A method and system for analyzing the safe operation of grid-connected offshore wind farms

Abstract

This invention provides a method and system for analyzing the safe operation of grid-connected offshore wind farms. The method includes: acquiring the network topology and operating parameters of the grid-connected offshore wind farm to construct a power flow feasible domain injection space and a node admittance matrix; obtaining the power flow feasible domain based on these two parameters to determine stability criteria; generating an operating sample set by determining the active power injection space of the grid-connected offshore wind system and the number of training samples; and combining the operating sample set and stability criteria to obtain a stability margin set, which is then used to train a margin prediction model under a nonlinear neural network for hyperparameter selection and early shutdown strategy training. Finally, the active power injection space of the grid-connected offshore wind system is used to obtain an operating condition set under the power flow feasible domain, and a stability margin prediction set is obtained under the margin prediction model to construct the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system for safe operation analysis. This invention can improve the accuracy of safe operation analysis of grid-connected offshore wind farms.

Description

Technical Field

[0001] This invention relates to the field of wind farm operation technology, and in particular to a method and system for analyzing the safe operation of grid-connected offshore wind farms. Background Technology

[0002] When disturbances occur in traditional power systems, thermal power generating units, due to their large rotational inertia, can naturally resist frequency shifts to maintain the dynamic stability of the power system. However, offshore wind turbines are typically connected to the grid via power electronic converters, forming an electromagnetic coupling link. They cannot provide rotational inertia and damping support to the power system. Therefore, in scenarios with high wind power penetration, the power system of grid-connected offshore wind farms is more sensitive to disturbances, experiences more severe frequency fluctuations, and has a reduced power angle stability margin, leading to more frequent low-frequency oscillations.

[0003] Existing technologies determine storm control strategies for offshore wind turbines by using operational data and vibration data of offshore wind turbines in the current sea area. They then use deep learning models to analyze storm control strategies and storm parameters to obtain operational risk data for offshore wind turbines. Subsequently, a safety risk assessment model is constructed to obtain a safety assessment level from the operational risk data. Finally, based on the safety assessment level, a safety operation analysis is conducted to ensure the safe operation of offshore wind farms.

[0004] However, the existing technology lacks modeling of the dynamic stability physical mechanism of the grid-connected offshore wind system, and only relies on the predicted operational risk data to obtain the safety assessment level for safety operation analysis. Therefore, it cannot construct an effective low-frequency oscillation stability criterion and safe operation boundary, thus it cannot accurately characterize the controllable operating range of the low-frequency oscillation risk of the offshore wind system, nor can it fully reflect the safety margin of the grid-connected offshore wind system in terms of stability, which reduces the accuracy of the safety operation analysis of the grid-connected offshore wind farm. Summary of the Invention

[0005] The present invention aims to provide a method and system for analyzing the safe operation of grid-connected offshore wind farms, so as to solve the above-mentioned technical problems and improve the accuracy of the analysis of the safe operation of grid-connected offshore wind farms.

[0006] To address the aforementioned technical problems, this invention provides a method for analyzing the safe operation of grid-connected offshore wind farms, comprising the following steps: Obtain the network topology and operating parameters of the grid-connected offshore wind farm, and establish an admittance model of the offshore wind farm and transmission line based on the network topology and operating parameters; Based on the network topology and operating parameters, a power flow feasible region injection space and a node admittance matrix are constructed, and the power flow feasible region is obtained based on the power flow feasible region injection space and the node admittance matrix. Stability criteria are obtained based on the admittance model of the offshore wind farm and transmission line. The active power injection space and training sample number of the grid-connected offshore wind system are determined based on the current tidal feasible domain, and an operating sample set is generated based on the active power injection space and the training sample number. Based on the running sample set and the stability criterion, a stability margin set is obtained. Then, based on the stability margin set and the running sample set, hyperparameter selection and early stopping strategy training are performed under a preset nonlinear neural network to obtain a margin prediction model. The active power injection space is densely sampled under the power flow feasible domain to obtain the operating condition set, and the stability margin prediction set is obtained under the margin prediction model based on the operating condition set. The low-frequency oscillation safety domain boundary of the grid-connected offshore wind system is constructed based on the stability margin prediction set, and a safety operation analysis is performed based on the low-frequency oscillation safety domain boundary.

[0007] The above-mentioned scheme, based on the tidal current feasible domain injection space and the nodal admittance matrix, iteratively solves the tidal current feasible domain under the Newton-Raphson iteration method, which can reduce misjudgments caused by inconsistencies with the electrical state of the offshore wind system. Subsequently, this scheme determines the stability criterion based on the tidal current feasible domain and obtains the stability margin set by generating an operating sample set using the Monte Carlo method. Based on the stability margin set and the operating sample set, hyperparameter selection is performed to train and optimize the preset nonlinear neural network for early shutdown strategy, so that the trained margin prediction model can effectively learn the complex nonlinear mapping relationship from operating conditions to stability margin, improving the accuracy of subsequent stability margin prediction. Subsequently, this scheme performs dense sampling of the active power injection space under the tidal current feasible domain to generate an operating condition set, and uses the above margin prediction model to obtain the stability margin prediction set. This stability margin prediction set can accurately distinguish between safe and unstable regions. Based on the safety domain boundary of the low-frequency oscillation safety domain of the grid-connected offshore wind system constructed based on the stability margin prediction set, safe operation analysis can be performed, which can effectively assess the stability of the offshore wind system and improve the accuracy of safe operation analysis of grid-connected offshore wind farms.

[0008] Furthermore, the step of determining the stability criterion, the active power injection space of the grid-connected offshore wind system, and the number of training samples based on the feasible power flow domain, and generating an operational sample set under a preset Monte Carlo method based on the active power injection space of the grid-connected offshore wind system and the number of training samples, includes: establishing an admittance model of the offshore wind farm and transmission line based on the network topology and operating parameters; obtaining the stability criterion based on the admittance model of the offshore wind farm and transmission line; determining the active power injection space and the number of training samples of the grid-connected offshore wind system based on the feasible power flow domain, and generating an operational sample set under a preset Monte Carlo method based on the active power injection space of the grid-connected offshore wind system and the number of training samples.

[0009] The aforementioned scheme establishes admittance models for offshore wind farms and transmission lines based on network topology and operating parameters. Both these models are frequency domain models. Subsequently, the stability criteria obtained from these models accurately reflect the stability characteristics of low-frequency oscillations in the offshore wind system under different operating conditions, making the subsequent margin prediction model more accurate. Furthermore, this scheme determines the active power injection space and training sample number of the grid-connected offshore wind system based on the tidal current feasible domain. This ensures that the generated operating sample set is limited to a statically safe and reasonable range within the tidal current feasible domain, avoiding the generation of invalid samples and improving the predictive reliability of the subsequent margin prediction model. This, in turn, enhances the accuracy of the safety operation analysis of grid-connected offshore wind farms.

[0010] Furthermore, the step of obtaining the stability criterion based on the admittance model of the offshore wind farm and transmission line includes: performing difference processing on the admittance model of the offshore wind farm and transmission line to obtain the equivalent admittance model of the grid-connected offshore wind system, and performing a Schul transform on the equivalent admittance model of the grid-connected offshore wind system to obtain the Schul complement of the admittance model; determining the oscillation frequency range that makes the Schul complement of the admittance model greater than zero, and using the oscillation frequency range as the stability criterion.

[0011] The aforementioned scheme obtains an equivalent admittance model for the grid-connected offshore wind system by subtracting the admittance model from that of the offshore wind farm and the transmission line. This equivalent admittance model reflects the dynamic interaction between the wind farm and the power grid in the offshore wind system. Subsequently, this scheme performs a Schuler transform on the equivalent admittance model to obtain the Schul complement of the admittance model. This transforms the stability problem of whether the offshore wind system will experience low-frequency oscillations into a problem of judging the stability of the one-dimensional function of the Schul complement in the frequency domain. The oscillation frequency range that makes the Schul complement greater than zero, which maintains the stability of the offshore wind system, is directly used as the stability criterion, reducing the complexity of stability judgment. Therefore, this scheme uses a Schuler transform to reduce the equivalent admittance model of the grid-connected offshore wind system to a one-dimensional representation of the admittance model Schul complement. Based on this reduced-dimensional admittance model Schul complement, a stability criterion for judging the low-frequency oscillation stability of the offshore wind system under different operating conditions can be obtained quickly and accurately, thereby improving the efficiency of the safety operation analysis of grid-connected offshore wind farms.

[0012] Further, the step of obtaining a stability margin set based on the operating sample set and the stability criterion, and then performing hyperparameter selection and early stopping strategy training processing under a preset nonlinear neural network based on the stability margin set and the operating sample set to obtain a margin prediction model, includes: obtaining a stability margin set based on the operating sample set and the stability criterion; integrating the stability margin set and the operating sample set to obtain a sea breeze system training dataset, and performing normalization and sample balancing processing on the sea breeze system training dataset to generate a sea breeze system training set and a sea breeze system validation set; performing hyperparameter selection and early stopping strategy training processing under a preset nonlinear neural network based on the sea breeze system training set and the sea breeze system validation set to obtain an optimized nonlinear neural network, and using the optimized nonlinear neural network as the margin prediction model.

[0013] The above scheme first obtains a stability margin set based on the operating sample set and stability criteria. The stability margin set and the operating sample set are then integrated to obtain a training dataset for the offshore wind system. Normalization and sample balancing are then performed on the training dataset to eliminate differences in data dimensions and improve data distribution, generating high-quality training and validation sets for the offshore wind system. This lays a reliable foundation for the subsequent training of the pre-defined nonlinear neural network. Finally, based on the offshore wind system training and validation sets, this scheme performs hyperparameter selection and early shutdown strategy training under the pre-defined nonlinear neural network. The resulting optimized nonlinear neural network serves as the margin prediction model. Hyperparameter selection systematically finds the optimal model structure parameters, while the early shutdown strategy monitors the training process through the offshore wind system validation set, effectively preventing overfitting on the training set. This ensures that the obtained margin prediction model possesses both strong nonlinear fitting ability and good generalization ability, enabling it to quickly and reliably assess the stability of massive operating conditions and obtain a stability margin prediction set, thereby improving the accuracy of the safety operation analysis of grid-connected offshore wind farms.

[0014] Furthermore, the step of constructing the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the stability margin prediction set, and performing safety operation analysis based on the low-frequency oscillation safety domain boundary, includes: screening a set of safe operation stability margins under a preset safe operation margin condition from the stability margin prediction set, and determining the low-frequency oscillation safety domain based on the safe operation stability margin set; constructing the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the low-frequency oscillation safety domain boundary, and performing safety operation analysis based on the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system.

[0015] The aforementioned scheme first selects a set of safe operating stability margins from the stability margin prediction set under a preset safe operating margin condition. Based on this set, a low-frequency oscillation safety domain is determined. This low-frequency oscillation safety domain then forms the boundary of a grid-connected offshore wind farm's low-frequency oscillation safety domain with clear and well-defined stability partitions. This boundary accurately delineates the safe and stable operating area. This scheme performs safety operation analysis based on the boundary of this grid-connected offshore wind farm's low-frequency oscillation safety domain. It can intuitively reflect whether the current or planned operating point is located within this domain and accurately quantify its distance from the instability boundary, thus enabling more refined safety operation analysis and improving the accuracy of safety operation analysis for grid-connected offshore wind farms.

[0016] Furthermore, the step of constructing the power flow feasible domain injection space and node admittance matrix based on the network topology and operating parameters, and obtaining the power flow feasible domain based on the power flow feasible domain injection space and the node admittance matrix, includes: forming a node admittance matrix based on the network topology and operating parameters; and constructing the power flow feasible domain injection space based on the operating status of each wind turbine in the network topology and operating parameters.

[0017] The above scheme forms a node admittance matrix based on network topology and operating parameters. This node admittance matrix can accurately reflect the connection relationship and electrical parameters of each component in the offshore wind power grid. Based on the operating status of each wind turbine in the network topology and operating parameters, a power flow feasible domain injection space is constructed, which can reflect the impact of different operating conditions on the stability of low-frequency oscillations of the system, ensure the reliability of the power flow feasible domain obtained subsequently, and thus improve the accuracy of the safety operation analysis of grid-connected offshore wind farms.

[0018] Further, the step of constructing a power flow feasible domain injection space and a node admittance matrix based on the network topology and operating parameters, and obtaining the power flow feasible domain based on the power flow feasible domain injection space and the node admittance matrix, includes: based on the node admittance matrix, performing iterative solutions using the Newton iteration method on each power group in the power flow feasible domain injection space under a preset initial node voltage value until a preset safe operation convergence threshold is met, to obtain the power flow result; and obtaining the power flow feasible domain under different operating conditions based on the power flow result under a preset given node voltage constraint.

[0019] The aforementioned scheme first injects the power flow feasible region into the space based on the nodal admittance matrix. For each power setting under a preset initial nodal voltage, iterative solutions are obtained using the Newton-Raphson method until a preset safe operation convergence threshold is met, yielding the power flow results. This process utilizes the Newton-Raphson method to calculate the power flow for each power setting, ensuring that the obtained power flow results are accurate solutions to the power balance equation of the offshore wind system. Furthermore, the preset safe operation convergence threshold guarantees the reliability of the obtained power flow results. Subsequently, based on these power flow results, this scheme obtains the power flow feasible region under different operating conditions under a preset given nodal voltage constraint, ensuring the reliability of the subsequently obtained operating sample set and thus improving the accuracy of the safe operation analysis of grid-connected offshore wind farms.

[0020] This invention also provides a safety operation analysis system for grid-connected offshore wind farms, used to implement any of the above-mentioned safety operation analysis methods for grid-connected offshore wind farms, including: an initial construction module for the offshore wind system, used to acquire the network topology and operating parameters of the grid-connected offshore wind farm, and construct a tidal current feasible domain injection space and a node admittance matrix based on the network topology and operating parameters; a tidal current feasible domain acquisition module, used to iteratively solve the tidal current feasible domain injection space and the node admittance matrix under a preset Newton iteration method until a preset safety operation convergence threshold is met, to obtain the tidal current feasible domain; and an operation sample generation module, used to determine the stability criterion, the active power injection space of the grid-connected offshore wind system, and the number of training samples based on the tidal current feasible domain, and to construct a tidal current feasible domain based on the active power injection space of the grid-connected offshore wind system and the number of training samples. The system generates an operational sample set using a preset Monte Carlo method; a margin prediction model building module is used to obtain a stability margin set based on the operational sample set and the stability criterion, and to perform hyperparameter selection and early shutdown strategy training processing under a preset nonlinear neural network based on the stability margin set and the operational sample set to obtain a margin prediction model; a stability margin prediction module is used to perform intensive sampling of the active power injection space of the grid-connected offshore wind system under the tidal feasible domain to obtain an operational condition set, and to obtain a stability margin prediction set based on the operational condition set under the margin prediction model; a safe operation analysis module is used to construct the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the stability margin prediction set, and to perform safe operation analysis based on the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system.

[0021] The present invention also provides a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements any of the above-mentioned methods for analyzing the safe operation of grid-connected offshore wind farms.

[0022] The present invention also provides a computer-readable storage medium, comprising: a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute any of the above-mentioned methods for analyzing the safe operation of a grid-connected offshore wind farm.

[0023] The above scheme reduces misjudgments caused by inconsistencies with the electrical state of the offshore wind system by iteratively solving the tidal current feasible domain injection space and nodal admittance matrix using the Newton-Raphson method. Subsequently, stability criteria are determined based on this tidal current feasible domain, and a stability margin set is obtained using an operating sample set generated by the Monte Carlo method. Hyperparameter selection is performed based on this stability margin set and the operating sample set to train and optimize a preset nonlinear neural network for early shutdown strategy training. This enables the trained margin prediction model to effectively learn the complex nonlinear mapping relationship from operating conditions to stability margin, improving the accuracy of subsequent stability margin predictions. Subsequently, the active power injection space is densely sampled under the tidal current feasible domain to generate an operating condition set, and the stability margin prediction set is obtained using the aforementioned margin prediction model. This stability margin prediction set can accurately distinguish between safe and unstable regions. Safe operation analysis is then performed on the boundary of the low-frequency oscillation safety domain of the grid-connected offshore wind system constructed based on this stability margin prediction set, effectively assessing the stability of the offshore wind system and improving the accuracy of safe operation analysis of grid-connected offshore wind farms. Attached Figure Description

[0024] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1 A flowchart illustrating the technical implementation of a method for analyzing the safe operation of a grid-connected offshore wind farm, as provided in an embodiment of the present invention. Figure 2 An equivalent circuit model of a grid-connected offshore wind system is provided for an embodiment of the present invention to analyze the safe operation of a grid-connected offshore wind farm. Figure 3 This is a schematic diagram of the architecture of a grid-connected offshore wind farm safety operation analysis system provided in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0027] 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 application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0028] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0029] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0030] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0031] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0032] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0033] Please see Figure 1 This embodiment provides a method for analyzing the safe operation of a grid-connected offshore wind farm, including the following steps: Step S1: Obtain the network topology and operating parameters of the grid-connected offshore wind farm, and establish an admittance model of the offshore wind farm and transmission line based on the network topology and operating parameters; Step S2: Construct the power flow feasible region injection space and node admittance matrix based on the network topology and operating parameters, and obtain the power flow feasible region based on the power flow feasible region injection space and the node admittance matrix; Step S3: Obtain stability criteria based on the offshore wind farm and transmission line admittance model; determine the active power injection space and training sample number of the grid-connected offshore wind system according to the tidal current feasible region, and generate an operating sample set according to the active power injection space and the training sample number; Step S4: Based on the running sample set and the stability criterion, obtain a stability margin set, and perform hyperparameter selection and early stopping strategy training processing under a preset nonlinear neural network according to the stability margin set and the running sample set to obtain a margin prediction model. Step S5: Densely sample the active power injected into the space under the power flow feasible domain to obtain the operating condition set, and obtain the stability margin prediction set under the margin prediction model based on the operating condition set. Step S6: Construct the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the stability margin prediction set, and perform safety operation analysis based on the low-frequency oscillation safety domain boundary.

[0034] The above embodiment, based on the tidal current feasible domain injection space and node admittance matrix, can reduce misjudgments caused by inconsistencies with the electrical state of the offshore wind system. Subsequently, this embodiment determines the stability criterion based on the tidal current feasible domain and obtains a stability margin set by combining the operating sample set. Based on the stability margin set and the operating sample set, hyperparameter selection is performed to train and optimize the preset nonlinear neural network for early shutdown strategy, enabling the trained margin prediction model to effectively learn the complex nonlinear mapping relationship from operating conditions to stability margin, thereby improving the accuracy of subsequent stability margin prediction. Subsequently, this embodiment performs dense sampling of the active power injection space under the tidal current feasible domain to generate an operating condition set, and uses the above margin prediction model to obtain a stability margin prediction set. This stability margin prediction set can accurately distinguish between safe and unstable regions. Based on the safety operation analysis of the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system constructed based on the stability margin prediction set, the stability of the offshore wind system can be effectively assessed, improving the accuracy of the safety operation analysis of the grid-connected offshore wind farm.

[0035] It should be noted that in deep learning training, hyperparameters are configuration parameters that are pre-set before model training begins. These include the number of layers in the neural network and the number of neurons in each layer to control model complexity, the learning rate and batch size to control parameter update step size, and the dropout rate to actively suppress model overfitting. The above embodiments, through searches and evaluations such as grid search and random search, enable hyperparameter selection to automatically find and determine the model configuration of the nonlinear neural network best suited to the current stability prediction task, ensuring that the resulting margin prediction model has high prediction accuracy.

[0036] It should also be noted that the early termination strategy is a method to prevent overfitting by monitoring the validation set of the sea breeze system during the training of the preset nonlinear neural network model. This ensures that the nonlinear neural network model trained in the above embodiment is kept in the state with the strongest generalization ability, and ensures that the margin prediction model obtained later can not only reproduce known samples, but also reliably predict the stability margin of unknown working conditions, thus guaranteeing the reliability and generalization of the model.

[0037] The above embodiment describes a specific method for optimizing the training of a nonlinear neural network: after each or several rounds of training, the model is evaluated simultaneously on both the sea breeze system training set and an independent sea breeze system validation set; performance metrics such as prediction error on the sea breeze system validation set are recorded; when it is found that the metric no longer improves or even begins to deteriorate in multiple rounds of training, the training will be forcibly terminated even though the model's error on the sea breeze system training set may still be decreasing; after training stops, the algorithm automatically saves the model parameters corresponding to the best performance on the sea breeze system validation set as the final optimized nonlinear neural network, and uses this optimized nonlinear neural network as the margin prediction model.

[0038] Furthermore, the step of determining the stability criterion, the active power injection space of the grid-connected offshore wind system, and the number of training samples based on the feasible power flow domain, and generating an operational sample set under a preset Monte Carlo method based on the active power injection space of the grid-connected offshore wind system and the number of training samples, includes: establishing an admittance model of the offshore wind farm and transmission line based on the network topology and operating parameters; obtaining the stability criterion based on the admittance model of the offshore wind farm and transmission line; determining the active power injection space and the number of training samples of the grid-connected offshore wind system based on the feasible power flow domain, and generating an operational sample set under a preset Monte Carlo method based on the active power injection space of the grid-connected offshore wind system and the number of training samples.

[0039] The above embodiments establish admittance models for offshore wind farms and transmission lines based on network topology and operating parameters. Both these models are frequency domain models. Subsequently, this embodiment reduces the dimensionality of the offshore wind farm and transmission line admittance models to obtain the Schul complement admittance model. The stability criterion obtained from this Schul complement accurately reflects the stability characteristics of low-frequency oscillations in the offshore wind system under different operating conditions, making the subsequent margin prediction model more accurate. Furthermore, this embodiment determines the active power injection space and training sample number of the grid-connected offshore wind system based on the tidal current feasible domain. This ensures that the generated operating sample set is limited to a statically safe and reasonable range within the tidal current feasible domain, avoiding the generation of invalid samples, improving the predictive reliability of the subsequently generated margin prediction model, and thus improving the accuracy of the safe operation analysis of grid-connected offshore wind farms.

[0040] Furthermore, the step of obtaining the stability criterion based on the admittance model of the offshore wind farm and transmission line includes: performing difference processing on the admittance model of the offshore wind farm and transmission line to obtain the equivalent admittance model of the grid-connected offshore wind system, and performing a Schul transform on the equivalent admittance model of the grid-connected offshore wind system to obtain the Schul complement of the admittance model; determining the oscillation frequency range that makes the Schul complement of the admittance model greater than zero, and using the oscillation frequency range as the stability criterion.

[0041] The above embodiment obtains an equivalent admittance model for the grid-connected offshore wind system by subtracting the admittance model from that of the offshore wind farm and the transmission line. This equivalent admittance model reflects the dynamic interaction between the wind farm and the power grid in the offshore wind system. Subsequently, this embodiment performs a Schul transform on the equivalent admittance model to obtain the Schul complement of the admittance model. This transforms the stability problem of whether the offshore wind system will experience low-frequency oscillations into a problem of judging the stability of the one-dimensional function of the admittance model's Schul complement in the frequency domain. The oscillation frequency range that makes the admittance model's Schul complement greater than zero, which maintains the stability of the offshore wind system, is directly used as the stability criterion, reducing the complexity of stability judgment. Therefore, this embodiment uses a Schul transform to reduce the equivalent admittance model of the grid-connected offshore wind system to a one-dimensional representation of the admittance model's Schul complement. Based on this reduced-dimensional admittance model's Schul complement, a stability criterion for judging the low-frequency oscillation stability of the offshore wind system under different operating conditions can be obtained quickly and accurately, thereby improving the efficiency of the safety operation analysis of the grid-connected offshore wind farm.

[0042] In one embodiment, the admittance model of an offshore wind farm and transmission line is as follows: Figure 2 The equivalent circuit model of the grid-connected offshore wind system is shown in the figure, which is a method for analyzing the safe operation of a grid-connected offshore wind farm. The figure includes an offshore wind farm admittance model and an AC transmission line admittance model. This is a model for the admittance of offshore wind farms. This is an admittance model for AC transmission lines.

[0043] The above embodiments are based on the admittance model of offshore wind farms. AC transmission line admittance model By performing interpolation, the equivalent admittance model of the grid-connected offshore wind system is obtained. ,and Subsequently, the equivalent admittance model of the grid-connected offshore wind system was used. The Schul transform yields the one-dimensional representation of the admittance model, Schul complement. ,in, and .

[0044] The above embodiments allow the admittance model to be supplemented by Schul. And admittance model Shur complement The real part is The imaginary part is ,like , then it means At angular frequency There is a zero-crossing point at that point, then at... The sea breeze system is stable at this time. The sea breeze system is at risk of oscillation and instability, and its oscillation frequency is This determines the admission model's Schuler compensation. The range of oscillation frequencies greater than zero is used as the stability criterion.

[0045] Further, the step of obtaining a stability margin set based on the operating sample set and the stability criterion, and then performing hyperparameter selection and early stopping strategy training processing under a preset nonlinear neural network based on the stability margin set and the operating sample set to obtain a margin prediction model, includes: obtaining a stability margin set based on the operating sample set and the stability criterion; integrating the stability margin set and the operating sample set to obtain a sea breeze system training dataset, and performing normalization and sample balancing processing on the sea breeze system training dataset to generate a sea breeze system training set and a sea breeze system validation set; performing hyperparameter selection and early stopping strategy training processing under a preset nonlinear neural network based on the sea breeze system training set and the sea breeze system validation set to obtain an optimized nonlinear neural network, and using the optimized nonlinear neural network as the margin prediction model.

[0046] The above embodiment first obtains a stability margin set based on the operating sample set and stability criteria. The stability margin set and the operating sample set are then integrated to obtain a training dataset for the offshore wind system. Subsequently, the training dataset is normalized and balanced to eliminate differences in data dimensions and improve data distribution, generating high-quality training and validation sets for the offshore wind system. This lays a reliable foundation for the subsequent training of the pre-set nonlinear neural network. Finally, this embodiment uses the offshore wind system training and validation sets to perform hyperparameter selection and early shutdown strategy training under the pre-set nonlinear neural network. The resulting optimized nonlinear neural network serves as the margin prediction model. Hyperparameter selection systematically finds the optimal model structure parameters, while the early shutdown strategy monitors the training process through the offshore wind system validation set, effectively preventing overfitting on the offshore wind system training set. This ensures that the obtained margin prediction model possesses both strong nonlinear fitting ability and good generalization ability, enabling it to quickly and reliably assess the stability of massive operating conditions and obtain a stability margin prediction set, thereby improving the accuracy of the safety operation analysis of grid-connected offshore wind farms.

[0047] It should be noted that the above embodiments are based on the sea breeze system training set and the sea breeze system validation set, and are subjected to hyperparameter selection and early stopping strategy training under a preset nonlinear neural network to obtain an optimized nonlinear neural network. Specifically, the optimized nonlinear neural network is obtained by minimizing the preset regression loss function using the sea breeze system training set and combining it with the sea breeze system validation set to perform hyperparameter selection and early stopping strategy.

[0048] Furthermore, the step of constructing the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the stability margin prediction set, and performing safety operation analysis based on the low-frequency oscillation safety domain boundary, includes: screening a set of safe operation stability margins under a preset safe operation margin condition from the stability margin prediction set, and determining the low-frequency oscillation safety domain based on the safe operation stability margin set; constructing the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the low-frequency oscillation safety domain boundary, and performing safety operation analysis based on the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system.

[0049] The above embodiment first filters the stability margin prediction set under the preset safe operating margin condition to form a safe operating stability margin set, and then determines the low-frequency oscillation safety domain based on the safe operating stability margin set. Based on this low-frequency oscillation safety domain, a boundary of the grid-connected offshore wind farm's low-frequency oscillation safety domain with clear and well-defined stability partitions is constructed. This low-frequency oscillation safety domain boundary accurately characterizes the safe and stable operating area. This embodiment performs safe operation analysis based on the boundary of the grid-connected offshore wind farm's low-frequency oscillation safety domain, which can intuitively reflect whether the current or planned operating point is located within the low-frequency oscillation safety domain of the grid-connected offshore wind farm, and can accurately quantify its distance from the instability boundary, thereby enabling more refined safe operation analysis and improving the accuracy of safe operation analysis of grid-connected offshore wind farms.

[0050] Furthermore, the step of constructing the power flow feasible domain injection space and node admittance matrix based on the network topology and operating parameters, and obtaining the power flow feasible domain based on the power flow feasible domain injection space and the node admittance matrix, includes: forming a node admittance matrix based on the network topology and operating parameters; and constructing the power flow feasible domain injection space based on the operating status of each wind turbine in the network topology and operating parameters.

[0051] The above embodiments form a node admittance matrix based on network topology and operating parameters. This node admittance matrix can accurately reflect the connection relationship and electrical parameters of each component in the offshore wind power grid. Based on the operating status of each wind turbine in the network topology and operating parameters, a power flow feasible domain injection space is constructed, which can reflect the impact of different operating conditions on the stability of low-frequency oscillations of the system, ensure the reliability of the power flow feasible domain obtained subsequently, and thus improve the accuracy of the safety operation analysis of grid-connected offshore wind farms.

[0052] Further, the step of constructing a power flow feasible domain injection space and a node admittance matrix based on the network topology and operating parameters, and obtaining the power flow feasible domain based on the power flow feasible domain injection space and the node admittance matrix, includes: based on the node admittance matrix, performing iterative solutions using the Newton iteration method on each power group in the power flow feasible domain injection space under a preset initial node voltage value until a preset safe operation convergence threshold is met, to obtain the power flow result; and obtaining the power flow feasible domain under different operating conditions based on the power flow result under a preset given node voltage constraint.

[0053] The above embodiment first injects the power flow feasible region into the space based on the nodal admittance matrix. For each power setting under a preset initial nodal voltage, Newton's iteration method is used for iterative solution until a preset safe operation convergence threshold is met, yielding the power flow result. This process utilizes Newton's iteration method to calculate the power flow for each power setting, ensuring that the obtained power flow result is an accurate solution satisfying the power balance equation of the offshore wind system. Furthermore, the preset safe operation convergence threshold guarantees the reliability of the obtained power flow result. Subsequently, based on this power flow result, this embodiment obtains the power flow feasible region under different operating conditions under a preset given nodal voltage constraint, ensuring the reliability of the subsequently obtained operating sample set, thereby improving the accuracy of the safe operation analysis of grid-connected offshore wind farms.

[0054] In one embodiment, the network topology and operating parameters of a grid-connected offshore wind farm are obtained, and a node admittance matrix is ​​formed based on the network topology and operating parameters. Based on the network topology and operating parameters, the operating status of each wind turbine's output active power is used to construct the power flow feasible domain injection space. , The total number of wind turbines in the offshore wind farm is used as a basis for determining the tidal current feasible domain injection space based on the operating status. The range of values ​​for is denoted as . ,in, This represents the minimum output power of each wind turbine in an offshore wind farm. This represents the maximum output power of each wind turbine in the offshore wind farm.

[0055] Based on nodal admittance matrix Injecting the feasible domain of trends into space Each power group is set at a preset initial node voltage value. The solution is then obtained by iteratively solving using Newton's method until the preset safe operation convergence threshold is met. The result of the current flow is obtained in the first stage. During the iteration of the second Newton-Raphson method, the imbalance of various types of nodes is calculated. , and If the imbalance of various nodes , and The maximum absolute value is less than the safe operation convergence threshold. That is, when If the model training converges, the iterative solution process using the Newton-Raphson method ends and the power flow result is output; otherwise, the iteration count is incremented and the Newton-Raphson method iterative process continues.

[0056] Based on the power flow results, a preset node voltage constraint is applied. The feasible power flow domain under different operating conditions is obtained. , This refers to the operating conditions of the sea breeze system.

[0057] The above embodiments construct node admittance matrices based on network topology and operating parameters. This ensures that the model is compatible with actual grid-connected offshore wind farms at the real physical level. Subsequently, the feasible domain of tidal current is injected with space using the Newton-Raphson iteration method. Each power setting within the system is solved with high precision, and strictly adheres to node voltage constraints. The selection process is performed to obtain the final feasible power flow domain. It is a collection of all static safe operating points. This ensures that subsequent steps are carried out only under feasible operating conditions, eliminating misjudgments caused by inadequate basic electrical conditions for actual grid-connected offshore wind farms, and improving the accuracy of subsequent safe operation analysis.

[0058] Establish an admittance model for offshore wind farms based on network topology and operating parameters. AC transmission line admittance model ,and Offshore wind farm admittance model AC transmission line admittance model The difference processing yielded the equivalent admittance model of the grid-connected offshore wind system. ,and .

[0059] Subsequently, the equivalent admittance model of the grid-connected sea breeze system was used. The Schul transform yields the one-dimensional representation of the admittance model, Schul complement. ,in, and .

[0060] As can be seen, this embodiment uses the Schur transform to transform the two-dimensional equivalent admittance model of the grid-connected offshore wind system. Schur complement of the admittance model reduced to one-dimensional representation This allows for the establishment of a low-frequency oscillation stability criterion, based on the principle of equivalent resonant circuits, in subsequent processes. This enables the rapid and accurate determination of the low-frequency oscillation stability of the sea breeze system under different operating conditions.

[0061] Schuler's compensation model And actually it is The imaginary part is ,like , then it means At angular frequency There is a zero-crossing point at that point, then at... The sea breeze system is stable at this time. The sea breeze system is at risk of oscillation and instability, and its oscillation frequency is This determines the admission model's Schuler compensation. The range of oscillation frequencies greater than zero is used as the stability criterion. Therefore, the above embodiments transform the complex system stability problem into a matter of... Zero crossing and The positive and negative judgments improve the accuracy and feasibility of the obtained stability criteria for the sea-wind system.

[0062] Subsequently, based on the feasible domain of the trend Determine the space for active power injection of the grid-connected offshore wind system and number of training samples ,and Based on the active power injection of the grid-connected offshore wind system into the space and number of training samples Randomly generate the running sample set under the preset Monte Carlo method. .

[0063] Then based on the running sample set The stability criterion is used to obtain the running sample set. Each sample stability margin All samples stability margin Integrate into a stable margin set.

[0064] The above stability margin set and operating sample set The training dataset of the sea breeze system was obtained by integration. And the training dataset for the sea breeze system. Normalization and sample balancing were performed to divide the system into a training set and a validation set for the sea breeze system.

[0065] Based on the aforementioned training and validation sets of the sea breeze system, hyperparameter selection and early termination strategy training are performed on a pre-defined nonlinear neural network algorithm model. Specifically, the training set minimizes a pre-defined regression loss function, and the validation set is used to optimize the pre-defined nonlinear neural network algorithm model through hyperparameter selection and early termination strategies, resulting in an optimized nonlinear neural network. This optimized nonlinear neural network is then used as the margin prediction model. .

[0066] The improved nonlinear neural network algorithm based on hyperparameter selection and early stopping strategy optimization in the above embodiments can improve the versatility of the nonlinear neural network algorithm and the accuracy of its prediction results.

[0067] The active power of the grid-connected offshore wind system will then be injected into the space. In the feasible area of ​​trends Several operating condition sets were obtained through dense sampling, and these operating condition sets were then... The input is fed into the trained, optimized nonlinear neural network, i.e., into the margin prediction model. In the middle, the corresponding stability margin prediction value is quickly calculated. This allows us to obtain a stable margin prediction set.

[0068] Finally, the stability margin prediction set is set under the preset safety margin conditions. The system first selects a set of safe operation stability margins and then determines the low-frequency oscillation safety domain based on these sets. Based on the low-frequency oscillation safety domain, it constructs the boundary of the low-frequency oscillation safety domain for the grid-connected offshore wind system and performs a safety operation analysis based on this boundary.

[0069] As can be seen, the above embodiments, which construct the low-frequency oscillation safety domain based on the nonlinear neural network algorithm model, can capture the nonlinear relationship between input and output over a wide range. They can construct the low-frequency oscillation safety domain of the grid-connected offshore wind system with high accuracy and low computational cost, obtain a continuous and smooth boundary of the low-frequency oscillation safety domain of the grid-connected offshore wind system, and have good scalability and engineering applicability. They are suitable for safety operation analysis under the condition of multiple wind farms connected to the grid, and can improve the accuracy of safety operation analysis of grid-connected offshore wind farms.

[0070] Please see Figure 3 This embodiment also provides a grid-connected offshore wind farm safety operation analysis system, used to implement any of the above-mentioned grid-connected offshore wind farm safety operation analysis methods, including: The initial construction module of the offshore wind system is used to obtain the network topology and operating parameters of the grid-connected offshore wind farm, and to establish an admittance model of the offshore wind farm and transmission line based on the network topology and operating parameters. The feasible power flow domain acquisition module is used to construct a feasible power flow domain injection space and a node admittance matrix based on the network topology and operating parameters, and to obtain the feasible power flow domain based on the feasible power flow domain injection space and the node admittance matrix. The sample generation module is used to obtain stability criteria based on the admittance model of the offshore wind farm and transmission line; determine the active power injection space and training sample number of the grid-connected offshore wind system according to the tidal current feasible domain; and generate an operating sample set according to the active power injection space and the training sample number. The margin prediction model building module is used to obtain a stable margin set based on the running sample set and the stability criterion, and to perform hyperparameter selection and early stopping strategy training processing under a preset nonlinear neural network based on the stable margin set and the running sample set to obtain the margin prediction model. The stability margin prediction module is used to perform intensive sampling of the active power injection space under the power flow feasible domain to obtain the operating condition set, and obtain the stability margin prediction set based on the operating condition set under the margin prediction model. The safe operation analysis module is used to construct the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the stability margin prediction set, and to perform safe operation analysis based on the low-frequency oscillation safety domain boundary.

[0071] The above embodiments, based on the tidal current feasible domain injection space and the nodal admittance matrix, iteratively solve the tidal current feasible domain using the Newton-Raphson method, which can reduce misjudgments caused by inconsistencies with the electrical state of the offshore wind system. Subsequently, based on this tidal current feasible domain, stability criteria are determined, and a stability margin set is obtained by generating an operating sample set using the Monte Carlo method. Based on this stability margin set and the operating sample set, hyperparameter selection is performed to train and optimize a preset nonlinear neural network for early shutdown strategy, enabling the trained margin prediction model to effectively learn the complex nonlinear mapping relationship from operating conditions to stability margin, thereby improving the accuracy of subsequent stability margin predictions. Subsequently, the active power injection space is densely sampled under the tidal current feasible domain to generate an operating condition set, and the stability margin prediction set is obtained using the above margin prediction model. This stability margin prediction set can accurately distinguish between safe and unstable regions, and the safety operation analysis of the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system constructed based on this stability margin prediction set can effectively assess the stability of the offshore wind system and improve the accuracy of safety operation analysis of grid-connected offshore wind farms.

[0072] It is understood that the above system embodiments correspond to the method embodiments of the present invention, and can implement the grid-connected offshore wind farm safety operation analysis method provided by any of the above method embodiments of the present invention. It should be noted that the system embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between modules indicates that they have a communication connection, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without creative effort.

[0073] Based on the above embodiments of the safe operation analysis method for grid-connected offshore wind farms, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the safe operation analysis method for grid-connected offshore wind farms according to any embodiment of the present invention.

[0074] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0075] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0076] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0077] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the grid-connected offshore wind farm safety operation analysis method described in any of the above-described method embodiments of the present invention.

[0078] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0079] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for analyzing the safe operation of a grid-connected offshore wind farm, characterized in that, Includes the following steps: Obtain the network topology and operating parameters of the grid-connected offshore wind farm, and establish an admittance model of the offshore wind farm and transmission line based on the network topology and operating parameters; Based on the network topology and operating parameters, a power flow feasible region injection space and a node admittance matrix are constructed, and the power flow feasible region is obtained based on the power flow feasible region injection space and the node admittance matrix. Stability criteria are obtained based on the admittance model of the offshore wind farm and transmission line. The active power injection space and training sample number of the grid-connected offshore wind system are determined based on the current tidal feasible domain, and an operating sample set is generated based on the active power injection space and the training sample number. Based on the running sample set and the stability criterion, a stability margin set is obtained. Then, based on the stability margin set and the running sample set, hyperparameter selection and early stopping strategy training are performed under a preset nonlinear neural network to obtain a margin prediction model. The active power injection space is densely sampled under the power flow feasible domain to obtain the operating condition set, and the stability margin prediction set is obtained under the margin prediction model based on the operating condition set. The low-frequency oscillation safety domain boundary of the grid-connected offshore wind system is constructed based on the stability margin prediction set, and a safety operation analysis is performed based on the low-frequency oscillation safety domain boundary.

2. The method for analyzing the safe operation of a grid-connected offshore wind farm according to claim 1, characterized in that, The stability criterion obtained based on the admittance model of the offshore wind farm and transmission line includes: The difference between the offshore wind farm and the transmission line admittance model is processed to obtain the equivalent admittance model of the grid-connected offshore wind system, and the admittance model Schul complement is obtained by performing a Schul transform based on the equivalent admittance model of the grid-connected offshore wind system. The range of oscillation frequencies that makes the admittance model's Shur complement greater than zero is determined, and this range of oscillation frequencies is used as a stability criterion.

3. The method for analyzing the safe operation of a grid-connected offshore wind farm according to claim 1, characterized in that, The process involves obtaining a stability margin set based on the running sample set and the stability criterion, and then performing hyperparameter selection and early stopping strategy training under a preset nonlinear neural network based on the stability margin set and the running sample set to obtain a margin prediction model, including: A stability margin set is obtained based on the running sample set and the stability criterion; The stability margin set and the running sample set are integrated to obtain the sea breeze system training dataset. The sea breeze system training dataset is then normalized and sample balanced to generate the sea breeze system training set and the sea breeze system validation set. Based on the sea breeze system training set and the sea breeze system validation set, hyperparameter selection and early termination strategy training are performed under a preset nonlinear neural network to obtain an optimized nonlinear neural network, which is then used as a margin prediction model.

4. The method for analyzing the safe operation of a grid-connected offshore wind farm according to claim 3, characterized in that, The process of constructing a low-frequency oscillation safety domain boundary for the grid-connected offshore wind system based on the stability margin prediction set, and performing safety operation analysis based on the low-frequency oscillation safety domain boundary, includes: The set of stability margin predictions is used to select a set of safe operating stability margins under a preset safe operating margin condition, and the safe operating stability margin set is used to determine the safe domain of low-frequency oscillations. Based on the low-frequency oscillation safety domain, the boundary of the low-frequency oscillation safety domain of the grid-connected offshore wind system is constructed, and a safety operation analysis is performed based on the boundary of the low-frequency oscillation safety domain of the grid-connected offshore wind system.

5. The method for analyzing the safe operation of a grid-connected offshore wind farm according to claim 1, characterized in that, The step of constructing the power flow feasible domain injection space and node admittance matrix based on the network topology and operating parameters, and obtaining the power flow feasible domain based on the power flow feasible domain injection space and node admittance matrix, includes: A node admittance matrix is ​​formed based on the network topology and operating parameters; Based on the network topology and the operating status of each wind turbine in the operating parameters, a power flow feasible domain injection space is constructed.

6. The method for analyzing the safe operation of a grid-connected offshore wind farm according to claim 5, characterized in that, The step of constructing a power flow feasible region injection space and a node admittance matrix based on the network topology and operating parameters, and obtaining the power flow feasible region based on the power flow feasible region injection space and the node admittance matrix, includes: Based on the node admittance matrix, each power group in the power flow feasible region injection space is set at a preset node voltage initial value and Newton's iteration method is used to iterate until the preset safe operation convergence threshold is met, and the power flow result is obtained. Based on the power flow results, the feasible power flow domains under different operating conditions are obtained under a preset given node voltage constraint.

7. A safety operation analysis system for grid-connected offshore wind farms, characterized in that, A method for analyzing the safe operation of a grid-connected offshore wind farm as described in any one of claims 1 to 6, comprising: The initial construction module of the offshore wind system is used to obtain the network topology and operating parameters of the grid-connected offshore wind farm, and to establish an admittance model of the offshore wind farm and transmission line based on the network topology and operating parameters. The feasible power flow domain acquisition module is used to construct a feasible power flow domain injection space and a node admittance matrix based on the network topology and operating parameters, and to obtain the feasible power flow domain based on the feasible power flow domain injection space and the node admittance matrix. The sample generation module is used to obtain stability criteria based on the admittance model of the offshore wind farm and transmission line; determine the active power injection space and training sample number of the grid-connected offshore wind system according to the tidal current feasible domain; and generate an operating sample set according to the active power injection space and the training sample number. The margin prediction model building module is used to obtain a stable margin set based on the running sample set and the stability criterion, and to perform hyperparameter selection and early stopping strategy training processing under a preset nonlinear neural network based on the stable margin set and the running sample set to obtain the margin prediction model. The stability margin prediction module is used to perform intensive sampling of the active power injection space under the power flow feasible domain to obtain the operating condition set, and obtain the stability margin prediction set based on the operating condition set under the margin prediction model. The safe operation analysis module is used to construct the low-frequency oscillation safety domain boundary of the grid-connected offshore wind system based on the stability margin prediction set, and to perform safe operation analysis based on the low-frequency oscillation safety domain boundary.

8. A terminal device, characterized in that, The method includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a method for analyzing the safe operation of a grid-connected offshore wind farm as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform a method for analyzing the safe operation of a grid-connected offshore wind farm as described in any one of claims 1 to 6.

Images