New energy converter low-pass control dominated voltage oscillation parameter feasible region construction method, system and equipment and medium

Abstract

The invention belongs to the technical field of converter stability analysis and control, and discloses a new energy converter low-pass control dominated voltage oscillation parameter feasible region construction method, system and equipment and a medium, so as to solve the problem that accurate description of a system stability boundary is difficult to realize in the prior art. The method comprises the following steps: establishing a converter grid-connected system model considering low-voltage ride-through control, and generating a data set comprising a multi-parameter combination and a corresponding repeated low-voltage ride-through state tag based on the model; taking the data set as training data, and training the repeated low-crossing risk discrimination model by adopting a fast kernel learning algorithm to obtain a trained model; and inputting converter grid-connected system parameters needing to be analyzed into the trained model for classification prediction, and drawing a parameter region without repeated low-voltage crossing based on a result to obtain a multi-parameter feasible region schematic diagram. According to the method, the technical blank of carrying out accurate quantification and visualization on the repeated low-penetration stable boundary in a complex parameter scene is filled.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of converter stability analysis and control, and specifically relates to a method, system, device and medium for constructing a voltage oscillation parameter feasible region dominated by low penetration control of a new energy converter. BACKGROUND

[0002] With the expansion of the scale of new energy grid connection, power electronic converters as the key interface have new stability problems due to the interaction between their complex control dynamics and the power grid. Specifically, to ensure different operation modes when the power grid experiences a temporary voltage drop, new energy converters are generally equipped with a low voltage ride through (LVRT) function. However, in engineering practice, due to factors such as controller parameter configuration, the converter may appear in the phenomenon of repeated switching at the edge of the LVRT state, i.e., "repeated low voltage penetration". This phenomenon can cause sustained oscillation of the grid-connected point voltage, seriously threatening the safe and stable operation of the system. Currently, research on this problem mainly focuses on mechanism analysis and influence factor exploration: on the one hand, by establishing a mathematical model, the dynamic process of voltage and current fluctuations is analyzed; on the other hand, the correlation between grid short-circuit capacity, unit output level, etc. and oscillation behavior is explored. It is generally agreed that the output power of the converter, the system strength of the grid-connected point, and the current loop control parameters are the key variables affecting the occurrence of repeated low penetration. At the same time, data-driven analysis methods have been increasingly widely applied in the field of stability evaluation due to their advantages in dealing with complex nonlinear problems.

[0003] In terms of stability domain or feasible region construction methods, traditional methods mainly rely on polynomial fitting and boundary search methods. However, polynomial fitting is highly dependent on system parameters, has high algorithm sensitivity, and lacks generalization ability; while boundary search methods have the problems of heavy computational burden and low efficiency, making it difficult to meet the application requirements of online or rapid evaluation. Although data-driven methods have achieved good results in stability analysis and feasible region construction of a single operating mode, their application to this specific and complex stability problem of repeated low penetration is still insufficient, and there is a lack of systematic solutions. More importantly, repeated low penetration is essentially a large disturbance stability problem, with significant strong nonlinear characteristics, making it difficult to directly apply classical linear system tools such as state space eigenvalue analysis, Nyquist stability criterion, etc. based on linearized models for accurate quantitative analysis and boundary quantification.

[0004] In summary, existing research on the feasible region of new energy converters mainly uses the control variable method, focusing on revealing the trend relationship between repeated low penetration and a single parameter, but it is difficult to accurately characterize the stability boundary of the system, especially lacking comprehensive analysis methods and feasible region construction methods for repeated low penetration characteristics in the scenario where multiple parameters are coupled and interact with each other. Summary of the Invention

[0005] Based on the aforementioned shortcomings and deficiencies in the prior art, one of the objectives of this invention is to at least solve one or more of the aforementioned problems in the prior art. In other words, one of the objectives of this invention is to provide a method, system, device, and medium for constructing the feasible domain of voltage oscillation parameters dominated by low-voltage control in new energy converters that meets one or more of the aforementioned requirements, so as to achieve rapid and accurate characterization of multi-parameter stability domains.

[0006] To achieve the above-mentioned objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for constructing the feasible region of voltage oscillation parameters dominated by low-down control in a new energy converter, comprising the following steps: S1. Establish a converter grid-connected system model that takes into account low voltage ride-through control, and generate a dataset containing multiple parameter combinations and their corresponding repeated low voltage ride-through state labels based on the model. S2. Using the dataset as training data, a repeated low-crossing risk discrimination model is trained using a fast kernel learning algorithm to obtain the trained model. S3. Input the parameters of the converter grid-connected system to be analyzed into the trained model for classification and prediction, and draw the parameter region where repeated low-voltage crossings did not occur based on the results to obtain a schematic diagram of the multi-parameter feasible region.

[0007] As a preferred embodiment, the converter grid-connected system model considering low-voltage ride-through control includes: The LVRT control module is used to switch between normal operation mode and LVRT mode based on the voltage amplitude at the common coupling point. The current inner loop control module is used to achieve tracking control of the converter output current; The phase-locked loop module is used to track the phase information of the grid voltage in real time, so as to achieve synchronization between the converter and the grid.

[0008] As a preferred approach, the data used in step S1 to generate a dataset containing multiple parameter combinations and their corresponding repeated low-voltage crossing state labels based on the converter grid-connected system model includes: Short-circuit ratio SCR Current inner loop proportional control parameters K p Voltage feedforward delay factor T s LVRT proportional coefficient, and the active power output of the converter in steady state. P and reactive power Q .

[0009] As a preferred approach, the fast kernel learning algorithm employs a polynomial kernel function.

[0010] As a preferred embodiment, the method of training the repeated low-crossing risk discrimination model using a fast kernel learning algorithm includes the following steps: S21. Execute the first part of the algorithm based on the given polynomial degree s: input the training data and the polynomial degree s; select a set of mutually independent center points from the input data; calculate the kernel matrix based on the center points; solve the parameter vector of the repeated low-crossing risk discrimination model through a generalized inverse operation, and output the training results; S22. Determine the optimal polynomial degree s' and execute the second part of the algorithm: Divide the training data into a first subset and a second subset; Using the first subset and different polynomial degrees, execute the first part of the algorithm to obtain a set of training model sequences; Using the second subset, select the final polynomial degree s' and its corresponding training model from the model sequence by solving an optimization problem, as the trained model.

[0011] As a preferred approach, the objective function of the optimization problem is: , In the formula, The degree of the obtained optimal polynomial. This represents the parameter values ​​that minimize the objective function. Represents a symbolic function. The number of data points in the training set. Input samples for the dataset, Output samples for the dataset. For the second subset of data, This represents the total number of basis functions used in model training. For the first subset of datasets and the corresponding polynomial degree The resulting training model, Let be the range of values ​​for the degree of the polynomial.

[0012] As a preferred embodiment, the step of obtaining the multi-parameter feasible region schematic diagram includes: The parameters of the converter grid-connected system to be analyzed are input into the trained model for classification and prediction. Based on the prediction results, the regions formed by all parameter combinations that were determined not to have repeatedly crossed low ground are plotted to generate the multi-parameter feasible region schematic diagram.

[0013] In a second aspect, the present invention provides a feasible domain construction system for voltage oscillation parameters dominated by low-voltage control of a new energy converter, for implementing the feasible domain construction method for voltage oscillation parameters as described in the first aspect.

[0014] Thirdly, the present invention provides an electronic device, the computer device including a memory, a processor and a computer program, wherein when the computer program is executed by the processor, it implements the voltage oscillation parameter feasible domain construction method as described in the first aspect.

[0015] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the voltage oscillation parameter feasible domain construction method as described in the first aspect.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1. The proposed construction method based on fast kernel learning in this invention significantly shortens computation time by calculating the kernel matrix using center points and performing a generalized inverse operation, thus omitting the complex parameter iteration optimization process in classical kernel learning. This method maintains excellent discriminative performance even in small-sample scenarios, effectively reducing the costs associated with large-scale simulations and data acquisition, and is more suitable for practical engineering scenarios requiring rapid evaluation and online application.

[0017] 2. This invention, through the systematic construction of a dataset, can simultaneously consider and analyze the interactive effects of multiple key variables such as short-circuit ratio, control parameters, and operating power, clearly revealing the stability laws under the coupling effect of multiple parameters. This overcomes the limitation of traditional control variable methods, which can only analyze the trend of a single parameter, and effectively fills the technical gap in accurately quantifying and visualizing the stability boundary of repeated low-voltage crossings in complex parameter scenarios, ensuring the comprehensiveness and depth of stability analysis.

[0018] 3. Repeated low-passing is a strongly nonlinear stability problem under large disturbances. This invention utilizes a data-driven kernel learning method, which can directly capture and model its complex nonlinear dynamic characteristics without linearizing the system model. This overcomes the limitations of traditional tools such as eigenvalue analysis based on linearized models in this type of problem, and improves the accuracy and reliability of the analysis of complex stability problems.

[0019] Further or more detailed beneficial effects will be described in conjunction with specific embodiments in the detailed implementation. Attached Figure Description

[0020] 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 of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1This is a flowchart illustrating the feasible region construction method for voltage oscillation parameters according to Embodiment 1 of the present invention.

[0022] Figure 2 This is a schematic diagram of the converter grid-connected system model described in Embodiment 1 of the present invention.

[0023] Figure 3 This is a structural diagram of the electronic device described in Embodiment 3 of the present invention.

[0024] Figure 4 This is the feasible region scatter plot described in Embodiment 5 of the present invention.

[0025] Figure 5 This is the convex hull diagram described in Embodiment 5 of the present invention.

[0026] Icon labels: 300. Electronic devices; 301. Processor; 302. Communication bus; 303. User interface; 304. Network interface; 305. Memory. Detailed Implementation

[0027] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0028] In the following description, several embodiments of the present invention are provided. Different embodiments can be substituted or combined. Therefore, the present invention can also be considered to include all possible combinations of the same and / or different embodiments described. Thus, if one embodiment includes features A, B, and C, and another embodiment includes features B and D, then the present invention should also be considered to include embodiments containing one or more other possible combinations of A, B, C, and D, even if such embodiments are not explicitly described in the following text.

[0029] The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made to the function and arrangement of the described elements without departing from the scope of the invention. Various processes or components may be appropriately omitted, substituted, or added to the various examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.

[0030] To facilitate a better understanding of the embodiments of the present invention, its application scenarios will be explained before providing a detailed explanation of the specific implementation methods.

[0031] The feasible region construction method for voltage oscillation parameters described in the embodiments of this specification is applied to the planning, design, and operation analysis process of new energy power generation grid-connected systems. In these scenarios, the application of the feasible region construction method for voltage oscillation parameters aims to: 1. Guide controller parameter tuning: During the design phase of converter control systems in new energy power plants such as wind farms and photovoltaic power stations, provide a safe and stable parameter configuration range for LVRT control parameters, current loop control parameters, etc., to avoid repeated low-voltage ride-through risks caused by improper parameter settings from the source.

[0032] 2. Assessing system access stability: When power grid dispatching or planning departments assess the impact of new large-scale renewable energy access on local power grid stability, they can quickly plot the operational feasibility domain under different access point strengths (short-circuit ratios), providing a quantitative basis for access scheme decisions.

[0033] 3. Conduct safety and stability early warning and analysis: As an online or offline stability analysis tool, it is used to scan the current or expected operating conditions, quickly identify whether there is a risk of repeatedly crossing the stability boundary, thereby achieving early warning and preventive control.

[0034] 4. Optimize system operation mode: Provide station operators with safe operating boundaries under different grid conditions and output levels, thereby optimizing power generation plans and operation strategies while ensuring system stability.

[0035] The following is a brief explanation of the low-voltage ride-through control, converter grid-connected system model, fast kernel learning algorithm, polynomial kernel function, repeated low-voltage ride-through risk discrimination model, and multi-parameter feasible region involved in several embodiments of this specification: Low-voltage ride-through control refers to the mandatory operational capability of a new energy converter to maintain grid-connected operation for a period of time without disconnecting from the grid when a grid fault causes a voltage drop at the grid connection point. This invention focuses particularly on the phenomenon of "repeated low-voltage ride-through" caused by dynamic control interactions during voltage recovery.

[0036] The converter grid-connected system model is a mathematical model established for simulation analysis and dataset generation in accordance with this invention. This model accurately accounts for the dynamic characteristics of the low-voltage ride-through control module, the current inner loop control module, the phase-locked loop module, and the grid-side impedance, and is the basis for reproducing repeated low-voltage ride-through phenomena and generating sample data.

[0037] Fast kernel learning algorithms map data to a high-dimensional feature space for linear analysis based on kernel functions. Their "fast" characteristic is reflected in the fact that model parameters can be solved directly through a single generalized inverse operation, avoiding the convergence problem that may occur in traditional iterative optimization and significantly improving computational efficiency.

[0038] A polynomial kernel function is a type of kernel function that controls the complexity of the model by using the polynomial degree. It can effectively capture the complex nonlinear relationship between input parameters and the risk of repeated low-passes.

[0039] The repeated low-voltage risk discrimination model refers to a binary classification model trained using the fast kernel learning algorithm. This model takes a combination of multiple parameters of the converter grid-connected system as input and outputs a binary discrimination result (1 or 0) to predict whether the system will experience repeated low-voltage phenomena under the specified parameter conditions.

[0040] The multi-parameter feasible region refers to the region comprised of all parameter points deemed "stable" (i.e., not subject to repeated low-voltage crossings) by the aforementioned repeated low-voltage crossing risk discrimination model within a high-dimensional parameter space composed of multiple key system parameters (such as short-circuit ratio, control parameters, output power, etc.). This feasible region is presented in the form of a schematic diagram, clearly defining the parameter boundaries for the safe and stable operation of the system.

[0041] Example 1: like Figure 1 As shown, this embodiment provides a method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage ride-through control in a new energy converter, including the following steps: Step S1: Establish a converter grid-connected system model that takes into account low voltage ride-through control, and generate a dataset containing multiple parameter combinations and their corresponding repeated low voltage ride-through state labels based on the model.

[0042] Specifically, the structure of the converter grid-connected system model considering low-voltage ride-through control can be found in [reference needed]. Figure 2 It includes the following modules: 1. The LVRT control module is used to switch between normal operation mode and LVRT mode based on the amplitude of the common coupling point voltage. When the amplitude of the common coupling point voltage is less than 0.9... p.u When the system enters LVRT mode, the reference current values ​​for normal operation mode and LVRT mode are shown in formulas (1) and (2), respectively: (1), (2), In equations (1) and (2), and These are the inner current loops. d , q Shaft current reference value, This is the LVRT scaling factor. This represents the voltage amplitude at point PCC.

[0043] 2. The inner current loop control module is used to achieve tracking control of the converter output current. The dynamic equation of the inner current loop is: (3), In equation (3), U sd and U sq These are the output voltages of the converter. d , q Axial components, I d and I q These are the currents at PCC. d , q Axial components, U d and U q The voltage at PCC is respectively d , q Axial components, G CC ( s )= k CCp + k CCi / s The transfer function for the inner current loop PI controller. k ccp The proportional gain of the inner-loop PI controller. k cci Integral coefficient of the current inner loop PI controller s For the Laplace operator, It is the power frequency angular frequency. For filtering inductors, G VF ( s )=1 / ( Ts +1) is the voltage feedforward filter transfer function, where T This represents the filter delay coefficient.

[0044] 3. The phase-locked loop (PLL) module is used to track the phase information of the grid voltage in real time, achieving synchronization between the converter and the grid. Its control equation is: (4), In equation (4), and These represent the rotational speed and phase angle of the PLL coordinate system relative to the system's synchronous rotating coordinate system, respectively. This is the transfer function of the phase-locked loop PI controller. Represents the Laplace operator; The dynamic equation of the filter inductor is: (5); Ignoring the filter capacitor, the dynamic equation of the transmission line is: (6), In equation (6), E The external power grid voltage is infinite. It is the power frequency angular frequency. For line inductance, This represents the line resistance. With this, the model of the converter grid-connected system considering LVRT control is complete.

[0045] Step S2: Using the dataset as training data, train the repeated low-crossing risk discrimination model using the fast kernel learning algorithm to obtain the trained model.

[0046] Specifically, in constructing the dataset, this embodiment selected the following key parameters as model input features: short-circuit ratio. SCR Current inner loop proportional control parameters K p Voltage feedforward delay factor T s LVRT proportional gain, and the active power output of the converter in steady state. P and reactive power Q This multi-parameter systematic construction scheme ensures that the dataset can comprehensively cover the main factors affecting the risk of repeated low-passing, thus laying the foundation for training a discriminative model with good representativeness and generalization ability.

[0047] Based on the above parameter combinations, a total of 600 valid data samples were generated through uniform random sampling within a reasonable range of parameter values. To accurately reflect the risk of repeated low-voltage wear, sample labels were defined: 200 samples were labeled "1", representing repeated low-voltage wear that occurred; the remaining 400 samples were labeled "0", representing that repeated low-voltage wear did not occur. This dataset covers different grid strengths, control parameters, and power output conditions, thus comprehensively characterizing various potential scenarios of repeated low-voltage wear risk.

[0048] Furthermore, this embodiment employs a uniform random sampling method to effectively avoid excessive concentration of data within a certain parameter range, ensuring the uniformity of the dataset distribution. Subsequently, the entire dataset is proportionally divided into a training set and a test set, where the training set is used for model training, and the test set is used to objectively evaluate the generalization ability and classification accuracy of the trained model on unseen data.

[0049] Specifically, the fast kernel learning algorithm described in this embodiment uses a polynomial kernel function.

[0050] In kernel learning algorithms, the core principle lies in the final learned model. f ( x ) can be represented as a kernel function The linear combination of is expressed as shown in equation (7): (7), In equation (7), The parameters to be trained are column vectors. This is the input data for the dataset.

[0051] Based on this, the polynomial kernel function used in this embodiment is defined as: (8), In equation (8), ≥1 represents the degree of the polynomial; Furthermore, the function in equation (8) satisfies the following kernel matrix K It can only be used as a kernel function when it meets the condition of being semi-positive definite.

[0052] (9).

[0053] Classical kernel learning algorithms determine model parameters (Equation (7)) by solving an optimization problem, the objective of which is to minimize the regularization term. Risk of experience The sum of these, the mathematical expression of the optimization objective is: (10) In equation (10), For the training set, The number of data points in the training set. λ The coefficient of the regularization term, For the kernel matrix K The norm of the established regenerative kernel Hilbert space.

[0054] The regularization term in equation (10) It plays a crucial role: when the Hilbert space is infinite-dimensional, it ensures the existence and uniqueness of solutions to optimization problems; when the space is finite-dimensional, it allows for the adjustment of parameters... λ This allows for control over model complexity, thereby achieving a balance between fitting bias and variance and improving the model's generalization ability. Theoretical research indicates that for a polynomial kernel function, when the number of samples... and polynomial degree When it increases, the coefficient of the regularization term λ It can decay exponentially or at any rate.

[0055] Based on the definition of the polynomial kernel function in the classical kernel learning algorithm, this embodiment adopts a more efficient fast kernel learning algorithm in the optimization process, avoiding the convergence problem that may occur in traditional iterative methods. Its execution flow is mainly divided into the following two parts: Part 1: Given the polynomial degree and training set data, solve the model directly using the following steps: 1. Input: Training set data and polynomial degree ∈N (N represents the set of non-negative integers).

[0056] 2. Select the center point: Let (C is the number of combinations) represents the number of center points selected from the test set, and a set of independent input data. This represents the center point of the input data space.

[0057] 3. Construct the kernel matrix: based on the center point The positive semi-definite kernel matrix can be calculated. .

[0058] 4. Generalized Inverse Solution: Let the output vector of the test set be... The vector of undetermined parameters in the training model is Then there is ,in It represents the generalized inverse operation.

[0059] 5. Output: Training results The algorithm terminates.

[0060] The core advantage of this section lies in its ability to handle any given polynomial degree. The training result can be obtained through only one generalized inverse operation. This avoids the non-convergence problem that may occur during parameter optimization, bypasses the traditional time-consuming iterative optimization process, and significantly improves computational efficiency and reliability.

[0061] Part Two: Determining the Optimal Polynomial Degree Perform the following steps: 1. Data partitioning: Divide the training set data... Divide it into two, and you get the base number. sub-data and base Subdataset .

[0062] 2. Generate model sequences: and Substituting each value into the first part of the fast kernel learning algorithm described above, we obtain a set of training model sequences. , where [·] represents rounding down.

[0063] 3. Select the optimal model: Utilize And the following optimization problem yields the final polynomial degree. And the corresponding training model.

[0064] (11), In equation (11), The degree of the obtained optimal polynomial. This represents the parameter values ​​that minimize the objective function. The sign function is used to determine the sign of the difference between the basis function output and the target value, thus indicating the direction of the match between the model prediction and the true result. In practical applications, care should be taken to ensure that the selected input data points are accurate. Since the input matrices are independent of each other, when the dimension of the input matrix is ​​high, a projection algorithm can be used to find a set of orthogonal bases with the largest distance between them from a large amount of input data. The number of data points in the training set. Input samples for the dataset, Output samples for the dataset. For the second subset of data, This represents the total number of basis functions used in model training. For the first subset of datasets and the corresponding polynomial degree The resulting training model, Let be the range of values ​​for the degree of the polynomial.

[0065] Step S3: Input the required converter grid-connected system parameters into the trained model for classification and prediction, and draw the parameter region where repeated low-voltage crossings did not occur based on the results to obtain a multi-parameter feasible region diagram. The specific operation is as follows: After the model training is completed, its performance needs to be verified first. To this end, the complete dataset constructed in step S1 is divided into a training set and a test set. In this embodiment, 80% of the data is used for model training, and the remaining 20% ​​is used as the test set. The training set is used to enable the model to learn the complex mapping relationship between input features and repeated low-risk classification; the test set is specifically used to evaluate the generalization ability and classification accuracy of the trained model, thereby objectively confirming its reliability in practical applications.

[0066] After verifying that the model performance meets the requirements, feasible region characterization is performed. The required converter grid-connected system parameter combination is used as the model input, and the trained discriminant model is used for classification and prediction. In this model, an output result of 1 indicates that the equipment will enter a repeated low-voltage state under the operating conditions of this parameter set; an output result of 0 indicates that the equipment is operating stably and no repeated low-voltage state has occurred.

[0067] Furthermore, by systematically adjusting a set of continuous parameters to be studied (such as short-circuit ratio, control parameters, etc.), a large number of parameter combinations covering the parameter space are generated, and batch prediction and discrimination are performed using the model. Finally, by visualizing the region formed by all parameter points that are judged to be stable by the model (i.e., output is 0) in the parameter space, a multi-parameter feasible region diagram that clearly represents the safe operation boundary of the system can be obtained.

[0068] Example 2: This embodiment provides a feasible region construction system for voltage oscillation parameters dominated by low-voltage control in a new energy converter, which is used to implement the feasible region construction method for voltage oscillation parameters as described in Embodiment 1.

[0069] Example 3: like Figure 3 As shown, this embodiment provides an electronic device, which may include: at least one processor, at least one network interface, a user interface, a memory, and at least one communication bus.

[0070] The communication bus can be used to enable communication between the various components mentioned above.

[0071] The user interface may include buttons, and optional user interfaces may also include standard wired interfaces and wireless interfaces.

[0072] The network interface may include, but is not limited to, Bluetooth modules, NFC modules, Wi-Fi modules, etc.

[0073] The processor may include one or more processing cores. It connects various parts of the electronic device via various interfaces and lines, executing instructions, programs, code sets, or instruction sets stored in memory, and accessing data stored in memory to perform various functions and process data. Optionally, the processor can be implemented using at least one hardware form of DSP, FPGA, or PLA. The processor may integrate one or more of the following: CPU, GPU, and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also be implemented as a separate chip without being integrated into the processor.

[0074] The memory may include RAM or ROM. Optionally, the memory may include a non-transitory computer-readable medium. The memory can be used to store instructions, programs, code, code sets, or instruction sets. The memory may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor. The memory, as a computer storage medium, may include an operating system, a network communication module, a user interface module, and application programs. The processor can be used to call the application programs stored in the memory and execute the steps of the voltage oscillation parameter feasible domain construction method mentioned in the foregoing embodiments.

[0075] Example 4: This embodiment provides a computer-readable storage medium storing instructions that, when executed on a computer or processor, cause the computer or processor to perform the above-described instructions. Figure 1 One or more steps in the illustrated embodiment. If the constituent modules of the above-described electronic device are implemented as software functional units and sold or used as independent products, they can be stored in the computer-readable storage medium.

[0076] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this specification are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in or transmitted through a computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., Digital Versatile Discs (DVDs)), or semiconductor media (e.g., Solid State Disks (SSDs)).

[0077] Those skilled in the art will understand that all or part of the processes in the method of Embodiment 1 described above can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks. Unless otherwise specified, the technical features of this embodiment and the implementation scheme can be combined arbitrarily.

[0078] Example 5: This embodiment aims to verify the effectiveness, feasibility, and practical value of the feasible region construction method for voltage oscillation parameters dominated by low-voltage control in a new energy converter, as described in this specification.

[0079] First, based on the parameter values ​​in Table 1 and the converter grid-connected system model structure described in Example 1, a numerical simulation model of the converter grid-connected system considering LVRT was built in Matlab / Simulink software.

[0080] Table 1: .

[0081] To simulate voltage disturbances that a real-world grid-connected renewable energy system might face, the simulation conditions are set as follows: At 1.0 second into the system's operation, a three-phase symmetrical short-circuit fault occurs in the simulated grid, causing the voltage amplitude at the point of common coupling to drop to [value missing]. 0.2 p.u. (i.e., instantaneous reduction) 0.8 p.u. The fault lasted for 0.1 seconds and was cleared after 1.1 seconds, with the voltage returning to 1. .0 p.u. The normal level.

[0082] This embodiment selects three key parameters that have a significant impact on the repeated low-voltage wear phenomenon, namely the short-circuit ratio. SCR Current inner loop proportionality coefficient K p With steady-state active power P This serves as the dimension characterizing the multi-parameter feasible region in this embodiment. The remaining parameters are set to fixed values, such as the voltage feedforward delay coefficient. T s The steady-state reactive power is 0.05. Q The value is 0.1, and the reactive current coefficient of LVRT is 1.5, in order to focus on studying the coupling effect of the above three parameters.

[0083] Subsequently, following the specific steps described in Example 1, a training dataset is first generated and a repeated low-crossing risk discrimination model based on fast kernel learning is trained. Then, the trained model is applied to the specific scenario of this example. This is achieved by systematically scanning the SCR... K p and P The parameter space generates a large number of parameter combinations, which are then input into the model for classification and prediction. An output of 0 indicates that the system is stable under this parameter combination and that repeated low-level crossings have not occurred.

[0084] Finally, all parameter points determined to be stable are visualized in a three-dimensional parameter space, and the resulting feasible region scatter plot is shown below. Figure 4 As shown, its corresponding convex hull boundary graph is as follows: Figure 5 As shown. Figure 5 The region enclosed by the convex hull is the multi-parameter feasible region under the disturbance conditions described in this embodiment, which can operate safely and avoid repeated low-level penetration.

[0085] The results show that the method of the present invention successfully achieves the processing of SCR, K p and P The stability boundary under the combined influence of three parameters can be quickly and accurately characterized. Furthermore, the system stability law can be intuitively summarized from the generated feasible region: the current inner loop proportionality coefficient... K p The larger the steady-state active power PThe smaller the SCR (Short Circuit Ratio), the larger the SCR, the stronger the system stability and the less likely it is to experience voltage oscillations caused by repeated low-circuit breakdowns.

[0086] In summary, this embodiment verifies the effectiveness of the method described in this specification in constructing a multi-parameter feasible region through specific simulation experiments, and proves that the feasible region can clearly reveal the stability law under the coupling effect of multiple parameters, which has clear guiding significance for parameter tuning and safe operation in engineering practice.

[0087] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0088] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0089] The above description is merely an exemplary embodiment of the present invention and should not be construed as limiting the scope of the invention. Any equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of embodiments of the invention upon considering the specification and practicing the disclosure herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not described herein. The specification and embodiments are to be considered exemplary only, and the scope and spirit of the invention are defined by the claims.

Claims

1. A method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage control in a new energy converter, characterized in that, Including the following steps: S1. Establish a converter grid-connected system model that takes into account low voltage ride-through control, and generate a dataset containing multiple parameter combinations and their corresponding repeated low voltage ride-through state labels based on the model. S2. Using the dataset as training data, a repeated low-crossing risk discrimination model is trained using a fast kernel learning algorithm to obtain the trained model. S3. Input the parameters of the converter grid-connected system to be analyzed into the trained model for classification and prediction, and draw the parameter region where repeated low-voltage crossings did not occur based on the results to obtain a schematic diagram of the multi-parameter feasible region.

2. The method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage control in a new energy converter according to claim 1, characterized in that, The converter grid-connected system model considering low-voltage ride-through control includes: The LVRT control module is used to switch between normal operation mode and LVRT mode based on the voltage amplitude at the common coupling point. The current inner loop control module is used to achieve tracking control of the converter output current; The phase-locked loop module is used to track the phase information of the grid voltage in real time, so as to achieve synchronization between the converter and the grid.

3. The method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage drive control in a new energy converter according to claim 2, characterized in that, The data used in step S1 to generate the dataset containing multiple parameter combinations and their corresponding repeated low-voltage crossing state labels based on the converter grid-connected system model includes: Short-circuit ratio SCR Current inner loop proportional control parameters K p Voltage feedforward delay factor T s LVRT proportional coefficient, and the active power output of the converter in steady state. P and reactive power Q .

4. The method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage control in a new energy converter according to claim 3, characterized in that: The fast kernel learning algorithm uses a polynomial kernel function.

5. The method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage control in a new energy converter according to claim 4, characterized in that, The method of training a repeated low-pass risk discrimination model using a fast kernel learning algorithm includes the following steps: S21. Execute the first part of the algorithm based on the given polynomial degree s: input the training data and the polynomial degree s; select a set of mutually independent center points from the input data; calculate the kernel matrix based on the center points; solve the parameter vector of the repeated low-crossing risk discrimination model through a generalized inverse operation, and output the training results; S22. Determine the optimal polynomial degree s' and execute the second part of the algorithm: Divide the training data into a first subset and a second subset; Using the first subset and different polynomial degrees, execute the first part of the algorithm to obtain a set of training model sequences; Using the second subset, select the final polynomial degree s' and its corresponding training model from the model sequence by solving an optimization problem, as the trained model.

6. The method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage drive control in a new energy converter according to claim 5, characterized in that, The objective function of the optimization problem is: ， In the formula, The degree of the obtained optimal polynomial. This represents the parameter values ​​that minimize the objective function. Represents a symbolic function. The number of data points in the training set. Input samples for the dataset, Output samples for the dataset. For the second subset of data, This represents the total number of basis functions used in model training. For the first subset of datasets and the corresponding polynomial degree The resulting training model, Let be the range of values ​​for the degree of the polynomial.

7. The method for constructing the feasible region of voltage oscillation parameters dominated by low-voltage control in a new energy converter according to claim 6, characterized in that, The steps for obtaining the multi-parameter feasible region schematic diagram include: The parameters of the converter grid-connected system to be analyzed are input into the trained model for classification and prediction. Based on the prediction results, the regions formed by all parameter combinations that were determined not to have repeatedly crossed low ground are plotted to generate the multi-parameter feasible region schematic diagram.

8. A system for constructing the feasible region of voltage oscillation parameters dominated by low-voltage control in a new energy converter, characterized in that, Used to implement the feasible region construction method for voltage oscillation parameters as described in any one of claims 1 to 7.

9. A computer device, the computer device comprising a memory, a processor, and a computer program, characterized in that, When the computer program is executed by the processor, it implements the voltage oscillation parameter feasible region construction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the voltage oscillation parameter feasible region construction method as described in any one of claims 1 to 7.

Images