Large-interference stability analysis method and system for offshore wind power electromechanical frequency conversion grid connection

Abstract

The invention relates to the technical field of motor frequency conversion stability control, in particular to a large-interference stability analysis method and system for offshore wind power electromechanical frequency conversion grid connection, and the method comprises the steps: building a nonlinear dynamic model of a motor and a generator set frequency conversion system; performing coordinate translation by taking a steady-state working point of the nonlinear dynamic model as a coordinate origin, and dividing an operation characteristic space into fuzzy subspaces; carrying out local linearization on the model in each fuzzy subspace, and constructing a TS fuzzy model set; constructing a corresponding quadratic Lyapunov function based on the TS fuzzy model set, simultaneously solving a maximum attraction domain boundary, and performing projection to generate a two-dimensional stable operation region graph; and according to the maximum state variable boundary in the graph, adjusting an output power instruction of an offshore wind power system controller to maintain the stability of the system. According to the invention, the problem in the prior art that the system is unstable and the equipment is damaged when the frequency conversion system of the rotating unit is subjected to high-power disturbance is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of motor frequency conversion stability control technology, and in particular to a method and system for analyzing the stability of offshore wind turbines connected to the grid under large disturbances. Background Technology

[0002] In the field of frequency conversion stability control technology for offshore wind power, the core energy conversion link is the low-frequency power transmission of large rotating turbines through frequency converters. This technology achieves dynamic balance of energy transmission by adjusting the motor speed using a frequency converter. Its core challenge lies in the sharp increase in the nonlinear dynamic characteristics of the system when offshore wind farms are subjected to large-power sudden disturbances such as typhoon surges.

[0003] Existing technologies primarily employ linearized approximation models to predict stability boundaries or rely on empirical thresholds to set protection parameters. While capable of handling routine operating conditions, these methods inherently neglect the strong nonlinearity of electromechanical coupling effects in rotating systems and the spatiotemporal correlation characteristics of disturbances, leading to catastrophic misjudgments of critical instability boundaries. Furthermore, existing control models cannot accurately quantify the dynamic stability limits of a system subjected to instantaneous surge impacts, causing frequency converters to frequently trigger false protection or experience protection delays under real-world extreme marine conditions. This deficiency directly results in catastrophic shutdowns of offshore wind power systems, such as full DC-link collapse and overvoltage burnout of power devices, posing significant safety hazards and causing large-scale equipment damage. The fundamental problem lies in the fact that existing analytical methods cannot characterize the energy interaction details between high-frequency disturbances and the unit's mechanical inertia, nor do they possess the mathematical closure description capability for the precise boundaries of the stability domain of multivariable coupled random disturbances.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the general background of this disclosure and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a method and system for analyzing the stability of offshore wind turbines connected to the grid under large disturbances, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A method for large-disturbance stability analysis of offshore wind turbine frequency converter grid connection, the method comprising: A 26-dimensional nonlinear dynamic model of the frequency conversion system of motor and generator set is established. The nonlinear dynamic model includes at least the d-axis current and q-axis current of the motor, the subtransient electromotive force of the generator and the power angle state variable. Using the steady-state operating point of the nonlinear dynamic model as the origin of the coordinate system, coordinate translation is performed. The inverter side d-axis current, converter DC bus voltage, motor stator winding d-axis current, and generator stator winding d-axis current are selected as partitioning parameters to divide the operating feature space into 16 fuzzy subspaces. The nonlinear dynamic model is locally linearized in each of the fuzzy subspaces to construct a TS fuzzy model set, which is associated with each local linear model through a weighting function. Based on the TS fuzzy model set, a corresponding quadratic Lyapunov function is constructed. The range of nonlinear term perturbation is expanded in each round. The maximum perturbation boundary is searched based on the iterative solution method to solve for the maximum attraction domain boundary of the system. The boundary of the maximum attraction domain is projected onto the electromechanical power plane to generate a two-dimensional stable operating region map; Based on the maximum state variable boundary in the two-dimensional stable operating region diagram, the output power command of the offshore wind power system controller is adjusted to maintain system stability.

[0007] Furthermore, the nonlinear dynamic model is configured to establish a continuous energy conversion dynamic path from the absorption of low-frequency electrical energy by the motor to generate mechanical torque, the conversion of the mechanical torque into power frequency electrical energy by the generator, and the buffering of power fluctuations on both sides of the DC link of the converter.

[0008] Furthermore, the operating feature space is divided into 16 fuzzy subspaces, including: Based on the positive and negative dimensions of the inverter-side d-axis current obtained after coordinate translation, the first type of classification reference region is divided into the operating characteristic space. The DC bus voltage of the converter is correlated with the positive and negative dimensions of the operating characteristic space to generate a second type of dividing reference region; The third type of dividing reference region is formed by combining the positive and negative dimensions of the d-axis current of the motor stator winding in the operating characteristic space; The fourth type of dividing reference region is constructed by linking the d-axis current of the generator stator winding with the positive and negative dimensions of the operating characteristic space; The positive and negative subspaces of the first type of partitioning reference region, the positive and negative subspaces of the second type of partitioning reference region, the positive and negative subspaces of the third type of partitioning reference region, and the positive and negative subspaces of the fourth type of partitioning reference region are combined by Cartesian product. Based on the traversal mapping of all logical branches in the Cartesian product combination, 16 fuzzy subspaces are defined in the running feature space; Complete region segmentation is performed on the coordinate-translated running feature space based on the fuzzy subspace.

[0009] Furthermore, a TS fuzzy model set is constructed, including: Within each of the fuzzy subspaces formed by dividing the operating feature space, the extreme values ​​of the division parameters are substituted into the nonlinear dynamic model to generate a linear state-space equation that matches the corresponding fuzzy subspace. The linear state-space equations are bound to the corresponding fuzzy subspaces to construct a set of linear models corresponding to all the fuzzy subspaces; For each of the local linear models in the set of linear models, a rule is designed to activate the weight function based on the degree of change of the system dynamic parameters in the corresponding fuzzy subspace; Based on the rules, the output vectors of the linear model set are dynamically aggregated using the activation weight function to generate a continuous TS fuzzy model set covering the entire operating feature space.

[0010] Furthermore, the simultaneous solution of the quadratic Lyapunov function satisfying the linear matrix inequality (LMI) constraint includes: For all the local linear models in the TS fuzzy model set, a set of matrix inequality constraints reflecting the energy state of the system is constructed based on Lyapunov theory. The set of matrix inequality constraints includes positive definite conditions that guarantee the asymptotic stability of the system and state derivative decay constraints. By combining the correlation between the rule activation weight function and the local linear model, a dynamic coupling relationship is constructed between the matrix inequality constraint sets. The dynamic coupling relationship reflects the transition coordination conditions of the system's dynamic behavior under different fuzzy subspaces. Based on the continuity of the dynamic coupling relationship in all the operating feature spaces, the matrix inequality constraint set corresponding to all the local linear models in the TS fuzzy model set is solved simultaneously to obtain a unified quadratic coefficient matrix. Based on the unified quadratic coefficient matrix, an analytical expression of the quadratic Lyapunov function that satisfies all linear matrix inequalities (LMI) constraints is synthesized. The perturbation range parameters of the nonlinear dynamic model are expanded iteratively in rounds according to a preset step size, and the simultaneous solution process of the matrix inequality constraint set and the quadratic coefficient matrix is ​​restarted until a feasible solution cannot be obtained in the current round, at which point the iteration process is terminated. The boundary defined by the disturbance range parameter corresponding to the last successfully solved round is determined as the boundary of the maximum attraction domain of the nonlinear dynamic model.

[0011] Furthermore, a two-dimensional stable operating region map is generated, including: Obtain the geometric parameters of the closed hypersurface formed by the boundary of the maximum attraction domain defined by the quadratic Lyapunov function in the 26-dimensional state space; The geometric parameters of the closed hypersurface are reduced in dimension along the direction dominated by electromechanical coupling characteristics and mapped to an orthogonal coordinate axis composed of the d-axis current of the inverter side and the DC voltage coordinate axis of the converter, as well as the d-axis current of the motor stator winding and the d-axis current coordinate axis of the generator stator winding. A convex envelope operation is performed on the mapped boundary point set to form an envelope boundary curve. The region enclosed by the envelope boundary curve represents the operating range of the system that maintains transient stability under power disturbance. Using the envelope boundary curve as the critical surface, a rectangular stable operating region map including the upper limit of active power and the upper limit of reactive power is generated. The horizontal axis of the rectangular stable operating region map is the active power axis, and the vertical axis is the reactive power axis.

[0012] Furthermore, adjust the output power command, including: Real-time monitoring of the actual output active and reactive power of the offshore wind power system at the real-time operating point coordinates in the current electromechanical power plane; By comparing the real-time operating point coordinates with the boundary position relationship of the two-dimensional stable operating area map, when it is detected that the real-time operating point is approaching the boundary of the maximum state variable of the two-dimensional stable operating area map, a power limiting protection action is triggered. The intersection of the tangential trajectory of the real-time running point in the direction exceeding the boundary of the maximum state variable and the boundary of the two-dimensional stable running area map is calculated as the target safe power point; The active power value and reactive power value corresponding to the target safe power point are converted into a power command format that can be executed by the offshore wind power system controller; The output power command is dynamically modified according to the command interface of the offshore wind power system controller, and the power operating point of the offshore wind power system is constrained within the boundary of the two-dimensional stable operating area map.

[0013] Furthermore, dynamic rate optimization is performed on the output power command, including: The rate of change of the power slope at the real-time operating point in the electromechanical power plane is monitored, and the rate of change of the power slope is calculated based on the derivative data of the rotational speed deviation after coordinate translation on the time series. When the power limiting protection action is triggered, the power command adjustment rate parameter is selected from the adjustment strategy library pre-stored in the offshore wind power system controller according to the absolute value of the power slope change rate.

[0014] A large-disturbance stability analysis system for offshore wind turbine frequency converter grid connection, the system comprising: The model building module establishes a 26-dimensional nonlinear dynamic model of the frequency conversion system of the motor and generator set. The spatial partitioning module uses the steady-state operating point of the nonlinear dynamic model as the origin of the coordinate system to perform coordinate translation, dividing the running feature space into 16 fuzzy subspaces. The local linear module performs local linearization on the nonlinear dynamic model in each fuzzy subspace, constructs the TS fuzzy model set, and the TS fuzzy model set is associated with each local linear model through a weighting function; The constraint solving module, based on the TS fuzzy model set, constructs the corresponding quadratic Lyapunov function, expands the range of nonlinear term perturbation in each round, and searches for the maximum perturbation boundary based on iterative solution to solve the maximum attraction domain boundary of the system; The image generation module projects the boundary of the maximum attraction domain onto the electromechanical power plane to generate a two-dimensional stable operating region map; The command stabilization module adjusts the output power command of the offshore wind power system controller to maintain system stability based on the maximum state variable boundary in the two-dimensional stable operating area diagram.

[0015] Furthermore, the image generation module includes: The parameter acquisition unit acquires the geometric parameters of the closed hypersurface formed by the boundary of the maximum attraction domain defined by the quadratic Lyapunov function in the 26-dimensional state space. The dimension reduction mapping unit reduces the geometric parameters of the closed hypersurface along the dominant direction of electromechanical coupling characteristics and maps them to an orthogonal coordinate axis composed of the inverter side d-axis current and converter DC voltage coordinate axis, as well as the motor stator winding d-axis current and generator stator winding d-axis current coordinate axis. The envelope curve unit performs convex envelope operations on the mapped set of boundary points to form an envelope boundary curve. The region enclosed by the envelope boundary curve represents the operating range of the system that maintains transient stability under power disturbances. The image generation unit uses the envelope boundary curve as the critical surface to generate a rectangular stable operating region map that includes the upper limit values ​​of active power and reactive power. The horizontal axis of the rectangular stable operating region map is the active power axis, and the vertical axis is the reactive power axis.

[0016] The technical solution of this invention can achieve the following technical effects: By establishing a 26-dimensional nonlinear dynamic model of the variable frequency system of motors and generator sets, and combining TS fuzzy modeling with 16 fuzzy subspace partitions and LMI solution driven by Lyapunov theory, the problem that existing technologies cannot accurately quantify the stable operating boundary when the variable frequency system of rotating units is subjected to high-power disturbances, such as typhoon surges, leading to system instability and equipment damage is solved.

[0017] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a large-disturbance stability analysis method for offshore wind turbine frequency converter grid connection; Figure 2 This is a topology diagram of an offshore wind farm based on frequency conversion of electric motors and generator sets; Figure 3 A flowchart illustrating the process of dividing a fuzzy subspace; Figure 4 A flowchart illustrating the process of constructing a TS fuzzy model set; Figure 5 A flowchart illustrating the process of establishing a quadratic Lyapunov function. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0022] Example 1; like Figure 1 and Figure 2 As shown, this application provides a large-disturbance stability analysis method for offshore wind turbine frequency converter grid connection, the method including: S10: Establish a 26-dimensional nonlinear dynamic model of the frequency conversion system of the motor and generator set. The nonlinear dynamic model shall include at least the d-axis current and q-axis current of the motor, the subtransient electromotive force of the generator and the power angle state variables. S20: Using the steady-state operating point of the nonlinear dynamic model as the origin of the coordinate system, the coordinates are translated. The inverter side d-axis current, converter DC bus voltage, motor stator winding d-axis current, and generator stator winding d-axis current are selected as partitioning parameters to divide the operating feature space into 16 fuzzy subspaces. S30: Locally linearize the nonlinear dynamic model in each fuzzy subspace, construct the TS fuzzy model set, and associate each local linear model with the TS fuzzy model set through a weighting function; S40: Based on the TS fuzzy model set, construct the corresponding quadratic Lyapunov function, expand the range of nonlinear term perturbation in each round, and search for the maximum perturbation boundary based on iterative solution to solve the maximum attraction domain boundary of the system. S50: Project the boundary of the maximum attraction domain onto the electromechanical power plane to generate a two-dimensional stable operating region map; S60: Adjust the output power command of the offshore wind power system controller to maintain system stability based on the maximum state variable boundary in the two-dimensional stable operation area diagram.

[0023] Specifically, firstly, based on the physical parameters and manufacturer data of the motor, generator set, and their frequency converter, a complete 26-dimensional nonlinear dynamic model is established. The model focuses on the motor's d-axis and q-axis currents, the generator's subtransient electromotive force, and power angle as core states. It describes the coupling evolution of the frequency converter under disturbances using electromagnetic induction relationships, stator circuit dynamics, rotor and shaft mechanical oscillation equations, excitation system dynamics, and the averaged switching behavior of the converter. During model establishment, it is preferable to retain both electromagnetic and mechanical coupling terms, and key time constants and coupling coefficients are calibrated through parameter identification and small-signal experiments to ensure model accuracy. Secondly, the system is selected under typical grid-connected operating conditions. Using the steady-state operating point as the coordinate translation reference, coordinate translation and normalization are achieved by subtracting the steady-state values ​​from all state variables to ensure numerical stability for subsequent local linearization and fuzzy partitioning. Based on this, four parameters—inverter-side d-axis current, converter DC bus voltage, motor stator winding d-axis current, and generator stator winding d-axis current—are used as partitioning references. Binary partitioning based on their positive and negative signs yields 16 Cartesian combination sub-regions. To ensure smooth transitions between sub-regions while maintaining discriminative power, overlapping membership functions are preferred to define the positive and negative sub-regions for each dimension, thus forming 16 fuzzy subspaces. Subsequently, within each fuzzy subspace, the subspace... Local operating points between sub-models yield locally linear state-space descriptions of the original nonlinear model. Local linearization can be achieved by analytical differentiation or numerical difference to obtain the local state matrix and input matrix. These locally linear descriptions are then bound to their corresponding membership functions and dynamically combined using weighted rules with normalized weights to obtain a continuous TS fuzzy model set covering the entire operating feature space. Based on this TS model set, a unified solution process for the quadratic Lyapunov energy function is constructed: positive definite energy function candidates are set at each sub-model, and corresponding matrix inequality constraints are derived based on the Lyapunov stability criterion. Membership weights are used to establish coupling constraints between sub-models, thus forming a... A complete family of linear matrix inequalities is used to solve these LMI constraints using a positive semidefinite programming solver, obtaining a quadratic coefficient matrix with the derivative of the enabling function negative definite in each subspace. To determine the actual withstand boundary of the system under nonlinear disturbances, this embodiment adopts an iterative strategy to expand the disturbance amplitude. Starting with an initial small-amplitude nonlinear term disturbance, the disturbance amplitude is increased by a predetermined step size in each round. In each round, all LMI constraints are solved again simultaneously until the solution becomes infeasible or the change in the obtained attraction domain boundary is lower than a threshold, for example, the boundary change is less than 1%. At the end of the iteration, the maximum permissible disturbance boundary is recorded, and the maximum attraction domain boundary of the system is obtained from this.For the obtained high-dimensional boundary, it is preferable to perform dense sampling of the Lyapunov function level set and boundary samples in the high-dimensional state space, mapping each high-dimensional state vector to the corresponding active and reactive power output of the motor, and performing dimensionality reduction mapping on the boundary point set along the electromechanical coupling dominant direction or by directly selecting the two coordinate components of the power. The convex hull operation is performed on the mapped two-dimensional point set to obtain the envelope boundary curve. Based on the curve, the critical values ​​of active and reactive power are extracted at the intersection of the horizontal and vertical coordinates on the electromechanical power plane, and a rectangular stable operating region map parallel to the coordinate axes is generated. The four sides of the rectangle are taken as the upper limit value under the required conservative margin to facilitate the implementation of the controller interface. Finally, the maximum state variable boundary in the two-dimensional stable operating region diagram is transformed into a power command strategy that the offshore wind power system controller can execute: During grid-connected operation, the monitoring layer monitors the positional relationship between the current active or reactive power point and the stable region in real time at a specified sampling frequency. When the operating point approaches or exceeds the safety boundary, the monitoring layer calculates the target safe power point and sends the active and reactive power limiting and rate constraints to the converter controller by adapting the command format with the frequency converter. After receiving the new command, the converter smoothly tracks according to its closed-loop control logic and avoids mechanical shock or electrical transient oscillation by rate shaping.

[0024] The technical solution of this invention establishes a 26-dimensional nonlinear dynamic model of the frequency conversion system of motor and generator set. Combining TS fuzzy modeling with 16 fuzzy subspaces and LMI solution driven by Lyapunov theory, it solves the problem that the existing technology cannot accurately quantify the stable operating boundary when the frequency conversion system of rotating unit is subjected to high-power disturbances, such as typhoon surges, which leads to system instability and equipment damage.

[0025] Furthermore, the nonlinear dynamic model is configured to establish a continuous energy conversion dynamic path from the absorption of low-frequency electrical energy by the motor to generate mechanical torque, the conversion of mechanical torque into power frequency electrical energy by the generator, and the buffering of power fluctuations on both sides of the DC link of the converter.

[0026] As a preferred embodiment of the above, firstly, the electric motor, as the initial stage of energy conversion, absorbs low-frequency electrical energy through its stator windings, drives the rotor, and generates mechanical torque. This process needs to consider the dynamic response of the excitation current within the motor and the influence of the surrounding magnetic flux to ensure that the motor maintains high efficiency and stable output. The state variable model of the motor details the current, electromagnetic torque, and related dynamic response characteristics. Next, the generator, as the intermediate stage of energy conversion, converts the absorbed mechanical torque into power frequency electrical energy. The dynamic response of the generator is described based on the changes in speed, torque, and transient magnetic field components to ensure that it can quickly respond to changing mechanical inputs at its high-efficiency operating point. The state model of the generator is located on the d-axis. The q-axis electromotive force is fully described, enabling the system to maintain stable power output around the clock. In particular, the adjustment of rotor damping and excitation voltage ensures the adaptability of transient electromotive force and stable operation of the system throughout the process. In the continuous energy conversion path, the converter is responsible for ensuring stable bidirectional power flow. The DC link of the converter is set to realize instantaneous energy buffering and smooth out fluctuations that may occur during power transmission between the motor and the generator. Specifically, through real-time monitoring of DC bus voltage and current, the converter is equipped with multi-level filtering and active control strategies to ensure that the entire conversion path can flexibly adjust its operating state under changing load conditions to support the stable output of the entire offshore wind power system.

[0027] Furthermore, such as Figure 3 As shown, the running feature space is divided into 16 fuzzy subspaces, including: Based on the positive and negative dimensions of the inverter-side d-axis current in the operating characteristic space obtained after coordinate translation, the regions are divided into the first category of reference regions. The DC bus voltage of the associated converter generates a second type of reference region based on the positive and negative dimensions of the operating characteristic space; A third type of reference region is formed by combining the positive and negative dimensions of the d-axis current of the motor stator winding in the operating characteristic space; The positive and negative dimensions of the d-axis current of the stator winding of the linkage generator in the operating characteristic space are used to construct a fourth type of reference region. The positive and negative subspaces of the first, second, third, and fourth categories of the datum region are combined by Cartesian product. Based on the traversal mapping of all logical branches in the Cartesian product combination, 16 fuzzy subspaces are defined in the running feature space; Complete region segmentation is performed on the coordinate-translated running feature space based on the fuzzy subspace.

[0028] As a preferred embodiment of the above, firstly, the inverter-side d-axis current obtained from the coordinate translation step is subdivided, with its positive and negative dimensions serving as the first type of division reference region. This region setting facilitates the identification of the inverter's power transmission state under different current directions, helping the system to identify the dynamic impact of the corresponding current. Simultaneously, this division basis provides strong support for performance optimization of the frequency conversion system in current vector control. Secondly, a second type of division reference region involving the converter's DC bus voltage is defined, also differentiated based on its positive and negative dimensions. Since the DC bus voltage reflects the instantaneous energy storage capacity of energy transmission, and changes in this voltage directly affect the stable operation of the inverter, this region division allows for better management of the effects of voltage fluctuations on the overall dynamic response of the system. The third type of division reference region is based on the positive and negative dimensions of the motor stator winding d-axis current. The current information on the motor side is designed to reflect changes in the motor's energy input. This type of division helps predict and adjust the motor's operating characteristics under different loads and control states, ensuring the motor's... The fourth category of classification involves the positive and negative dimensions of the generator stator winding d-axis current. This dimension provides feedback information for the generator in frequency conversion control, which can significantly improve the generator's ability to dynamically adapt to changes in grid conditions. By distinguishing the generator stator d-axis current, the synchronization of the power generation process and the stability of power output are improved. Then, in each of the above reference regions, the corresponding positive and negative subspaces are combined in sequence and arranged interactively according to the Cartesian product principle, finally forming 16 unique fuzzy subspaces. In these spaces, each subspace reflects the specialized dynamic response of the system under a specific state, ensuring that the control strategy has comprehensive and detailed adaptability when considering multiple state variables. By logically traversing and fully defining each fuzzy subspace generated by the Cartesian product combination, the full coverage of the feature space is achieved. This division not only ensures the reasonable handling of each operating state, but also makes the expression of the entire dynamic characteristics possible, supporting subsequent model analysis and the execution of control algorithms.

[0029] Furthermore, such as Figure 4 As shown, the TS fuzzy model set is constructed, including: Within each fuzzy subspace formed by dividing the feature space, the extreme values ​​of the partitioning parameters are substituted into the nonlinear dynamic model to generate a linear state-space equation that matches the corresponding fuzzy subspace. By binding the linear state-space equations with the corresponding fuzzy subspaces, a set of linear models corresponding to all fuzzy subspaces is constructed. For each local linear model in the set of linear models, design rules to activate the weight function based on the degree of change of the system dynamic parameters in the corresponding fuzzy subspace; The output vectors of the linear model set are dynamically aggregated based on the rule-based activation weight function to generate a continuous TS fuzzy model set covering the entire running feature space.

[0030] As a preferred embodiment of the above, after the running feature space is divided into multiple fuzzy subspaces, this implementation selects the local operating point of the nonlinear dynamic model in each subspace for detailed mathematical processing, transforming the complex nonlinear dynamic behavior into local linear state-space equations. The core of this process lies in utilizing the local dynamic characteristics of the system to decompose the nonlinear problem into an easily tractable linear problem, ensuring that the dynamic change characteristics in each subspace can be accurately described. After completing the construction of the linear state-space equations, this implementation binds them to the corresponding fuzzy subspaces to form a set of linear models adapted to the characteristics of each subspace. This step is performed by identifying the specific dynamic patterns of each subspace, enabling the linear model to directly respond to state changes within the subspace, thereby enhancing the sensitivity and accuracy of the overall model. Next, this implementation designs rules to activate weight functions based on the degree of change of the system dynamic parameters in each fuzzy subspace. These functions are used to dynamically evaluate the operation of the state parameters in each subspace. The system dynamically adjusts the current state of each subspace and activates the corresponding linear model. This weighting function, through a weighted dynamic adaptive mechanism, allows the current state of the subspace to be reflected in a wider range of model responses in real time, improving the control system's instantaneous adjustment capability and response speed. Finally, the implementation method activates the weighting function based on the rules designed above, dynamically aggregates the linear model set, and generates a continuous TS fuzzy model set by integrating some model output vectors. This fuzzy model set covers the entire operating feature space and can effectively fill the dynamic response gaps between subspaces, ensuring that the system's stability can be maintained and optimized even under maximum disturbance. In a specific example, when a wind farm is affected by a sudden strong wind, the TS fuzzy model set can quickly determine the state characteristics of each fuzzy subspace and automatically adjust the control strategy in a timely manner through its dynamic weight adjustment mechanism. For example, when the power angle deviation or current deviation changes, the model set can optimize the power command, thereby ensuring the continuous and stable operation of the wind turbine.

[0031] Furthermore, such as Figure 5 As shown, the simultaneous solution of the quadratic Lyapunov function satisfying the linear matrix inequality (LMI) constraint includes: For all local linear models in the TS fuzzy model set, a set of matrix inequality constraints reflecting the energy state of the system is constructed based on Lyapunov theory. The set of matrix inequality constraints includes positive definite conditions that guarantee the asymptotic stability of the system and state derivative decay constraints. By combining the correlation between the rule-activated weight function and the local linear model, a dynamic coupling relationship is constructed between the matrix inequality constraint sets. The dynamic coupling relationship reflects the transition coordination conditions of the system's dynamic behavior under different fuzzy subspaces. Based on the continuity of the dynamic coupling relationship in the entire operating feature space, the matrix inequality constraint set corresponding to all local linear models in the TS fuzzy model set is solved simultaneously to obtain a unified quadratic coefficient matrix. An analytical expression of a quadratic Lyapunov function that satisfies all linear matrix inequalities (LMI) is synthesized based on a unified quadratic coefficient matrix. The perturbation range parameters of the nonlinear dynamic model are expanded iteratively in rounds according to a preset step size, and the process of solving the matrix inequality constraint set and the quadratic coefficient matrix is ​​restarted until a feasible solution cannot be obtained in the current round, at which point the iteration process is terminated. The boundary defined by the perturbation range parameters corresponding to the last successfully solved round is determined as the boundary of the maximum attraction domain of the nonlinear dynamic model.

[0032] As a preferred embodiment of the above, firstly, in the TS fuzzy model set, a set of matrix inequality constraints is designed for each local linear model. These constraints focus on reflecting the energy state of the system and ensuring asymptotic stability under dynamic disturbances. Specifically, the positive definite condition guarantees the energy convergence characteristics of the system in steady state, while the state derivative decay constraint controls the stable convergence speed of the system in transition state. After constructing the constraint set, this embodiment combines rule-activated weight functions to establish dynamic coupling relationships between each local linear model. This relationship is defined by analyzing the changes in state parameters within each fuzzy subspace, reflecting the transition coordination conditions of the system between different subspaces and the dynamic coupling relationship. This approach makes the transitions between different states of the model smoother, ensuring that the dynamic behavior of the system under different fuzzy subspaces is consistent. Based on the above dynamic coupling relationship, this implementation analyzes and solves the overall continuity of the running feature space. By simultaneously solving the matrix inequality constraint set corresponding to all local linear models in the TS fuzzy model set, a unified quadratic coefficient matrix is ​​obtained. This unified matrix is ​​the result of integrating various local models, providing a stability verification scheme from a global perspective. Finally, relying on the unified quadratic coefficient matrix, this implementation synthesizes the analytical expression of the quadratic Lyapunov function that satisfies all LMI constraints. This Lyapunov function can not only comprehensively describe the system... The system's energy changes under various dynamic states can also provide specific stability assessment criteria. In practical applications, for example, when a wind farm faces power angle deviation caused by drastic weather changes, this function can effectively determine the system's stable state and guide timely adjustments to the control strategy. The disturbance range parameters of the nonlinear dynamic model are expanded iteratively in rounds with a preset step size. Starting from the initial conditions, a set of small-amplitude nonlinear disturbance parameters are set to simulate peak load changes that may be encountered in actual operation. When starting the iteration process, it is preferable to select a disturbance increment of 1% to 5% of the nominal value to ensure that the expansion of the disturbance range in each round can accurately capture dynamic characteristics under the requirements of system stability analysis. In each iteration, the strength of the nonlinear term is gradually increased based on the current state and parameters of the system, thereby expanding the perturbation range and testing the robustness and adaptability of the model. After adjusting the perturbation range in each iteration, the simultaneous solution process of the matrix inequality constraint set and the quadratic coefficient matrix is ​​immediately restarted. Through reloading in each round, the entire analysis process is ensured to reflect the stability of the system under different perturbation levels in a timely manner. This operation ensures that the quadratic coefficient matrix of the Lyapunov function and the system's state constraints are always in a dynamic update state, ensuring the optimal solution under each perturbation range. The solution process uses optimization algorithms, such as semidefinite programming solvers, to solve the linear matrix inequalities and capture the new boundaries after the perturbation.If a feasible solution cannot be obtained in the current iteration, i.e., the matrix inequality solver returns an infeasible state, the iteration process terminates. At this point, the system has exceeded its stability convergence capability, meaning the disturbance's impact has exceeded the system's automatic adjustment and energy absorption capacity. Finally, the boundary defined by the disturbance range parameter corresponding to the last successfully solved iteration is determined as the maximum attraction domain boundary of the nonlinear dynamic model. This boundary provides the maximum possible range of state changes the system can withstand, serving as a guideline for subsequent control and actual operation, ensuring that the motor and generator set can still maintain stable operation when facing large disturbances.

[0033] Furthermore, generating a two-dimensional stable operating region map includes: Obtain the geometric parameters of the closed hypersurface formed by the boundary of the maximum attraction domain defined by the quadratic Lyapunov function in the 26-dimensional state space; The geometric parameters of the closed hypersurface are reduced in dimension along the dominant direction of electromechanical coupling characteristics and mapped to an orthogonal coordinate axis composed of the inverter side d-axis current and converter DC voltage coordinate axis, as well as the motor stator winding d-axis current and generator stator winding d-axis current coordinate axis. A convex envelope operation is performed on the mapped boundary point set to form an envelope boundary curve. The region enclosed by the envelope boundary curve represents the operating range of the system that maintains transient stability under power disturbance. Using the envelope boundary curve as the critical surface, a rectangular stable operating region map is generated, which includes the upper limit values ​​of active power and reactive power. The horizontal axis of the rectangular stable operating region map is the active power axis, and the vertical axis is the reactive power axis.

[0034] As a preferred embodiment of the above, firstly, by obtaining the boundary of the maximum attraction domain defined by the quadratic Lyapunov function, the implementation method determines the geometric parameters of the closed hypersurface formed in the 26-dimensional state space from a global perspective. These parameters comprehensively describe the stability boundary of the system in the dynamic state. Through this global geometry, the energy state changes of the system under different states can be identified, providing a basis for subsequent dimensionality reduction mapping. The implementation method employs dimensionality reduction technology to reduce the parameters of the closed hypersurface along the dominant direction of electromechanical coupling characteristics, mapping them to an orthogonal coordinate axis formed by the inverter-side d-axis current and converter DC voltage coordinate axis, as well as the motor stator winding d-axis current and generator stator winding d-axis current coordinate axis. This dimensionality reduction process, focusing on the output characteristics of the motor, effectively simplifies high-dimensional complex data into an easily identifiable form in a two-dimensional coordinate space, ensuring that the main features of dynamic behavior can be directly represented in the two-dimensional space. The stability analysis is visually represented by performing convex envelope operations on the boundary point set obtained after dimensionality reduction to form an envelope boundary curve. The region enclosed by this curve directly represents the operating range in which the system maintains transient stability under power disturbances. This step ensures the accuracy of stability analysis, enabling the rapid determination of the system's safe operating range under different dynamic conditions. After the boundary curve is formed, a rectangular stable operating region map including the upper limits of active power and reactive power is generated using the envelope boundary curve as the critical surface. This rectangular map uses the horizontal axis as the active power axis and the vertical axis as the reactive power axis, intuitively reflecting the safe operating range under different power conditions. For example, in actual operation, when the power demand of the power grid fluctuates drastically due to weather changes, this stable operating region map can quickly identify the critical power range and provide guidance for system control. Through this visualization method, operators can quickly adjust the operating parameters of electromechanical equipment to ensure system stability.

[0035] Furthermore, adjusting the output power command includes: Real-time monitoring of the actual output active and reactive power of the offshore wind power system at the real-time operating point coordinates in the current electromechanical power plane; By comparing the real-time operating point coordinates with the boundary position of the two-dimensional stable operating area map, when the real-time operating point is detected to be approaching the boundary of the maximum state variable of the two-dimensional stable operating area map, the power limiting protection action is triggered. The intersection of the tangential trajectory of the real-time running point in the direction exceeding the maximum state variable boundary and the boundary of the two-dimensional stable running area map is calculated as the target safe power point; Convert the active power and reactive power values ​​corresponding to the target safe power point into a power command format that can be executed by the offshore wind power system controller; The output power command is dynamically corrected based on the command interface of the offshore wind power system controller, and the power operating point of the offshore wind power system is constrained within the boundary of the two-dimensional stable operating area map.

[0036] As a preferred embodiment of the above, firstly, the implementation method monitors the output power of the offshore wind power system in real time to obtain the actual output active and reactive power, and maps them onto the current electromechanical power plane. The identification of the real-time operating point coordinates provides a precise data foundation for dynamic control, ensuring that the system can adjust in a timely manner under rapidly changing power demands. When the real-time operating point gradually approaches the boundary of the two-dimensional stable operating area map, especially when it touches the maximum state variable boundary, the implementation method immediately triggers a power limiting protection action. This function ensures that the system does not exceed the safe power range by setting limits, effectively preventing potential instability caused by excessive power. Through the protection mechanism, once boundary approach is detected, the output power is promptly limited to a safe range. Next, the implementation method calculates the tangential trajectory of the current operating point in the direction exceeding the power boundary, and the intersection point with the boundary of the two-dimensional stable operating area map is taken as the target safe power point. The target safe power point serves as a new benchmark for the system during power adjustment, enabling rapid restoration of stable operation by minimizing power deviation. Once the target safe power point is determined, its corresponding active and reactive power values ​​are converted into a power command format executable by the offshore wind power system controller. This conversion process ensures accurate data exchange, laying the foundation for subsequent command execution. Finally, dynamic correction is performed based on the command interface of the offshore wind power system controller to ensure that the output power command is aligned with the target safe power point. Through this dynamic correction, the power operating point of the offshore wind power system is strictly constrained within the boundary of the two-dimensional stable operating area map, ensuring stable operation of the system under various load conditions. For example, in practical applications, when a power surge occurs, the implementation method can quickly monitor and adjust the power transmission status of the power grid, ensuring that the system can continue to maintain its operational stability when facing unforeseen natural disturbances such as strong winds.

[0037] Furthermore, dynamic rate optimization of the output power command includes: The rate of change of power slope at real-time operating points in the electromechanical power plane is monitored. The rate of change of power slope is calculated based on the derivative data of the rotational speed deviation after coordinate translation on the time series. When the power limiting protection action is triggered, the power command adjustment rate parameter is selected from the adjustment strategy library pre-stored in the offshore wind power system controller according to the absolute value of the power slope change rate.

[0038] As a preferred embodiment of the above, firstly, this implementation monitors the rate of change of the power slope at the real-time operating point in the electromechanical power plane. This rate of change is calculated based on the derivative data of the speed deviation over time. By continuously monitoring this key indicator, the system can control the instantaneous changes in the power state in real time, providing a precise basis for subsequent adjustments. When the power limiting protection action is triggered, the system selects a suitable power command adjustment rate parameter from the adjustment strategy library pre-stored in the offshore wind power system controller according to the absolute value of the rate of change of the power slope. The contents of the adjustment strategy library define corresponding linear adjustment rate curves according to different power slope intervals. This selection process ensures the specific adaptability of the adjustment strategy to the current state of the system and avoids the possibility of blind adjustment. The selected adjustment rate parameter is used in the closed-loop control logic of the offshore wind power system controller to dynamically correct the output power command. At this stage, not only is the real-time accuracy of the output command emphasized, but it is also necessary to suppress the electrical and mechanical shocks caused by abrupt changes. Therefore, the dynamic adjustment strategy adjusts the power in a gradual manner, so that the risks of mechanical oscillation and electrical transient oscillation between the motor and the generator set are effectively balanced. For example, if the output power demand increases sharply due to a sudden change in weather during operation, and the real-time monitored power slope is large, this implementation will automatically select a more moderate adjustment rate curve to ensure a smooth transition of power change, thereby making the adjustment process more stable and reducing the potential impact on the equipment.

[0039] Example 2; Based on the same inventive concept as the large disturbance stability analysis method for offshore wind turbine frequency converter grid connection in the foregoing embodiments, the present invention also provides a large disturbance stability analysis system for offshore wind turbine frequency converter grid connection, the system comprising: The model building module establishes a 26-dimensional nonlinear dynamic model of the frequency conversion system of the motor and generator set. The spatial partitioning module uses the steady-state operating point of the nonlinear dynamic model as the origin of the coordinate system to perform coordinate translation, dividing the running feature space into 16 fuzzy subspaces. The local linear module performs local linearization on the nonlinear dynamic model in each fuzzy subspace, constructs the TS fuzzy model set, and the TS fuzzy model set is associated with each local linear model through a weighting function; The constraint solving module, based on the TS fuzzy model set, constructs the corresponding quadratic Lyapunov function, expands the range of nonlinear term perturbation in each round, and searches for the maximum perturbation boundary based on iterative solution to solve the maximum attraction domain boundary of the system; The image generation module projects the boundary of the maximum attraction domain onto the electromechanical power plane to generate a two-dimensional stable operating region map; The command stabilization module adjusts the output power command of the offshore wind power system controller to maintain system stability based on the maximum state variable boundary in the two-dimensional stable operating area diagram.

[0040] The adjustment system described above in this invention can effectively realize a large-interference stability analysis method for offshore wind turbine frequency conversion grid connection, and the technical effects it can achieve are as described in the above embodiments, which will not be repeated here.

[0041] Furthermore, the image generation module includes: The parameter acquisition unit acquires the geometric parameters of the closed hypersurface formed by the boundary of the maximum attraction domain defined by the quadratic Lyapunov function in the 26-dimensional state space. The dimension reduction mapping unit reduces the geometric parameters of the closed hypersurface along the dominant direction of electromechanical coupling characteristics and maps them to an orthogonal coordinate axis composed of the inverter side d-axis current and converter DC voltage coordinate axis, as well as the motor stator winding d-axis current and generator stator winding d-axis current coordinate axis. The envelope curve unit performs convex envelope operations on the mapped set of boundary points to form an envelope boundary curve. The region enclosed by the envelope boundary curve represents the operating range of the system that maintains transient stability under power disturbances. The image generation unit uses the envelope boundary curve as the critical surface to generate a rectangular stable operating region map that includes the upper limit values ​​of active power and reactive power. The horizontal axis of the rectangular stable operating region map is the active power axis, and the vertical axis is the reactive power axis.

[0042] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0043] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for analyzing the stability of offshore wind turbines connected to the grid under large disturbances, characterized in that, The method includes: A 26-dimensional nonlinear dynamic model of the frequency conversion system of motor and generator set is established. The nonlinear dynamic model includes at least the d-axis current and q-axis current of the motor, the subtransient electromotive force of the generator and the power angle state variable. Using the steady-state operating point of the nonlinear dynamic model as the origin of the coordinate system, coordinate translation is performed. The inverter side d-axis current, converter DC bus voltage, motor stator winding d-axis current, and generator stator winding d-axis current are selected as partitioning parameters to divide the operating feature space into 16 fuzzy subspaces. The nonlinear dynamic model is locally linearized in each of the fuzzy subspaces to construct a TS fuzzy model set, which is associated with each local linear model through a weighting function. Based on the TS fuzzy model set, a corresponding quadratic Lyapunov function is constructed. The range of nonlinear term perturbation is expanded in each round. The maximum perturbation boundary is searched based on the iterative solution method to solve for the maximum attraction domain boundary of the system. The boundary of the maximum attraction domain is projected onto the electromechanical power plane to generate a two-dimensional stable operating region map; Based on the maximum state variable boundary in the two-dimensional stable operating region diagram, the output power command of the offshore wind power system controller is adjusted to maintain system stability.

2. The method for analyzing the large-disturbance stability of offshore wind turbine frequency converter grid connection according to claim 1, characterized in that, The nonlinear dynamic model is configured to establish a continuous energy conversion dynamic path from the absorption of low-frequency electrical energy by the motor to generate mechanical torque, the conversion of the mechanical torque into power frequency electrical energy by the generator, and the buffering of power fluctuations on both sides of the DC link of the converter.

3. The method for analyzing the large-disturbance stability of offshore wind turbine frequency converter grid connection according to claim 1, characterized in that, The running feature space is divided into 16 fuzzy subspaces, including: Based on the positive and negative dimensions of the inverter-side d-axis current obtained after coordinate translation, the first type of classification reference region is divided into the operating characteristic space. The DC bus voltage of the converter is correlated with the positive and negative dimensions of the operating characteristic space to generate a second type of dividing reference region; The third type of dividing reference region is formed by combining the positive and negative dimensions of the d-axis current of the motor stator winding in the operating characteristic space; The fourth type of dividing reference region is constructed by linking the d-axis current of the generator stator winding with the positive and negative dimensions of the operating characteristic space; The positive and negative subspaces of the first type of partitioning reference region, the positive and negative subspaces of the second type of partitioning reference region, the positive and negative subspaces of the third type of partitioning reference region, and the positive and negative subspaces of the fourth type of partitioning reference region are combined by Cartesian product. Based on the traversal mapping of all logical branches in the Cartesian product combination, 16 fuzzy subspaces are defined in the running feature space; Complete region segmentation is performed on the coordinate-translated running feature space based on the fuzzy subspace.

4. The method for analyzing the stability of offshore wind turbines connected to the grid under large disturbances according to claim 3, characterized in that, Constructing a TS fuzzy model set includes: Within each of the fuzzy subspaces formed by dividing the operating feature space, the extreme values ​​of the division parameters are substituted into the nonlinear dynamic model to generate a linear state-space equation that matches the corresponding fuzzy subspace. The linear state-space equations are bound to the corresponding fuzzy subspaces to construct a set of linear models corresponding to all the fuzzy subspaces; For each of the local linear models in the set of linear models, a rule is designed to activate the weight function based on the degree of change of the system dynamic parameters in the corresponding fuzzy subspace; Based on the rules, the output vectors of the linear model set are dynamically aggregated using the activation weight function to generate a continuous TS fuzzy model set covering the entire operating feature space.

5. The method for analyzing the stability of offshore wind turbines connected to the grid under large disturbances according to claim 4, characterized in that, Solving for the quadratic Lyapunov function that satisfies the linear matrix inequality (LMI) constraint includes: For all the local linear models in the TS fuzzy model set, a set of matrix inequality constraints reflecting the energy state of the system is constructed based on Lyapunov theory. The set of matrix inequality constraints includes positive definite conditions that guarantee the asymptotic stability of the system and state derivative decay constraints. By combining the correlation between the rule activation weight function and the local linear model, a dynamic coupling relationship is constructed between the matrix inequality constraint sets. The dynamic coupling relationship reflects the transition coordination conditions of the system's dynamic behavior under different fuzzy subspaces. Based on the continuity of the dynamic coupling relationship in all the operating feature spaces, the matrix inequality constraint set corresponding to all the local linear models in the TS fuzzy model set is solved simultaneously to obtain a unified quadratic coefficient matrix. Based on the unified quadratic coefficient matrix, an analytical expression of the quadratic Lyapunov function that satisfies all linear matrix inequalities (LMI) constraints is synthesized. The perturbation range parameters of the nonlinear dynamic model are expanded iteratively in rounds according to a preset step size, and the simultaneous solution process of the matrix inequality constraint set and the quadratic coefficient matrix is ​​restarted until a feasible solution cannot be obtained in the current round, at which point the iteration process is terminated. The boundary defined by the disturbance range parameter corresponding to the last successfully solved round is determined as the boundary of the maximum attraction domain of the nonlinear dynamic model.

6. The method for analyzing the large-disturbance stability of offshore wind turbine frequency converter grid connection according to claim 1, characterized in that, Generate a two-dimensional stable operating region map, including: Obtain the geometric parameters of the closed hypersurface formed by the boundary of the maximum attraction domain defined by the quadratic Lyapunov function in the 26-dimensional state space; The geometric parameters of the closed hypersurface are reduced in dimension along the direction dominated by electromechanical coupling characteristics and mapped to an orthogonal coordinate axis composed of the d-axis current of the inverter side and the DC voltage coordinate axis of the converter, as well as the d-axis current of the motor stator winding and the d-axis current coordinate axis of the generator stator winding. A convex envelope operation is performed on the mapped boundary point set to form an envelope boundary curve. The region enclosed by the envelope boundary curve represents the operating range of the system that maintains transient stability under power disturbance. Using the envelope boundary curve as the critical surface, a rectangular stable operating region map including the upper limit of active power and the upper limit of reactive power is generated. The horizontal axis of the rectangular stable operating region map is the active power axis, and the vertical axis is the reactive power axis.

7. The method for analyzing the stability of offshore wind turbines connected to the grid under large disturbances according to claim 1, characterized in that, The output power adjustment command includes: Real-time monitoring of the actual output active and reactive power of the offshore wind power system at the real-time operating point coordinates in the current electromechanical power plane; By comparing the real-time operating point coordinates with the boundary position relationship of the two-dimensional stable operating area map, when it is detected that the real-time operating point is approaching the boundary of the maximum state variable of the two-dimensional stable operating area map, a power limiting protection action is triggered. The intersection of the tangential trajectory of the real-time running point in the direction exceeding the boundary of the maximum state variable and the boundary of the two-dimensional stable running area map is calculated as the target safe power point; The active power value and reactive power value corresponding to the target safe power point are converted into a power command format that can be executed by the offshore wind power system controller; The output power command is dynamically modified according to the command interface of the offshore wind power system controller, and the power operating point of the offshore wind power system is constrained within the boundary of the two-dimensional stable operating area map.

8. The method for analyzing the stability of offshore wind turbines connected to the grid under large disturbances according to claim 7, characterized in that, Dynamic rate optimization of the output power command includes: The rate of change of the power slope at the real-time operating point in the electromechanical power plane is monitored, and the rate of change of the power slope is calculated based on the derivative data of the rotational speed deviation after coordinate translation on the time series. When the power limiting protection action is triggered, the power command adjustment rate parameter is selected from the adjustment strategy library pre-stored in the offshore wind power system controller according to the absolute value of the power slope change rate.

9. A large-disturbance stability analysis system for offshore wind turbine frequency converter grid connection, characterized in that, The system includes: The model building module establishes a 26-dimensional nonlinear dynamic model of the frequency conversion system of the motor and generator set. The spatial partitioning module uses the steady-state operating point of the nonlinear dynamic model as the origin of the coordinate system to perform coordinate translation, dividing the running feature space into 16 fuzzy subspaces. The local linear module performs local linearization on the nonlinear dynamic model in each fuzzy subspace, constructs the TS fuzzy model set, and the TS fuzzy model set is associated with each local linear model through a weighting function; The constraint solving module, based on the TS fuzzy model set, constructs the corresponding quadratic Lyapunov function, expands the range of nonlinear term perturbation in each round, and searches for the maximum perturbation boundary based on iterative solution to solve the maximum attraction domain boundary of the system; The image generation module projects the boundary of the maximum attraction domain onto the electromechanical power plane to generate a two-dimensional stable operating region map; The command stabilization module adjusts the output power command of the offshore wind power system controller to maintain system stability based on the maximum state variable boundary in the two-dimensional stable operating area diagram.

10. The large-disturbance stability analysis system for offshore wind turbine frequency converter grid connection according to claim 9, characterized in that, The image generation module includes: The parameter acquisition unit acquires the geometric parameters of the closed hypersurface formed by the boundary of the maximum attraction domain defined by the quadratic Lyapunov function in the 26-dimensional state space. The dimension reduction mapping unit reduces the geometric parameters of the closed hypersurface along the dominant direction of electromechanical coupling characteristics and maps them to an orthogonal coordinate axis composed of the inverter side d-axis current and converter DC voltage coordinate axis, as well as the motor stator winding d-axis current and generator stator winding d-axis current coordinate axis. The envelope curve unit performs convex envelope operations on the mapped set of boundary points to form an envelope boundary curve. The region enclosed by the envelope boundary curve represents the operating range of the system that maintains transient stability under power disturbances. The image generation unit uses the envelope boundary curve as the critical surface to generate a rectangular stable operating region map that includes the upper limit values ​​of active power and reactive power. The horizontal axis of the rectangular stable operating region map is the active power axis, and the vertical axis is the reactive power axis.

Images