singular values ​​are selected to reconstruct the main oscillation components of the speed.

[0025] The vibration signal is subjected to empirical mode decomposition, and the complex signal is decomposed into several intrinsic mode functions through an iterative screening process. The empirical mode decomposition first finds all the local extreme points of the signal, forms the upper and lower envelopes through cubic spline interpolation, calculates the mean envelope, subtracts the mean envelope from the original signal to obtain the first intrinsic mode function, and repeats this process until the remaining signal becomes a monotonic function. Hilbert transform is performed on each intrinsic mode function to obtain the instantaneous amplitude and instantaneous frequency to form vibration amplitude characteristic data. The electrical harmonic characteristic data and the main oscillation components of the speed are cross-correlated, the correlation coefficients under different time delays are calculated, the time delay corresponding to the maximum correlation coefficient is found, and the correlation between electrical oscillation and mechanical oscillation is determined. The cross-correlation analysis results are fused with the vibration amplitude characteristic data at the feature level, and the three characteristic data are normalized and combined using the weighted average method to obtain multi-source oscillation characteristic data. Finally, the multi-source oscillation characteristic data is modally separated, and the mixed oscillation characteristics are separated into independent oscillation modes using the independent component analysis method. Independent component analysis is based on the statistical independence of signals. It estimates the separation matrix by maximizing the non-Gaussianity method to obtain the characteristic quantities of each independent oscillation mode.

[0026] The data processing process is illustrated by taking a 1.5MW doubly-fed wind turbine as an example: FFT analysis is performed on the collected stator voltage signal. In addition to the fundamental wave of 50Hz, three main harmonic components of 15Hz, 30Hz and 45Hz are detected, with amplitudes of 5%, 3% and 2% of the fundamental wave respectively. Wavelet decomposition is performed on the 1500rpm generator speed signal. A fluctuation with an amplitude of 5rpm is detected in the third layer (6.25-12.5Hz), which is confirmed as the main oscillation component by singular value decomposition. Five intrinsic mode functions are obtained by empirical mode decomposition of the transmission shaft vibration signal, among which the second intrinsic mode function shows significant vibration characteristics near 7.5Hz with an amplitude of 0.8g. Through cross-correlation analysis, it is found that there is a time delay of 0.1s between the 15Hz electrical harmonic and the 7.5Hz mechanical vibration, and the correlation coefficient reaches 0.85, indicating that there is an obvious coupling relationship between the two oscillations. Finally, three main oscillation modes are obtained through modal separation: a 15 Hz electrical oscillation mode, a 7.5 Hz mechanical oscillation mode, and a subsynchronous oscillation mode generated by the coupling of the two.

[0027] In a specific embodiment, the process of executing step S103 may specifically include the following steps: (1) Calculate the characteristic roots of the oscillation modal characteristic quantities to obtain a set of system eigenvalues, and calculate the damping ratio of the system eigenvalue set to obtain modal damping ratio data; (2) Perform critical stability judgment processing on the modal damping ratio data to obtain oscillation instability degree data, and perform energy distribution analysis processing on the oscillation instability degree data to obtain oscillation energy distribution characteristics; (3) Perform frequency band decomposition processing on the oscillation energy distribution characteristics to obtain the oscillation intensity data of each frequency band, and perform threshold comparison processing on the oscillation intensity data of each frequency band to obtain the frequency band warning level data; (4) Perform risk assessment processing on the frequency band warning level data to obtain oscillation risk level data, and perform probability update processing on the oscillation risk level data through Bayesian reasoning to obtain risk probability distribution data; (5) Conduct comprehensive risk analysis and processing on the risk probability distribution data to obtain subsynchronous oscillation risk warning signals.

[0028] Specifically, the characteristic roots of the oscillation modal characteristic quantities are calculated, and the characteristic equation is constructed through the state space matrix, and the system eigenvalue set is solved. Each eigenvalue contains a real part and an imaginary part. The real part represents the attenuation rate of the oscillation, and the imaginary part represents the angular frequency of the oscillation. The damping ratio of each oscillation mode is calculated based on the eigenvalue. The damping ratio is equal to the absolute value of the real part of the eigenvalue divided by the modulus of the eigenvalue to obtain the modal damping ratio data matrix. The modal damping ratio data is judged for critical stability, and the damping ratio critical value is set to 5%. The oscillation mode with a damping ratio lower than the critical value is marked as a potentially unstable mode. According to the difference between the value of the damping ratio and the critical value, the degree of oscillation instability is calculated, and the energy distribution analysis of the unstable oscillation is performed. The energy distribution analysis is calculated using the following formula: in: is the oscillation energy distribution at frequency f; N is the number of oscillation modes; is the weight coefficient of the i-th mode; is the power spectral density of the ith mode at frequency f; is the attenuation coefficient; is the damping ratio of the ith mode at frequency f.

[0029] The calculated oscillation energy distribution characteristics are divided into multiple frequency bands according to the frequency, and the width of each frequency band is 5Hz. The oscillation intensity in each frequency band is calculated. The oscillation intensity is determined according to the energy integral value in the frequency band. The frequency band oscillation intensity is compared with the preset threshold value. The threshold value is divided into three levels: low risk (less than 30%), medium risk (30%-70%) and high risk (greater than 70%), and the frequency band warning level data is generated. The frequency band warning level data is risk assessed, and the oscillation risk level data is obtained by comprehensively considering factors such as oscillation frequency, energy size, and duration. The risk level is dynamically updated through Bayesian reasoning. Bayesian reasoning establishes prior probability based on historical data, and updates posterior probability based on real-time monitoring data to obtain risk probability distribution data.

[0030] Finally, a comprehensive analysis of the risk probability distribution data is conducted, and a subsynchronous oscillation risk warning signal is generated based on the risk probability and energy distribution characteristics of each frequency band. The warning signal contains information such as oscillation frequency, risk level, and development trend.

[0031] In a specific embodiment, the process of executing step S104 may specifically include the following steps: (1) Perform time-scale layered processing on the subsynchronous oscillation risk warning signal to obtain fast oscillation components and slow oscillation components, and extract electrical characteristics of the fast oscillation components to obtain the control requirements of the electrical subsystem; (2) Extract the mechanical characteristics of the slow oscillation component to obtain the control requirements of the mechanical subsystem, and perform coupling analysis on the control requirements of the electrical subsystem and the mechanical subsystem to obtain system coupling characteristic data; (3) Decompose the system coupling characteristic data into control targets to obtain the control target set of each subsystem, and perform coordination constraint processing on the control target set of each subsystem to obtain coordination control constraint conditions; (4) Perform control sequence generation processing on the coordinated control constraints to obtain a benchmark control sequence, and perform dynamic compensation processing on the benchmark control sequence to obtain a compensation control amount; (5) The compensation control quantity is comprehensively processed by coordinated control to obtain mechanical-electrical coordinated control instructions.

[0032] Specifically, the subsynchronous oscillation risk warning signal is processed in a time-scale hierarchical manner, and the wavelet packet decomposition method is used to divide the signal into different frequency bands. The frequency above 5Hz is defined as a fast oscillation component, which is mainly related to the electrical subsystem; the frequency between 0.1-5Hz is defined as a slow oscillation component, which is mainly related to the mechanical subsystem. The electrical characteristics of the fast oscillation component are extracted, and the parameters including voltage fluctuation rate, current harmonic content, and power fluctuation amplitude are analyzed. The voltage fluctuation allowable range is set to ±5%, the current harmonic content limit is 5%, and the power fluctuation limit is 10%, and the control demand data of the electrical subsystem is generated. The mechanical characteristics of the slow oscillation component are extracted, and the parameters including speed fluctuation amplitude, transmission shaft torsional vibration amplitude, and blade vibration frequency are analyzed. The speed fluctuation limit is set to ±2%, the torsional vibration amplitude limit is 0.5°, and the vibration acceleration limit is 2g, and the control demand data of the mechanical subsystem is formed. The electrical and mechanical control requirements are coupled and analyzed, and the electro-mechanical coupling matrix is ​​established. The matrix elements represent the coupling strength between different parameters. The coupling coefficient is calculated using the correlation analysis method to generate the system coupling characteristic data.

[0033] The control objectives of the system coupling characteristic data are decomposed, and the overall control objectives are decomposed into electrical control objectives and mechanical control objectives. The electrical control objectives include: stabilizing bus voltage, suppressing power fluctuations, and reducing harmonic content; the mechanical control objectives include: smoothing speed fluctuations, reducing transmission shaft stress, and suppressing structural vibration. The control objectives of each subsystem are processed with coordination constraints, and a set of constraints is established, including control quantity range constraints, change rate constraints, stability constraints, etc., to generate coordinated control constraints. Based on the coordinated control constraints, the rolling time domain optimization method is used to generate the benchmark control sequence. The control sequence includes: converter control instructions (such as dq axis current reference value), pitch control instructions (such as pitch angle given value), excitation control instructions (such as excitation voltage adjustment value), etc. Dynamic compensation is performed on the benchmark control sequence, the control parameters are adjusted according to the real-time operating status, the compensation value is calculated, and the compensation control quantity is obtained.

[0034] Finally, the compensation control quantity is coordinated and comprehensively processed to ensure the coordination and timing between the control quantities, forming a complete mechanical-electrical coordinated control instruction. The control instructions include: converter PWM control signal, pitch actuator control signal, excitation system control signal, etc.

[0035] Take a 1.5MW doubly-fed generator set as an example: a 15Hz subsynchronous oscillation warning signal is detected, and the oscillation amplitude is 8% of the rated power. Through time scale decomposition, it is separated into 12Hz electrical oscillation (amplitude 6%) and 3Hz mechanical oscillation (amplitude 4%). Electrical feature extraction shows: voltage fluctuation 3.5%, current harmonic 4.2%, power fluctuation 8%; mechanical feature extraction shows: speed fluctuation 1.8%, torsional amplitude 0.4°, vibration acceleration 1.5g. Coupling analysis shows that the coupling coefficient between electrical oscillation and mechanical oscillation is 0.72. After control target decomposition and constraint processing, the benchmark control sequence is generated: converter d-axis current adjustment -50A, q-axis current adjustment +30A, pitch angle adjustment -1.5°. Dynamic compensation compensates the d-axis current by +10A according to the real-time power fluctuation, and finally forms a coordinated control instruction: converter PWM control duty cycle adjustment 2%, pitch rate control 0.5° / s.

[0036] In a specific embodiment, the process of executing step S105 may specifically include the following steps: (1) Perform functional hierarchical division on the mechanical-electrical coordinated control instructions to obtain the main control layer instructions and the device layer instructions, and perform control task analysis on the main control layer instructions to obtain the key control parameter set; (2) Perform execution unit allocation processing on the device layer instructions to obtain the control sequence of each execution unit, and perform timing coordination processing on the control sequence of each execution unit to obtain control timing data; (3) Calculate the control quantity of the key control parameter set to obtain basic control data, and match the basic control data with the control timing data to obtain the actuator control scheme; (4) Perform control component extraction on the actuator control scheme to obtain a set of candidate control quantities, and perform constraint verification on the set of candidate control quantities to obtain the real-time control quantity of each actuator.

[0037] Specifically, the mechanical-electrical coordinated control instructions are divided into two levels: the main control layer and the equipment layer. The main control layer instructions are responsible for the formulation of the overall control strategy, including oscillation suppression targets, control priorities, coordination strategies, etc.; the equipment layer instructions are responsible for the control of specific execution units, including converter control, pitch control, excitation control, etc. The main control layer instructions are task parsed to extract key control parameters, such as oscillation suppression target value, control bandwidth, response time and other parameters, to form a key control parameter set. The equipment layer instructions are assigned to execution units, and the instructions are assigned to the corresponding actuators according to the characteristics of the control tasks. The converter control unit is responsible for electrical oscillation suppression, the pitch control unit is responsible for mechanical oscillation suppression, and the excitation control unit is responsible for voltage regulation. The control sequences of each execution unit are coordinated in time, the order and execution time of the control actions are determined, and control timing data including execution time, duration, and switching conditions are generated.

[0038] The control quantity is calculated for the key control parameter set, and the calculation process uses the following formula: in, is the control quantity at time t; M is the number of control parameters; is the weight factor of the kth parameter; Adjust the gain for the parameter; is the reference value; is the damping coefficient; is the deviation value; is the time attenuation factor. The calculated basic control data is matched with the control timing data, and the control quantity is organized according to the execution timing to form a complete actuator control scheme. The control scheme is extracted, and the control quantity is decomposed into specific control components required by each actuator, such as the modulation ratio of the converter, the angle increment of the pitch, the voltage regulation of the excitation, etc., to form a set of alternative control quantities. The set of alternative control quantities is constrained and checked to check whether each control quantity meets the physical constraints (such as the pitch angle is limited to 0-90 degrees), dynamic constraints (such as the pitch rate is limited to within 8 degrees / second) and stability constraints (such as maintaining a certain stability margin), and the real-time control quantity of each actuator that meets the constraint conditions is obtained. Take a 1.5MW doubly fed generator set to handle subsynchronous oscillation as an example: after detecting a 15Hz oscillation, the main control layer instruction sets the oscillation suppression target to reduce the oscillation amplitude to within 3% of the rated value, and the response time requirement is 100ms. Equipment layer instruction allocation: converter control time resolution 5ms, pitch control time resolution 50ms. Through the control quantity calculation, the converter d-axis current reference value is set to -50A, the q-axis current reference value is set to +30A; the pitch angle reference adjustment is -1.5 degrees. Considering the converter current limit (±1.2 times the rated value) and the pitch rate limit (8 degrees / second), it is finally determined that the converter d-axis current real-time control quantity is -45A, the q-axis current real-time control quantity is +25A, the pitch angle real-time adjustment rate is 6 degrees / second, and the adjustment direction is to decrease.

[0039] In a specific embodiment, the process of executing step S106 may specifically include the following steps: (1) Perform deviation calculation on the response data of the real-time control quantity of each actuator to obtain control deviation data, and perform evaluation index calculation on the control deviation data to obtain performance evaluation results; (2) Calculating the control parameter correction amount on the performance evaluation results to obtain the parameter correction amount, and limiting the parameter correction amount to obtain the effective correction amount; (3) performing control parameter updating processing on the effective correction amount to obtain a progressive correction parameter, and performing comprehensive processing on the progressive correction parameter to obtain a corrected control parameter set; (4) The modified control parameter set is distributed and processed to obtain the modified instructions of each controller, and the modified instructions of each controller are synchronously issued to complete the real-time control of the wind power grid-connected system.

[0040] Specifically, the deviation of the real-time control quantity response of each actuator is calculated, and the difference between the actual response value and the target value is calculated. The converter control deviation includes the deviation between the actual value of the dq axis current and the given value, with a sampling period of 1ms; the pitch control deviation includes the deviation between the actual pitch angle and the given value, with a sampling period of 10ms; the excitation control deviation includes the deviation between the actual excitation voltage and the given value, with a sampling period of 5ms. The calculated control deviation data is evaluated for performance indicators, including steady-state deviation (required to be less than 1%), overshoot (required to be less than 10%), adjustment time (required to be less than 200ms), oscillation attenuation rate (required to be greater than 0.1 / s), etc. The comprehensive performance evaluation results are obtained through weighted calculation. The control parameter correction amount is calculated based on the performance evaluation results. The calculation of the correction amount takes into account the size, change trend and duration of the control deviation. For converter control, when the d-axis current deviation exceeds 5%, the proportional gain correction is 0.2 times the deviation value; for pitch control, when the pitch angle deviation exceeds 0.5 degrees, the response gain correction is 0.3 times the deviation value; for excitation control, when the excitation voltage deviation exceeds 3%, the integral time correction is 0.1 times the original value. The calculated parameter correction is limited to ensure that the corrected parameter changes are within a safe range: the converter control parameter correction is limited to ±20%, the pitch control parameter correction is limited to ±15%, and the excitation control parameter correction is limited to ±10%, so as to obtain an effective correction.

[0041] The control parameters of the effective correction amount are updated using a progressive correction method, with each update amplitude not exceeding 5% of the current parameter value. The converter control parameters are updated every 5ms, the pitch control parameters are updated every 50ms, and the excitation control parameters are updated every 20ms. The updated parameters are verified for stability to ensure that the parameter update does not cause new oscillations. The progressive correction parameters are processed comprehensively to ensure the coordination between the parameters of each controller and form a complete set of correction control parameters. The correction control parameter set is distributed to each execution controller through a distributed control network, using CAN bus communication with a communication rate of 1Mbps. After each controller receives the correction instruction, it uses a synchronous trigger method to update the parameters. The trigger signal is uniformly distributed by the main controller to ensure that the parameter update actions of each controller are executed synchronously to complete the real-time control process of the entire wind power grid-connected system.

[0042] Take a 1.5MW doubly-fed generator set as an example: when a 15Hz subsynchronous oscillation is detected, the d-axis current setting value of the converter controller is -45A, the actual response value is -40A, and the deviation is 11.1%, which exceeds the allowable range of 5%; at the same time, the q-axis current setting value is 25A, and the actual response value is 22A, with a deviation of 12%. The performance evaluation shows that the adjustment time is 180ms and the overshoot is 15%, both of which fail to meet the control index requirements. According to the evaluation results, the correction amount of the converter proportional gain is calculated to be +0.15 (15% of the original value), and the actual correction amount is determined to be +0.12 after limiting processing. The progressive correction method is adopted to increase the proportional gain by 0.02 each time, and the correction is completed in 6 times. The corrected control parameters are sent to the converter controller through the CAN bus at a rate of 1Mbps. At the same time, the pitch controller and the excitation controller also synchronously complete the update of their respective parameters. After parameter correction, the d-axis current response deviation is reduced to 3%, the q-axis current response deviation is reduced to 4%, the adjustment time is shortened to 150ms, and the overshoot is reduced to 8%, meeting the control performance requirements.

[0043] The above describes the safe operation method of the wind power grid-connected system for subsynchronous oscillation in the embodiment of the present application. The following describes the safe operation system of the wind power grid-connected system for subsynchronous oscillation in the embodiment of the present application. Figure 2 In the embodiment of the present application, an embodiment of a safe operation system for a wind power grid-connected system for subsynchronous oscillation includes: The acquisition module is used to perform multi-source data acquisition and processing on the electrical parameters and mechanical parameters of the wind power grid-connected system to obtain the original data set; An extraction module is used to perform time-frequency feature extraction processing on the original data set to obtain oscillation mode feature quantities; An analysis module is used to perform stability analysis on the oscillation modal characteristic quantity to obtain a subsynchronous oscillation risk warning signal; A decoupling module, used for performing multi-time scale decoupling processing on the sub-synchronous oscillation risk warning signal to obtain a mechanical-electrical coordinated control instruction; The decomposition module is used to perform task decomposition processing on the coordinated control instructions to obtain the real-time control quantity of each actuator; The correction module is used to evaluate and correct the execution effect of the real-time control quantity of each actuator to obtain a corrected control parameter set, and to control the wind power grid-connected system in real time through the corrected control parameter set.

[0044] Through the coordinated cooperation of the above-mentioned components, by collecting and processing the electrical and mechanical parameters of the wind power grid-connected system through multi-source data, the system operation status information can be fully obtained, laying a data foundation for the subsequent oscillation feature extraction and analysis; the time-frequency feature extraction and processing of the original data set can accurately capture the oscillation characteristics of the system in both time domain and frequency domain, and improve the accuracy of oscillation mode identification; the stability analysis and processing of the oscillation mode feature quantity can timely discover the potential unstable factors in the system and realize the early warning of subsynchronous oscillation; by performing multi-time scale decoupling processing on the subsynchronous oscillation risk warning signal, the oscillation characteristics of different frequency bands in the system can be separated. The swing components can be identified and control strategies can be formulated in a targeted manner to improve the accuracy of control; the coordinated control instructions can be decomposed into tasks to reasonably allocate control tasks to various actuators to ensure the executability of control instructions; the execution effect of the real-time control quantity of each actuator can be evaluated, processed and corrected, and the control parameters can be dynamically adjusted to continuously improve the control performance; the entire control process forms a closed-loop adaptive control system, which not only takes into account the coupling characteristics of the electrical system and the mechanical system, but also realizes coordinated control on multiple time scales. At the same time, the real-time evaluation and correction mechanism is used to ensure the continuous optimization of the control effect, thereby effectively improving the operating reliability and stability of the wind power grid-connected system.

[0045] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for safe operation of a wind power grid-connected system for subsynchronous oscillation, characterized in that: The wind power grid-connected system safe operation method for subsynchronous oscillation includes: Perform multi-source data collection and processing on the electrical and mechanical parameters of the wind power grid-connected system to obtain the original data set; Performing time-frequency feature extraction processing on the original data set to obtain oscillation mode feature quantities; Performing stability analysis on the oscillation modal characteristic quantity to obtain a subsynchronous oscillation risk warning signal; Performing multi-time scale decoupling processing on the subsynchronous oscillation risk warning signal to obtain a mechanical-electrical coordinated control instruction; Decompose the coordinated control instructions to obtain the real-time control quantity of each actuator; The execution effect of the real-time control quantity of each actuator is evaluated, processed and corrected to obtain a corrected control parameter set, and the wind power grid-connected system is controlled in real time through the corrected control parameter set.

2. The method for safe operation of a wind power grid-connected system against subsynchronous oscillation according to claim 1, characterized in that: The electrical parameters and mechanical parameters of the wind power grid-connected system are collected and processed from multiple sources to obtain an original data set, including: The voltage signal and the current signal of the stator end of the wind power grid-connected system are synchronously sampled and processed to obtain the stator electrical parameter data, and the voltage signal and the current signal of the rotor end of the wind power grid-connected system are synchronously sampled and processed to obtain the rotor electrical parameter data; Performing digital filtering processing on the stator electrical parameter data and the rotor electrical parameter data to obtain filtered electrical signal data, and performing power parameter calculation processing on the filtered electrical signal data to obtain system real-time power data; Performing speed collection and processing on the wind turbine rotor speed and the generator rotor speed of the wind power grid-connected system to obtain mechanical speed data, and performing spectrum analysis and processing on the mechanical speed data to obtain speed spectrum characteristic data; Performing vibration characteristic acquisition processing on the transmission shaft vibration signal of the wind power grid-connected system to obtain original vibration data, and performing noise elimination processing on the original vibration data through wavelet transform to obtain vibration characteristic data after noise reduction; The system real-time power data, the speed spectrum characteristic data and the vibration characteristic data after noise reduction are subjected to data fusion processing to obtain the original data set.

3. The method for safe operation of a wind power grid-connected system against subsynchronous oscillation according to claim 1, characterized in that: The step of performing time-frequency feature extraction processing on the original data set to obtain oscillation mode feature quantities includes: Performing Fourier transform processing on the voltage signal and the current signal in the original data set to obtain electrical signal spectrum data, and performing harmonic component extraction processing on the electrical signal spectrum data to obtain electrical harmonic characteristic data; Performing wavelet decomposition processing on the speed signal in the original data set to obtain multi-scale speed characteristic data, and performing principal component analysis processing on the multi-scale speed characteristic data through singular value decomposition to obtain the main oscillation component of the speed; Performing empirical mode decomposition processing on the vibration signal in the original data set to obtain an intrinsic mode function set, and performing envelope analysis processing on the intrinsic mode function set to obtain vibration amplitude characteristic data; Performing cross-correlation analysis on the electrical harmonic characteristic data and the main oscillation component of the speed to obtain electro-mechanical coupling characteristic data, and performing feature fusion processing on the electro-mechanical coupling characteristic data and the vibration amplitude characteristic data to obtain multi-source oscillation characteristic data; The multi-source oscillation characteristic data is subjected to modal separation processing to obtain the oscillation modal characteristic quantity.

4. The method for safe operation of a wind power grid-connected system against subsynchronous oscillation according to claim 1, characterized in that: The performing stability analysis on the oscillation modal characteristic quantity to obtain a subsynchronous oscillation risk warning signal includes: Performing characteristic root calculation processing on the oscillation modal characteristic quantity to obtain a system characteristic value set, and performing damping ratio calculation processing on the system characteristic value set to obtain modal damping ratio data; Performing critical stability judgment processing on the modal damping ratio data to obtain oscillation instability degree data, and performing energy distribution analysis processing on the oscillation instability degree data to obtain oscillation energy distribution characteristics; Performing frequency band decomposition processing on the oscillation energy distribution characteristics to obtain oscillation intensity data of each frequency band, and performing threshold comparison processing on the oscillation intensity data of each frequency band to obtain frequency band warning level data; Performing risk assessment processing on the frequency band warning level data to obtain oscillation risk level data, and performing probability update processing on the oscillation risk level data through Bayesian reasoning to obtain risk probability distribution data; A comprehensive risk analysis is performed on the risk probability distribution data to obtain the sub-synchronous oscillation risk warning signal.

5. The method for safe operation of a wind power grid-connected system against subsynchronous oscillation according to claim 1, characterized in that: The multi-time scale decoupling process is performed on the sub-synchronous oscillation risk warning signal to obtain a mechanical-electrical coordinated control instruction, including: Performing time-scale layered processing on the subsynchronous oscillation risk warning signal to obtain a fast oscillation component and a slow oscillation component, and performing electrical feature extraction processing on the fast oscillation component to obtain an electrical subsystem control demand; Performing mechanical feature extraction processing on the slow oscillation component to obtain a mechanical subsystem control requirement, and performing coupling analysis processing on the electrical subsystem control requirement and the mechanical subsystem control requirement to obtain system coupling feature data; Performing control target decomposition processing on the system coupling characteristic data to obtain a control target set of each subsystem, and performing coordination constraint processing on the control target set of each subsystem to obtain a coordinated control constraint condition; Performing control sequence generation processing on the coordinated control constraint condition to obtain a reference control sequence, and performing dynamic compensation processing on the reference control sequence to obtain a compensation control amount; The compensation control amount is subjected to coordinated control comprehensive processing to obtain the mechanical-electrical coordinated control instruction.

6. The method for safe operation of a wind power grid-connected system against subsynchronous oscillation according to claim 1, characterized in that: The coordinated control instructions are task-decomposed to obtain the real-time control quantity of each actuator, including: Performing functional hierarchical division processing on the mechanical-electrical coordinated control instructions to obtain main control layer instructions and device layer instructions, and performing control task parsing processing on the main control layer instructions to obtain a key control parameter set; Performing execution unit allocation processing on the device layer instruction to obtain control sequences of each execution unit, and performing timing coordination processing on the control sequences of each execution unit to obtain control timing data; Performing control amount calculation processing on the key control parameter set to obtain basic control data, and performing matching processing on the basic control data and the control timing data to obtain an actuator control scheme; The control scheme of the actuator is subjected to control component extraction processing to obtain a set of candidate control quantities, and the set of candidate control quantities is subjected to constraint verification processing to obtain the real-time control quantities of each actuator.

7. The method for safe operation of a wind power grid-connected system against subsynchronous oscillation according to claim 1, characterized in that: The evaluating and correcting the execution effect of the real-time control amount of each actuator to obtain a corrected control parameter set, and controlling the wind power grid-connected system in real time by using the corrected control parameter set, includes: Performing deviation calculation processing on the response data of the real-time control amount of each actuator to obtain control deviation data, and performing evaluation index calculation processing on the control deviation data to obtain a performance evaluation result; Performing control parameter correction amount calculation processing on the performance evaluation result to obtain a parameter correction amount, and performing amplitude limiting processing on the parameter correction amount to obtain an effective correction amount; Performing control parameter update processing on the effective correction amount to obtain a progressive correction parameter, and performing comprehensive processing on the progressive correction parameter to obtain the correction control parameter set; The modified control parameter set is distributedly distributed to obtain the modified instructions of each controller, and the modified instructions of each controller are synchronously issued to complete the real-time control of the wind power grid-connected system.

8. A wind power grid-connected system safety operation system for subsynchronous oscillation, used to implement the wind power grid-connected system safety operation method for subsynchronous oscillation according to any one of claims 1 to 7, characterized in that: The wind power grid-connected system safety operation system for subsynchronous oscillation includes: The acquisition module is used to perform multi-source data acquisition and processing on the electrical parameters and mechanical parameters of the wind power grid-connected system to obtain the original data set; An extraction module is used to perform time-frequency feature extraction processing on the original data set to obtain oscillation mode feature quantities; An analysis module is used to perform stability analysis on the oscillation modal characteristic quantity to obtain a subsynchronous oscillation risk warning signal; A decoupling module, used for performing multi-time scale decoupling processing on the sub-synchronous oscillation risk warning signal to obtain a mechanical-electrical coordinated control instruction; The decomposition module is used to perform task decomposition processing on the coordinated control instructions to obtain the real-time control quantity of each actuator; The correction module is used to evaluate and correct the execution effect of the real-time control quantity of each actuator to obtain a corrected control parameter set, and to control the wind power grid-connected system in real time through the corrected control parameter set.

Images