Construction method and equipment of multi-machine equivalent model of wind power plant and medium

Abstract

The invention relates to the technical field of electric power system modeling, in particular to a construction method and device for a multi-machine equivalent model of a wind power plant and a medium, and the method comprises the steps: carrying out the overall parameter identification of a grid-connected model based on an improved multi-target Lichberg algorithm, and obtaining a control parameter; inputting the control parameters into the grid-connected model to obtain a reactive current track and a dq-axis current component of the doubly-fed asynchronous wind driven generator of the wind power plant, and clustering the doubly-fed asynchronous wind driven generator according to the reactive current track and the dq-axis current component to obtain a plurality of class clusters; and performing parameter equivalence and network equivalence on the doubly-fed asynchronous wind power generator in each class cluster to obtain an equivalent doubly-fed asynchronous wind power generator corresponding to each class cluster, and constructing a multi-machine equivalent model of the wind power plant by using the equivalent doubly-fed asynchronous wind power generator and the static var generator. According to the invention, the deviation between the simulation result of the equivalent model of the wind power plant and the real dynamic behavior of the wind power plant is reduced.

Description

Technical Field

[0001] This invention relates to the field of power system modeling technology, specifically to a method, equipment, and medium for constructing a multi-machine equivalent model of a wind farm. Background Technology

[0002] With the global trend towards a clean and low-carbon energy transition, wind power is increasingly penetrating the power grid. Doubly fed induction generators (DFIGs) have become the mainstream turbine type for wind farms due to their advantages such as variable speed and constant frequency, and independent active and reactive power regulation. However, with the large-scale grid connection of DFIG wind farms, their interaction with the grid becomes increasingly complex. The transient response of DFIGs under fault conditions directly affects grid security, making transient stability analysis of DFIG wind farms a critical technical requirement. However, existing research mostly focuses on modeling under single faults, with insufficient model construction and characteristic analysis for cascading faults. This makes it difficult to accurately characterize the dynamic behavior during fault switching, leading to significant discrepancies between the analysis results and actual conditions.

[0003] Parameter identification is crucial for modeling, but existing methods have significant limitations: First, they often focus on individual DFIGs or Static Var Generators (SVG), neglecting their collaborative coupling under fault conditions and failing to reflect overall transient characteristics. Second, they are mostly conducted under single fault scenarios (such as fixed voltage drops), failing to consider the randomness of actual faults, resulting in poor parameter universality and further reducing model reliability. Furthermore, wind farms contain dozens to hundreds of DFIG units; while full-detail modeling can preserve individual characteristics, its computational complexity is extremely high, making it difficult to meet engineering efficiency requirements. Therefore, simulation results from equivalent models of wind farms cannot reflect the true dynamic behavior of wind farms, resulting in large errors in the analysis and hindering guidance for the safe operation of wind farms and the power grid. Summary of the Invention

[0004] To address the significant discrepancy between simulation results of equivalent wind farm models and the actual dynamic behavior of wind farms, this invention aims to provide a method, equipment, and medium for constructing multi-machine equivalent models of wind farms. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of the present invention provide a method for constructing a multi-machine equivalent model of a wind farm, comprising: constructing a grid-connected model of a doubly-fed asynchronous wind turbine generator and a static var generator (SVM) of the wind farm; performing overall parameter identification on the grid-connected model based on an improved multi-objective Lichtenberg algorithm to obtain control parameters of the rotor-side converter, grid-side converter, and SVM of the doubly-fed asynchronous wind turbine generator; inputting the control parameters into the grid-connected model to obtain the reactive current trajectory and dq-axis current component of the doubly-fed asynchronous wind turbine generator of the wind farm, and clustering the doubly-fed asynchronous wind turbine generators according to the reactive current trajectory and dq-axis current component to obtain multiple clusters; performing parameter equivalence and network equivalence on the doubly-fed asynchronous wind turbine generators in each cluster to obtain an equivalent doubly-fed asynchronous wind turbine generator corresponding to each cluster, and constructing a multi-machine equivalent model of the wind farm using the equivalent doubly-fed asynchronous wind turbine generator and the SVM.

[0006] Optionally, the overall parameter identification of the grid-connected model is performed based on the improved multi-objective Lichtenberg algorithm to obtain the control parameters of the rotor-side converter, grid-side converter, and static var generator of the doubly-fed asynchronous wind turbine. This includes: constructing a parameter vector from the parameters to be identified of the rotor-side converter, grid-side converter, and static var generator of the doubly-fed asynchronous wind turbine; determining the fitness function for parameter identification based on the simulation data of the grid-connected model and the measured data of the wind farm; and performing overall parameter identification of the grid-connected model based on the improved multi-objective Lichtenberg algorithm until the fitness function tends to a stable value.

[0007] Optionally, the overall parameter identification of the grid-connected model based on the improved multi-objective Lichtenberg algorithm until the fitness function tends to a stable value includes: performing preliminary identification of the parameter vector based on the improved multi-objective Lichtenberg algorithm to obtain preliminary identification results, thereby narrowing the range of high-quality solutions; determining the parameter trajectory sensitivity of each parameter in the parameter vector based on the change in the controlled power of the doubly-fed asynchronous wind turbine or static var generator corresponding to the parameter to be identified and the change in the step size of the parameter to be identified; dividing the parameter to be identified into a first sensitivity parameter and a second sensitivity parameter based on the parameter trajectory sensitivity, wherein the parameter trajectory sensitivity of the first sensitivity parameter is higher than that of the second sensitivity parameter; and re-identifying the first sensitivity parameter based on the preliminary identification results until the fitness function tends to a stable value, and then re-identifying the second sensitivity parameter until the fitness function tends to a stable value, thereby obtaining the control parameters.

[0008] Optionally, determining the parameter trajectory sensitivity of each parameter in the parameter vector based on the change in the controlled power of the doubly-fed asynchronous wind turbine or static var generator corresponding to the parameter to be identified and the change in the step size of the parameter to be identified includes: determining the absolute value of the ratio between the change in the controlled power of the doubly-fed asynchronous wind turbine or static var generator corresponding to the parameter to be identified and the change in the step size of the parameter to be identified; and using the absolute value of the ratio as the parameter trajectory sensitivity of each parameter to be identified.

[0009] Optionally, based on the preliminary identification results, the first sensitivity parameter is identified again until the fitness function tends to a stable value, and then the second sensitivity parameter is identified again until the fitness function tends to a stable value. The resulting control parameters include: fixing all second sensitivity parameters to their corresponding preliminary identification results, and only running the improved multi-objective Lichtenberg algorithm to identify the first sensitivity parameters again until the fitness function tends to a stable value; fixing all first sensitivity parameters to their corresponding preliminary identification results, and only running the improved multi-objective Lichtenberg algorithm to identify the second sensitivity parameters again until the fitness function tends to a stable value.

[0010] Optionally, the fitness function for parameter identification based on the simulation data of the grid-connected model and the measured data of the wind farm includes: determining the root mean square error between the simulation data of the grid-connected model and the measured data of the wind farm as the fitness function for parameter identification.

[0011] Optionally, performing parameter and network equivalence on the doubly-fed asynchronous wind turbines in each cluster to obtain an equivalent doubly-fed asynchronous wind turbine corresponding to each cluster includes: taking a weighted average of the capacity and control parameters of the doubly-fed asynchronous wind turbines in each cluster to obtain the capacity and control parameters of an equivalent doubly-fed asynchronous wind turbine corresponding to each cluster, where the capacity of the equivalent doubly-fed asynchronous wind turbine is the weighted average of the capacities of the doubly-fed asynchronous wind turbines in the cluster, and the control parameters of the equivalent doubly-fed asynchronous wind turbine are the weighted average of the control parameters of the doubly-fed asynchronous wind turbines in the cluster; simplifying the collector line impedance of the equivalent doubly-fed asynchronous wind turbine using the radial equivalent independent method to obtain the collector line impedance of the equivalent doubly-fed asynchronous wind turbine; and using the capacity, control parameters, and collector line impedance of the equivalent doubly-fed asynchronous wind turbine, determining an equivalent doubly-fed asynchronous wind turbine corresponding to each cluster.

[0012] Optionally, constructing a multi-machine equivalent model of a wind farm using equivalent doubly-fed asynchronous wind turbines and a static var generator includes: equating each doubly-fed asynchronous wind turbine to four equivalent doubly-fed asynchronous wind turbines; and constructing a multi-machine equivalent model of the wind farm using four equivalent doubly-fed asynchronous wind turbines and one static var generator.

[0013] In a second aspect, embodiments of the present invention provide an electronic device, including: a processor and a memory; wherein the memory is used to store a computer program that can run on the processor; the processor is used to execute the program stored in the memory to implement the steps of the method for constructing a multi-machine equivalent model of a wind farm as mentioned in the first aspect.

[0014] Thirdly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method for constructing a multi-machine equivalent model of a wind farm as mentioned in the first aspect.

[0015] The present invention has the following beneficial effects: By constructing an integrated grid-connected model of a doubly-fed asynchronous wind turbine and a static var generator (SVM), the embodiments of the present invention fully consider the dynamic coupling relationship between the rotor-side converter, grid-side converter, and SVM of the doubly-fed asynchronous wind turbine in cascading faults. This allows for accurate capture of reactive power interaction behavior between devices during fault mode switching, significantly reducing the analytical bias caused by neglecting synergistic effects in traditional models. Furthermore, the present invention employs an improved multi-objective Lichtenberg algorithm for overall parameter identification, simultaneously identifying the control parameters of the rotor-side, grid-side, and SVM of the doubly-fed asynchronous wind turbine, comprehensively covering the dynamic characteristics of the core reactive power regulation equipment in the wind farm, further ensuring the authenticity and reliability of the model parameters. Finally, using the dynamic response characteristics such as the reactive current trajectory and dq-axis current component of the doubly-fed asynchronous wind turbine under fault conditions as clustering criteria, the invention can accurately capture the differentiation in unit response caused by differences in control strategies, with units within the same cluster exhibiting highly similar transient behaviors. Based on this, equivalent wind farm units are obtained by using control parameters within the cluster and network equivalence. The final multi-machine equivalent model reduces computational complexity, meeting the efficiency requirements of engineering-level simulation while accurately reflecting the true dynamic distribution of units under fault conditions. This narrows the discrepancy between the simulation results of the equivalent wind farm model and the actual dynamic behavior of the wind farm, thus reducing the error in the analysis results. Attached Figure Description

[0016] To more clearly illustrate the technical solutions and advantages 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.

[0017] Figure 1 This is a flowchart illustrating a method for constructing a multi-machine equivalent model of a wind farm, as provided in an embodiment of the present invention.

[0018] Figure 2This is a schematic diagram comparing the voltage curves of an identification model and a test model, provided as an embodiment of the present invention.

[0019] Figure 3 This is a schematic diagram of the active power curves of an identification model and a test model provided in an embodiment of the present invention.

[0020] Figure 4 A schematic diagram of the active power curve of another identification model and test model provided in an embodiment of the present invention.

[0021] Figure 5 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention. Detailed Implementation

[0022] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation methods, structures, features, and effects of a method, equipment, and medium for constructing a multi-machine equivalent model of a wind farm according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0024] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method, equipment, and medium for constructing a multi-machine equivalent model of a wind farm provided by the present invention.

[0025] Please see Figures 1 to 5 , Figure 1 This is a flowchart illustrating a method for constructing a multi-machine equivalent model of a wind farm, as provided in an embodiment of the present invention. Figure 2 This is a schematic diagram comparing the voltage curves of an identification model and a test model, provided as an embodiment of the present invention. Figure 3 This is a schematic diagram of the active power curves of an identification model and a test model provided in an embodiment of the present invention. Figure 4 A schematic diagram of the active power curve of another identification model and test model provided in an embodiment of the present invention. Figure 5 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention.

[0026] like Figure 1 As shown in the embodiments of the present invention, the method for constructing a multi-machine equivalent model of a wind farm includes:

[0027] Step S101: Construct a grid-connected model of the doubly-fed asynchronous wind turbine and static var generator of the wind farm.

[0028] Specifically, this embodiment of the invention utilizes a portable power logging analyzer (such as a TK portable power logging analyzer) to collect measured data on voltage, current, active power, and reactive power at the wind farm grid connection point during cascaded short-circuit faults. A DFIG-SVG grid-connected model is built in MATLAB / Simulink. This model structure includes the control loops of the DFIG's rotor-side converter (RSC) and grid-side converter (GSC), as well as the SVG's dual closed-loop control structure.

[0029] Step S102: Based on the improved multi-objective Lichtenberg algorithm, the overall parameters of the grid-connected model are identified to obtain the control parameters of the rotor-side converter, grid-side converter, and static var generator of the doubly fed asynchronous wind turbine.

[0030] Specifically, the improved multi-objective Lichtenberg algorithm in this embodiment of the invention introduces the following strategies:

[0031] The inertial weighting strategy that integrates the inverse incomplete Weierstrass infinite product Γ function and the B distribution has the following formula for calculating the inertial weight ω:

[0032] In the above formula, Indicates inertia weight, These represent the maximum inertia weight and the minimum inertia weight, respectively. This represents the scaling factor (related to the Euler-Marshall constant or algorithm parameters, i.e., the Euler-Marshall coefficient). The decay function related to the number of iterations T is derived from the definition of the Gamma function. This represents the scaling factor for the perturbation term. The random term of the Beta function is defined by spherical coordinate integral. This represents the critical radius (a parameter in LF growth). Represents the random exponent or an integer parameter related to the particle state, where k is the scaling factor and T is the current iteration number.

[0033] Furthermore, embodiments of the present invention utilize inertial weights to update particle rotation coordinate values. :

[0034] in, The new particle rotation coordinates are represented by α, β, and γ, which are random numbers in the range [0,1]. , where i is the position of the i-th particle, and W represents the aforementioned inertial weight. .

[0035] Furthermore, as an optional embodiment of the present invention, the overall parameter identification of the grid-connected model based on the improved multi-objective Lichtenberg algorithm is performed to obtain the control parameters of the rotor-side converter, grid-side converter, and static var generator of the doubly-fed asynchronous wind turbine generator. This includes: constructing a parameter vector from the parameters to be identified of the rotor-side converter, grid-side converter, and static var generator of the doubly-fed asynchronous wind turbine generator; determining the fitness function for parameter identification based on the simulation data of the grid-connected model and the measured data of the wind farm; and performing overall parameter identification of the grid-connected model based on the improved multi-objective Lichtenberg algorithm until the fitness function tends to a stable value.

[0036] Specifically, in this embodiment of the invention, the parameters of the multi-objective Lichtenberg algorithm are first set, including the population size N, the fractal coefficient Ref (e.g., 0.2), the viscosity coefficient S, and the extreme value of the inertia weight. =0.9, =0.4, Euler coefficient Maximum number of iterations =500.

[0037] Initialize particle positions: The initial position x_global_p of the i-th particle, i is within the preset upper and lower bounds [ , The value is randomly generated within the range [i(0)], as shown in the following formula:

[0038] In the above formula, Let x_global_p,i represent the initial position of the i-th particle. This indicates the preset lower boundary. This indicates the preset upper boundary.

[0039] Furthermore, in this embodiment of the invention, the Lichtenberg pattern (LF) binary matrix is ​​initialized, with the center point L(n / 2,n / 2, n / 2) set to 1, as shown in the following formula:

[0040] In the above formula, It is the critical radius in the Lichtenberg diagram LF simulation.

[0041] Furthermore, as an optional embodiment of the present invention, the overall parameter identification of the grid-connected model based on the improved multi-objective Lichtenberg algorithm until the fitness function changes tend to a stable value includes:

[0042] An improved multi-objective Lichtenberg algorithm is used to initially identify the parameter vector, obtaining preliminary identification results to narrow down the range of high-quality solutions. Based on the changes in the controlled power of the doubly-fed asynchronous wind turbine or static var generator corresponding to the parameters to be identified, and the changes in the step size of the parameters to be identified, the trajectory sensitivity of each parameter in the parameter vector is determined. Based on the trajectory sensitivity, the parameters to be identified are divided into first sensitivity parameters and second sensitivity parameters, with the trajectory sensitivity of the first sensitivity parameter being higher than that of the second sensitivity parameter. Based on the preliminary identification results, the first sensitivity parameter is identified again until the fitness function tends to a stable value, and then the second sensitivity parameter is identified again until the fitness function tends to a stable value, thus obtaining the control parameters.

[0043] Specifically, in this embodiment of the invention, all the parameters to be identified for RSC, GSC, and SVG (including inner and outer loop PI parameters, maximum current limits i_Pmax, i_Qmax, etc.) are constructed into a parameter vector θ, as shown in the following formula:

[0044] In the above formula, Indicates voltage error. It is the voltage component in the actual dq coordinate system. It is the reference voltage component; Indicates current error. The actual current dq component, Reference current dq component; This indicates the active power error, or the actual active power. Reference active power Error = ; Indicates reactive power error; Indicates the PI parameters of the RSC (rotor-side converter); Indicates the PI parameters of the GSC (grid-side converter); The PI parameter represents the SVG (Static Var Generator).

[0045] Furthermore, in this embodiment of the invention, the multi-objective Lichtenberg algorithm is run, and the fitness function Fitness(θ) is defined as the guide for optimization, as shown in the following equation:

[0046] In the above formula, Represents the three-dimensional LF matrix in coordinates The value at the location (0 or 1), where 1 indicates a discharged point. Represents the coordinate index in three-dimensional space (which may correspond to x, y, z or row, column, height). This represents a random number that is uniformly distributed within the range [0,1) or a specific range, and is used for probability branches. These represent the growth accumulator or displacement in the three coordinate directions (used to ultimately determine the position offset of the new particle).

[0047] Furthermore, the core iterative process of the multi-objective Lichtenberg algorithm specifically includes the following steps:

[0048] First, perform particle updates and checks: update the particle rotation coordinates and perform an aggregation check, as shown in the following equation: ; ;

[0049] In the above formula, It is the Lichtenberg pattern matrix (a three-dimensional binary matrix that records the distribution of discharge points) at the t-th iteration. These are coordinate indices in the matrix; it seems we are traversing a two-dimensional slice of the matrix. Indicates two different directions or modes ( ). The displacement vector components obtained from the previous growth rules ( The components in two directions of ). The electric field function or activation probability function represents the electric field at position. The "electric field strength" or "breakdown probability" at a point is used to determine whether that point becomes 1 (discharge). This represents the LF matrix at the previous time step. These represent the coordinates of the new discharge point in the three-dimensional LF matrix. Indicates the size of the LF matrix ( ). Indicates the center coordinates (the initial discharge point position). It is a displacement vector accumulated by the growth rules; This represents the critical radius, which controls the range of random offset. This represents a random number that is uniformly distributed between 0 and 1.

[0050] Then, the inertia weight ω is dynamically adjusted using the above inertia weight adjustment formula.

[0051] Secondly, this embodiment of the invention introduces the Gold-SA operator to participate in the particle cluster selection process, selecting the current optimal cluster, as shown in the following formula:

[0052] In the above formula, , , These represent the particle's position in three directions at time t−1. The momentum, velocity, or magnetic field components on the (depending on the physical analogy). It is the inertia weight (from the previous formula). It is a coefficient that controls the magnetic field or cross-coupling. is the rotation angle (Euler angle) in three-dimensional space, used to guide the search space to the optimal region. t is the number of iterations. n is the problem dimension or the number of particles.

[0053] Furthermore, after the global search reaches a certain stage, a fractal coefficient Ref is introduced to narrow the search range for a refined local search. The particle positions are then updated as follows:

[0054] In the above formula, Indicates the particle position. Indicates the number of samples. This represents the voltage and current measurements of the i-th sample. This represents the value of the voltage error function in the i-th sample. This represents the value of the current error function in the i-th sample. This represents the value of the power error function in the i-th sample. It represents the root mean square error of voltage, current, and power.

[0055] Thus, by using the above formula to perform preliminary identification of the parameter vectors of each parameter to be identified, preliminary identification results are obtained, thereby narrowing down the range of high-quality solutions.

[0056] Furthermore, in this embodiment of the invention, the identification results from the preliminary identification are used to identify high-sensitivity parameters and low-sensitivity parameters respectively. In an optional embodiment of the invention, determining the parameter trajectory sensitivity of each parameter in the parameter vector based on the change in the controlled power of the doubly-fed asynchronous wind turbine or static var generator corresponding to the parameter to be identified and the change in the step size of the parameter to be identified includes: determining the absolute value of the ratio between the change in the controlled power of the doubly-fed asynchronous wind turbine or static var generator corresponding to the parameter to be identified and the change in the step size of the parameter to be identified; and using the absolute value of the ratio as the parameter trajectory sensitivity of each parameter to be identified.

[0057] Specifically, the embodiments of the present invention employ the following formula to calculate the parameter trajectory sensitivity of each parameter to be identified:

[0058] In the above formula, This represents the sensitivity of the parameter trajectory to be identified. This represents the change in the controlled power of the doubly-fed asynchronous wind turbine or static var generator corresponding to the parameter to be identified. This indicates the step size change of the parameter to be identified.

[0059] After calculating the parameter trajectory sensitivity, the parameters to be identified with a trajectory sensitivity greater than a preset threshold are determined as first sensitivity parameters, and the parameters to be identified with a trajectory sensitivity not greater than the preset threshold are determined as second sensitivity parameters. The preset threshold can be set according to actual conditions, and this embodiment of the invention does not impose any limitations on it.

[0060] Furthermore, as an optional embodiment of the present invention, the first sensitivity parameter is identified again based on the preliminary identification results until the fitness function changes tend to a stable value, and then the second sensitivity parameter is identified again until the fitness function changes tend to a stable value. The control parameters obtained include: fixing all the second sensitivity parameters to their corresponding preliminary identification results, and only running the improved multi-objective Lichtenberg algorithm to identify the first sensitivity parameters again until the fitness function changes tend to a stable value; fixing all the first sensitivity parameters to their corresponding preliminary identification results, and only running the improved multi-objective Lichtenberg algorithm to identify the second sensitivity parameters again until the fitness function changes tend to a stable value.

[0061] Specifically, in this embodiment of the invention, the preliminary identification results of all second sensitivity parameters are first fixed. The improved multi-objective Lichtenberg algorithm is then run again on the first sensitivity parameters until the fitness function stabilizes and the goodness of fit (1 - Fitness(θ)) of the grid-connected model is greater than 95%. Then, the identified first sensitivity parameters are fixed, and the improved multi-objective Lichtenberg algorithm is run again on the remaining second sensitivity parameters. Finally, the optimal identification values ​​for all control parameters are obtained.

[0062] Furthermore, as an optional embodiment of the present invention, determining the fitness function for parameter identification based on the simulation data of the grid-connected model and the measured data of the wind farm includes: determining the root mean square error between the simulation data of the grid-connected model and the measured data of the wind farm as the fitness function for parameter identification.

[0063] Furthermore, as an optional embodiment of the present invention, determining the fitness function for parameter identification based on the simulation data of the grid-connected model and the measured data of the wind farm includes: determining the root mean square error between the simulation data of the grid-connected model and the measured data of the wind farm as the fitness function for parameter identification.

[0064] Specifically, in this embodiment of the invention, the fitness function for parameter identification is defined as the root mean square error between the measured data and the simulation data:

[0065] in, Let the error vector be the error in voltage, current, active power, and reactive power, and its definition is:

[0066] Where V and I are the voltage and current output responses of the grid-connected model, respectively, V* and I* are the measured voltage and current values ​​on site, respectively, and the subscripts d and q represent the dq axis components.

[0067] Step S103: Input the control parameters into the grid-connected model to obtain the reactive current trajectory and dq-axis current components of the doubly-fed asynchronous wind turbine generators in the wind farm. Then, cluster the doubly-fed asynchronous wind turbine generators according to the reactive current trajectory and dq-axis current components to obtain multiple clusters.

[0068] Specifically, in this embodiment of the invention, the finally identified control parameters are substituted into the grid-connected model to simulate a cascading fault scenario identical to the measured data, and the reactive current trajectories and dq-axis current components of all N DFIG units in the wind farm are recorded. , Then, based on the performance of each DFIG unit during the failure... , Time-series data is used as clustering features. Euclidean distance or dynamic time-warped distance is used as a similarity measure.

[0069] Using the HAC algorithm, with distance minimization as the criterion, N DFIG units are aggregated into K clusters (K=4 in this embodiment of the invention).

[0070] Step S104: Perform parameter equivalence and network equivalence on the doubly fed asynchronous wind turbines in each cluster to obtain an equivalent doubly fed asynchronous wind turbine corresponding to each cluster, and use the equivalent doubly fed asynchronous wind turbine and static var generator to construct a multi-machine equivalent model of the wind farm.

[0071] Specifically, further, as an optional embodiment of the present invention, performing parameter equivalence and network equivalence on the doubly-fed asynchronous wind turbines in each cluster to obtain an equivalent doubly-fed asynchronous wind turbine corresponding to each cluster includes: taking a weighted average of the capacity and control parameters of the doubly-fed asynchronous wind turbines in each cluster to obtain the capacity and control parameters of an equivalent doubly-fed asynchronous wind turbine corresponding to each cluster, wherein the capacity of the equivalent doubly-fed asynchronous wind turbine is the weighted average of the capacities of the doubly-fed asynchronous wind turbines in the cluster, and the control parameters of the equivalent doubly-fed asynchronous wind turbine are the weighted average of the control parameters of the doubly-fed asynchronous wind turbines in the cluster; simplifying the collector line impedance of the equivalent doubly-fed asynchronous wind turbine using the radial equivalent independent method to obtain the collector line impedance of the equivalent doubly-fed asynchronous wind turbine; and determining an equivalent doubly-fed asynchronous wind turbine corresponding to each cluster using the capacity, control parameters, and collector line impedance of the equivalent doubly-fed asynchronous wind turbine.

[0072] Specifically, in this embodiment of the invention, the capacity and control parameters (such as identified PI parameters) of each DFIG unit within a cluster are weighted and averaged to represent an equivalent DFIG unit. Then, the Radial Equivalent Independent (REI) method is used to simplify the collector line impedance within the cluster. Finally, the detailed wind farm model containing hundreds of units is simplified into a four-unit equivalent model consisting of four equivalent DFIG units and one SVG.

[0073] Furthermore, in this embodiment of the invention, the constructed multi-machine equivalent model is simulated under another set of cascading fault scenarios that were not involved in the identification, and its dynamic response (PCC point voltage, active power, reactive power) is compared with the field measured data (or detailed model simulation results). The verification results are as follows: Figures 2 to 4 As shown, the curve fitting degree between the multi-level equivalent identification model and the test model of the measured data exceeds 95%, effectively reproducing the transient characteristics of the system during the cascading fault ride-through process, proving the effectiveness and rationality of the method of the present invention.

[0074] This invention constructs an integrated grid-connected model of a doubly-fed asynchronous wind turbine (DFIG) and a static var generator (SVA). It fully considers the dynamic coupling relationship between the rotor-side converter, grid-side converter, and SVA in cascading faults, accurately capturing the reactive power interaction behavior between devices during fault mode switching, significantly reducing the analytical bias caused by neglecting synergistic effects in traditional models. Furthermore, this invention employs an improved multi-objective Lichtenberg algorithm for overall parameter identification, simultaneously identifying the control parameters of the rotor-side, grid-side, and SVA of the DFIG, comprehensively covering the dynamic characteristics of the core reactive power regulation equipment in the wind farm, further ensuring the authenticity and reliability of the model parameters. Finally, using the dynamic response characteristics such as the reactive current trajectory and dq-axis current component of the DFIG under fault conditions as clustering criteria, it can accurately capture the differentiation in unit response caused by differences in control strategies, with units within the same cluster exhibiting highly similar transient behaviors. Based on this, equivalent wind farm units are obtained by using control parameters within the cluster and network equivalence. The final multi-machine equivalent model reduces computational complexity, meeting the efficiency requirements of engineering-level simulation while accurately reflecting the true dynamic distribution of units under fault conditions. This narrows the discrepancy between the simulation results of the equivalent wind farm model and the actual dynamic behavior of the wind farm, thus reducing the error in the analysis results.

[0075] Corresponding to the method for constructing a multi-machine equivalent model of a wind farm provided in the above embodiments, based on the same technical concept, this embodiment of the invention also provides an electronic device for executing the above method for constructing a multi-machine equivalent model of a wind farm. Figure 5 This is a schematic diagram of the structure of an electronic device provided in one embodiment of the present invention, as shown below. Figure 5 As shown. Electronic devices can vary considerably due to differences in configuration or performance, and may include one or more processors 501 and memories 502. The memory 502 stores computer programs that can run on the processor 501, and the processor 501 executes the programs stored in the memory 502 to achieve the above. Figure 1 The various steps in the method embodiment are described. Memory 502 can be temporary or persistent storage. The application stored in memory 502 may include one or more modules (not shown), each module may include a series of computer-executable instructions for the electronic device.

[0076] Furthermore, the processor 501 may be configured to communicate with the memory 502 and execute a series of computer-executable instructions stored in the memory 502 on the electronic device. The electronic device may also include one or more power supplies 503, one or more wired or wireless network interfaces 504, one or more input / output interfaces 505, and one or more keyboards 506.

[0077] Specifically, in this embodiment, the electronic device includes a processor, a communication interface, a memory, and a communication bus; wherein, the processor, the communication interface, and the memory communicate with each other via the bus; the memory is used to store computer programs; and the processor is used to execute the programs stored in the memory to achieve the above. Figure 1 The various steps in the method embodiments are the same as those in the above method embodiments, and have the same beneficial effects. To avoid repetition, the embodiments of the present invention will not be described again here.

[0078] It should be noted that the electronic equipment provided in this embodiment of the invention and the method for constructing a multi-machine equivalent model of a wind farm provided in this embodiment of the invention are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned method for constructing a multi-machine equivalent model of a wind farm, and has the same or similar beneficial effects. Repeated parts will not be described again.

[0079] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0080] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for constructing a multi-machine equivalent model of a wind farm, characterized in that, The application relates to a wind farm parameter identification method. The application comprises the following steps: constructing a grid-connected model of a doubly-fed asynchronous wind generator and a static var generator of the wind farm; performing overall parameter identification on the grid-connected model based on an improved multi-objective Lichengberg algorithm to obtain control parameters of a rotor-side converter, a grid-side converter and the static var generator of the doubly-fed asynchronous wind generator; inputting the control parameters into the grid-connected model to obtain a reactive current trajectory and dq-axis current components of the doubly-fed asynchronous wind generator of the wind farm, and performing clustering on the doubly-fed asynchronous wind generator according to the reactive current trajectory and the dq-axis current components to obtain a plurality of clusters; 2. The method of claim 1, wherein, performing parameter equivalence and network equivalence on the doubly-fed asynchronous wind generator in each cluster to obtain an equivalent doubly-fed asynchronous wind generator corresponding to each cluster, and constructing a multi-machine equivalent model of the wind farm by using the equivalent doubly-fed asynchronous wind generator and the static var generator. The application performs overall parameter identification on the grid-connected model based on the improved multi-objective Lichengberg algorithm to obtain the control parameters of the rotor-side converter, the grid-side converter and the static var generator of the doubly-fed asynchronous wind generator, which comprises the following steps: constructing a parameter vector by using the to-be-identified parameters of the rotor-side converter, the grid-side converter and the static var generator of the doubly-fed asynchronous wind generator; determining a fitness function for parameter identification based on simulation data of the grid-connected model and measured data of the wind farm; 3. The method of claim 2, wherein, performing overall parameter identification on the grid-connected model based on the improved multi-objective Lichengberg algorithm until the fitness function changes to a stable value. The application performs overall parameter identification on the grid-connected model based on the improved multi-objective Lichengberg algorithm until the fitness function changes to a stable value, which comprises the following steps: performing preliminary identification on the parameter vector based on the improved multi-objective Lichengberg algorithm to obtain a preliminary identification result, so as to reduce the range of high-quality solution space; determining parameter trajectory sensitivity of each to-be-identified parameter in the parameter vector according to a variation of a controlled electric quantity of the doubly-fed asynchronous wind generator or the static var generator corresponding to the to-be-identified parameter and a step variation of the to-be-identified parameter; dividing the to-be-identified parameters into first sensitivity parameters and second sensitivity parameters according to the parameter trajectory sensitivity, wherein the parameter trajectory sensitivity of the first sensitivity parameters is higher than that of the second sensitivity parameters; 4. The method of claim 3, wherein, performing re-identification on the first sensitivity parameters according to the preliminary identification result until the fitness function changes to a stable value, and then performing re-identification on the second sensitivity parameters until the fitness function changes to a stable value, so as to obtain the control parameters. The application determines the parameter trajectory sensitivity of each to-be-identified parameter in the parameter vector according to a variation of a controlled electric quantity of the doubly-fed asynchronous wind generator or the static var generator corresponding to the to-be-identified parameter and a step variation of the to-be-identified parameter, which comprises the following steps: determining an absolute value of a ratio between the variation of the controlled electric quantity of the doubly-fed asynchronous wind generator or the static var generator corresponding to the to-be-identified parameter and the step variation of the to-be-identified parameter. An absolute value of the ratio is taken as a parameter trajectory sensitivity of each of the to-be-identified parameters.

5. The method of claim 3, wherein, The preliminary identification result is used to identify the first sensitivity parameter again until the fitness function tends to be stable, and then the second sensitivity parameter is identified again until the fitness function tends to be stable, to obtain the control parameter, and the method comprises: All the second sensitivity parameters are fixed as the corresponding preliminary identification results, and the improved multi-objective Levenberg-Marquardt algorithm is only used to identify the first sensitivity parameter again until the fitness function tends to be stable; All the first sensitivity parameters are fixed as the corresponding preliminary identification results, and the improved multi-objective Levenberg-Marquardt algorithm is only used to identify the second sensitivity parameter again until the fitness function tends to be stable.

6. The method of claim 2, wherein, The fitness function for parameter identification is determined based on the simulation data of the grid-connected model and the measured data of the wind farm, and the method comprises: The root mean square error between the simulation data of the grid-connected model and the measured data of the wind farm is determined as the fitness function for parameter identification.

7. The method of claim 1, wherein, The parameter equivalence and network equivalence of the doubly-fed asynchronous wind power generator in each cluster are performed to obtain an equivalent doubly-fed asynchronous wind power generator corresponding to each cluster, and the method comprises: The capacity and control parameters of the doubly-fed asynchronous wind power generator in each cluster are weighted and averaged to obtain the capacity and control parameters of an equivalent doubly-fed asynchronous wind power generator corresponding to each cluster, the capacity of the equivalent doubly-fed asynchronous wind power generator is a weighted average of the capacities of the doubly-fed asynchronous wind power generators in the cluster, and the control parameters of the equivalent doubly-fed asynchronous wind power generator are a weighted average of the control parameters of the doubly-fed asynchronous wind power generators in the cluster; The radial equivalent independent method is used to equivalently simplify the collector line impedance of the equivalent doubly-fed asynchronous wind power generator to obtain the collector line impedance of the equivalent doubly-fed asynchronous wind power generator; The capacity, control parameters and collector line impedance of the equivalent doubly-fed asynchronous wind power generator are used to determine an equivalent doubly-fed asynchronous wind power generator corresponding to each cluster.

8. The method of claim 1, wherein, The method comprises: Each of the doubly-fed asynchronous wind power generators is equivalently converted into four equivalent doubly-fed asynchronous wind power generators; The four equivalent doubly-fed asynchronous wind power generators and a static reactive power generator are used to construct a multi-machine equivalent model of the wind farm.

9. An electronic device, comprising: The method comprises: a processor and a memory; wherein the memory is used to store a computer program which can be run on the processor; the processor is used to execute the program stored on the memory to realize the steps of the method for constructing the multi-machine equivalent model of the wind farm according to any one of claims 1-8.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method for constructing the multi-machine equivalent model of the wind farm according to any one of claims 1-8.

Images