A new type of power grid dynamic equivalent method and system

Abstract

The present invention discloses a method and system for dynamic equivalent of a power grid of a new type of power system, the method comprising: obtaining an original model of the new type of power system, and dividing a study area and an external area; presetting a disturbance scenario and establishing a disturbance scenario data set; determining the number of clusters based on the electrical distance from the unit in the external area to the node in the study area; clustering and dividing the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model; optimizing the parameters of the simplified model in a loop until the termination condition is reached to obtain the optimal simplified model and its parameters. The system comprises a region division module, a disturbance module, a cluster number confirmation module, a cluster division module and a cross optimization module. By using the present invention, the computational complexity of the dynamic response mechanism can be simplified, and the problem that parameter identification optimization is prone to overfitting can be avoided. The present invention can be widely applied to the technical field of dynamic response analysis of new power systems.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic response analysis of a novel power system, and in particular to a power grid dynamic equivalent method and system for a novel power system. Background Art

[0002] Dynamic response analysis is of great significance for the stability study of new power systems. However, it is computationally demanding to conduct this type of analysis because its original model is complex and detailed. In fact, only a certain part of the power system needs to be the main focus of the study, which is defined as the study area. The rest of the system is called the external area, and its structure and internal parameters do not need to be focused on. In order to simplify the process of dynamic analysis, the external area is replaced by a simplified equivalent model, and it is guaranteed that the simplified model should have similar dynamic characteristics to the original model.

[0003] At present, the traditional methods for the study of power grid dynamic equivalents are mainly: (1) the coherent equivalent method, which includes three parts: identification of related groups, network simplification and aggregation of related group parameters, and simulates the dynamic curve of transient stability analysis based on the co-swing of the generator rotor; (2) the model equivalent method, which establishes the differential equation of the external system, then linearizes it to obtain the state equation, and further obtains the reduced-order state equation. In addition, with the widespread application of synchronized phasor measurement unit (PMU) technology in new power systems, parameter identification based on measured data is also considered to be an effective method for obtaining the parameters of the power grid equivalent model. Although the former is rigorous in theory, it requires complex calculations of the detailed model and parameters of the entire system, and for new power systems with increasing penetration of new energy, its dynamic response mechanism and mode are more complex. The latter can be applied to systems with rapidly changing operating conditions and structures, but it lacks a theoretical basis and is prone to overfitting problems.

[0004] Therefore, in the scenario of new energy station access in the new power system, the existing technology has the problem that the calculation of dynamic response mechanism is more complicated and the intelligent algorithm parameter identification and optimization is prone to overfitting. Summary of the invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a new type of power system grid dynamic equivalence method and system, which can simplify the calculation complexity of the dynamic response mechanism and avoid the problem of overfitting in parameter identification and optimization.

[0006] The first technical solution adopted by the present invention is: a new type of power grid dynamic equivalent method of the power system, comprising the following steps:

[0007] Obtain the original model of the new power system and divide the study area and external area;

[0008] Presetting disturbance scenarios in the study area and establishing a disturbance scenario data set;

[0009] Determining the number of clusters based on the electrical distances from the units in the external area to the nodes in the study area;

[0010] Clustering the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model;

[0011] The parameters of the simplified model are optimized iteratively based on the improved vertical and horizontal cross algorithm until the termination condition is reached, thereby obtaining the optimal simplified model and its parameters.

[0012] Further, the step of determining the number of clusters based on the electrical distances from the units in the external area to the nodes in the research area specifically includes:

[0013] Calculate the transfer reactance from each synchronous unit and new energy station in the external area to the nearest node of the study area to obtain transfer reactance data;

[0014] Calculate the mean and standard deviation of the transfer reactance data;

[0015] Dividing the transfer reactance data based on the mean value and the standard deviation to obtain a divided data set;

[0016] The number of clusters is determined based on the number of empty sets in the divided data set.

[0017] Furthermore, the step of clustering the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model specifically includes:

[0018] Based on the synchronous generator power angle time series data set in the disturbance scenario data set, calculating the swing mean and instability times of the power angle of each synchronous generator in the external area;

[0019] Based on the swing mean value of the power angle of each synchronous unit, the number of instability, the rated capacity, the steady-state reactance and the transient reactance, the external area is clustered by using the K-means clustering algorithm and the number of cluster families to obtain a first clustering result;

[0020] Based on the voltage time series data set of the grid connection points of the new energy stations in the disturbance scenario data set, the average voltage drop degree and the number of times of entering low voltage at the grid connection points of each new energy station in the external area are calculated;

[0021] Based on the average voltage drop degree of the grid connection points of each new energy station, the number of times of entering low voltage, the new energy type, the rated capacity and the stator and rotor reactance, the external area is clustered by using the K-means clustering algorithm and the number of cluster families to obtain a second clustering result;

[0022] The clusters in the external area are merged based on the first clustering result and the second clustering result to obtain a simplified model.

[0023] Furthermore, the step of optimizing the parameters of the simplified model based on the improved cross-cutting algorithm in a loop iteration until the termination condition is reached to obtain the optimal simplified model and its parameters specifically includes:

[0024] Initialize the parameters of the improved cross-cutting algorithm and determine the iterative optimization parameters of the simplified model;

[0025] An index of dynamic response similarity is constructed based on the voltage-frequency deviation coefficient;

[0026] Based on the index of the dynamic response similarity and the vertical crossover probability, respectively performing horizontal crossover and vertical crossover on the population, recording the optimal index value of the dynamic response similarity and updating the population;

[0027] The loop is iterated until the termination condition is reached, and the optimal simplified model and its parameters are output.

[0028] Furthermore, the index of the dynamic response similarity is expressed as follows:

[0029]

[0030] in, An indicator representing the similarity of dynamic responses; Represents a disturbance scenario Next research node The voltage-frequency deviation coefficient; represents the number of disturbance scenarios; Indicates the number of research nodes; Indicates the number of sampling measurements; Represents a disturbance scenario Next research node No. Sub-sampling measurement voltage; Represents a disturbance scenario Next research node The average voltage value of Represents a disturbance scenario Next research node No. Subsampling measurement frequency; Represents a disturbance scenario Next research node The frequency average of represents the voltage-frequency deviation coefficient matrix; Indicates the use of simplified models in disturbance scenarios Next research node No. Sub-sampling measurement voltage; Indicates the use of simplified models in disturbance scenarios Next research node No. Subsampling measurement frequency.

[0031] Furthermore, the longitudinal crossover is expressed as follows:

[0032]

[0033] in, Indicates new particles generated by longitudinal crossing; Indicates that the randomly selected dimension is Particles ; Indicates that the randomly selected dimension is Particles ; Representation range Random numbers on ; Indicates the size of the population; represents a particle; represents the population dimension; Indicates a natural number within a range; Indicates the number of equivalent machines in the synchronous unit; and Represents particles Two different dimension numbers.

[0034] Furthermore, the longitudinal crossing probability is determined by the degree of data confusion of the parameters contained in the particle, and its expression is as follows:

[0035]

[0036] in, represents the vertical crossover probability; represents the division coefficient; Indicates the size of the population; represents the population dimension; Represents particles The between-group covariance matrix containing the parameters; Represents particles The within-group covariance matrix containing the parameters; Represents the trace of the matrix.

[0037] The second technical solution adopted by the present invention is: a new type of power grid dynamic equivalent system, comprising:

[0038] The regional division module is used to obtain the original model of the new power system and divide the study area and the external area;

[0039] A disturbance module, used to preset disturbance scenarios in the study area and establish a disturbance scenario data set;

[0040] A cluster number confirmation module, which determines the cluster number based on the electrical distance between the units in the external area and the nodes in the research area;

[0041] A clustering module, which performs clustering on the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model;

[0042] The cross optimization module performs cyclic iterative optimization on the parameters of the simplified model based on the improved vertical and horizontal cross algorithm until the termination condition is reached to obtain the optimal simplified model and its parameters.

[0043] The beneficial effects of the method and system of the present invention are as follows: the present invention completes the preliminary division of the external area by calculating the average value and standard deviation of the transfer reactance data; determines the number of clustering families by the electrical distance from the unit in the external area to the node of the study area, and further refines the clustering of the external area in combination with the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model; uses an improved vertical and horizontal cross algorithm to perform cyclic iterative optimization on the parameters of the simplified model until the termination condition is reached, and obtains the optimal simplified model and its parameters; finally, the optimal simplified model and its parameters can realize the computational complexity of simplifying the dynamic response mechanism, and avoid the parameter identification optimization from easily falling into overfitting. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a flow chart of the steps of a novel power grid dynamic equivalent method of the present invention;

[0045] Figure 2 It is a structural block diagram of a power grid dynamic equivalent system of a novel power system of the present invention;

[0046] Figure 3 It is a regional division diagram of a new type of power grid dynamic equivalent method of the present invention;

[0047] Figure 4 It is a sub-area division diagram of a new type of power system dynamic equivalent method of the present invention; wherein ① represents the first sub-area; ② represents the second sub-area; ③ represents the third sub-area; ④ represents the fourth sub-area;

[0048] Figure 5 It is a schematic diagram of the number of buses and units included in each sub-area before and after the group merger of a new type of power system dynamic equivalent method of the present invention;

[0049] Figure 6 It is a flow chart of an improved vertical and horizontal cross algorithm of a novel power grid dynamic equivalent method of the present invention;

[0050] Figure 7 It is a fitness change curve diagram of a new type of power grid dynamic equivalent method of the present invention;

[0051] Figure 8 It is a voltage and frequency schematic diagram of an original model and a simplified model of a new type of power system dynamic equivalent method of the present invention. DETAILED DESCRIPTION

[0052] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only provided for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0053] Reference Figure 1 The present invention provides a novel power grid dynamic equivalent method for a power system, the method comprising the following steps:

[0054] S1. Obtain the original model of the new power system and divide the study area and external area;

[0055] Specifically, the study area refers to the analysis of the internal state and response of the system in this area according to the research objectives. After determining the boundary between the inside and outside of the system, the point to be studied is further determined by the location of the fault disturbance point, thereby dividing the study area of ​​the system. The remaining area is determined as the external area of ​​the power system.

[0056] Reference Figure 3 According to the research requirements, the specific embodiment of the present invention divides the solid line frame into the research area and the dotted line frame into the external area. The research area has , There are two nodes connected to the external area. ; The number of buses and units in the study area are 36 and 9 respectively, so bus failure and unit tripping events are taken as disturbance scenarios. .

[0057] S2. Preset a disturbance scenario in the study area and establish a disturbance scenario data set;

[0058] Specifically, suppose the study area has nodes are connected to the external area, and the total number of disturbance scenarios is In the original model, it is measured The voltage and frequency time series data sets of each node are and , in the external area The time series data set of the power angle of the synchronous unit is , The time series data set of the grid connection point voltage of each new energy station is , the expression of the disturbance scene data set is as follows:

[0059]

[0060] in, Indicates voltage; Indicates frequency; Indicates the power angle.

[0061] S3, determining the number of clusters based on the electrical distances from the units in the external area to the nodes in the study area;

[0062] S3.1. Calculate the transfer reactance from each synchronous unit and new energy station in the external area to the nearest node in the study area to obtain transfer reactance data;

[0063] S3.2. Calculate the average value of the transfer reactance data and standard deviation ;

[0064] S3.3, dividing the transfer reactance data based on the mean value and the standard deviation to obtain a divided data set;

[0065] Specifically, the transfer reactance data is divided into 2n data sets, and the expression is as follows:

[0066]

[0067] in, represents the average value of the transfer reactance data; represents the standard deviation of the transfer reactance data; Indicates the number of data sets after partitioning.

[0068] S3.4. Determine the number of clusters based on the number of empty sets in the divided data set.

[0069] Specifically, if there is empty sets, then the number of clusters is determined to be .

[0070] S4, clustering the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model;

[0071] S4.1. Based on the synchronous unit power angle timing data set in the disturbance scenario data set , calculate the swing mean and instability times of the power angle of each synchronous unit in the external area;

[0072] S4.2. Based on the swing mean value of the power angle of each synchronous unit, the number of instability, the rated capacity, the steady-state reactance and the transient reactance, the external area is clustered by using the K-means clustering algorithm and the number of cluster families to obtain a first clustering result;

[0073] S4.3. Voltage time series data set of the new energy station grid connection point based on the disturbance scenario data set , calculate the average voltage drop degree of the grid connection point voltage of each new energy station in the external area and the number of times it enters low voltage;

[0074] S4.4. Based on the average voltage drop degree of the grid connection points of each new energy station, the number of times of entering low voltage, the new energy type, the rated capacity and the stator and rotor reactance, the external area is clustered by the K-means clustering algorithm and the number of cluster families to obtain a second clustering result;

[0075] S4.5. Based on the first clustering result and the second clustering result, the clusters in the external area are merged to obtain a simplified model.

[0076] Specifically, for a For clustered units, the rated capacity of the equivalent units is the sum of the combined units, and the aggregation of other initial parameters adopts the weighted method based on rated capacity and transfer reactance, and the calculation formula is as follows:

[0077]

[0078] in, represents the entropy weight of the rated capacity; represents the entropy weight of the transfer reactance; Indicates the number of units in the cluster; Indicates rated capacity; It represents the combination of the reactance parameters of the group, the initial inertia constant of the synchronous unit, the damping coefficient of the synchronous unit, the initial active power-frequency droop coefficient of the new energy station and the reactive power-voltage droop coefficient of the new energy station; represents the transfer reactance.

[0079] According to the specific embodiment of the present invention, the external area can be divided into four sub-areas based on the transfer impedance in step S3, namely, the first sub-area ①, the second sub-area ②, the third sub-area ③ and the fourth sub-area ④. Figure 4 However, due to the complexity of the unit parameters in each sub-area, each sub-area is divided again, and the clusters are further clustered and merged. Before and after the cluster merger, the number of buses and units in each sub-area is as follows: Figure 5 As shown, sub-area 2 includes a photovoltaic power station, sub-area 4 includes a wind farm, and the others are all synchronous units.

[0080] S5. Optimizing the parameters of the simplified model iteratively based on the improved cross-cutting algorithm until a termination condition is reached, thereby obtaining an optimal simplified model and its parameters.

[0081] S5.1. Initialize the parameters of the improved cross-cutting algorithm and determine the iterative optimization parameters of the simplified model;

[0082] Specifically, refer to Figure 6 , set the maximum number of iterations to , generating a size of , the dimension is The initial population .for For a synchronous unit, its inertia constant and damping coefficient are used as optimization parameters; The active-frequency droop coefficient and reactive-voltage droop coefficient of each new energy station are used as optimization parameters, with a total of four groups of optimization parameters. The upper and lower limits of each dimension are based on the initial parameters of the simplified model obtained in step S4, and are taken as ±25% respectively. and the vertical crossover probability , where the horizontal crossover probability Generally set to 1.

[0083] In a specific embodiment of the present invention, the maximum number of iterations is Set to 50 times; population size Set to 50; According to the result of step S4, , ,but .

[0084] S5.2, constructing an index of dynamic response similarity based on voltage-frequency deviation coefficient;

[0085] Specifically, the original model and the simplified model were used to sample and measure the The voltage and frequency of the research node are obtained to obtain the voltage and frequency time series data set of the original model. , And the voltage and frequency time series data set of the simplified model , .

[0086] Calculate separately Under disturbance scenario The voltage-frequency deviation coefficient matrix of the studied nodes , which is expressed as follows:

[0087]

[0088] in, Represents a disturbance scenario Next research node The voltage-frequency deviation coefficient; represents the number of disturbance scenarios; Indicates the number of research nodes; Indicates the number of sampling measurements; Represents a disturbance scenario Next research node No. Sub-sampling measurement voltage; Represents a disturbance scenario Next research node The average voltage value of Represents a disturbance scenario Next research node No. Subsampling measurement frequency; Represents a disturbance scenario Next research node The frequency average of represents the voltage-frequency deviation coefficient matrix; Indicates the use of simplified models in disturbance scenarios Next research node No. Sub-sampling measurement voltage; Indicates the use of simplified models in disturbance scenarios Next research node No. Subsampling measurement frequency.

[0089] Defining metrics for dynamic response similarity , which is the fitness function of the algorithm iteration, is expressed as follows:

[0090]

[0091] in, A metric that represents the similarity of dynamic responses.

[0092] S5.3, based on the index of the dynamic response similarity and the vertical crossover probability, perform horizontal crossover and vertical crossover on the population respectively, record the optimal index value of the dynamic response similarity and update the population;

[0093] Specifically, since the particle parameters include both synchronous units and new energy stations, the expression of the vertical crossover is as follows:

[0094]

[0095] in, Indicates new particles generated by longitudinal crossing; Indicates that the randomly selected dimension is Particles ; Indicates that the randomly selected dimension is Particles ; Representation range Random numbers on ; Indicates the size of the population; represents a particle; represents the population dimension; Indicates a natural number within a range; Indicates the number of equivalent machines in the synchronous unit; and Represents particles Two different dimension numbers.

[0096] The longitudinal crossing probability is determined by the degree of data confusion of the parameters contained in the particle, and its expression is as follows:

[0097]

[0098] in, represents the vertical crossover probability; represents the division coefficient; Indicates the size of the population; represents the population dimension; Represents particles The between-group covariance matrix containing the parameters; Represents particles The within-group covariance matrix containing the parameters; Represents the trace of the matrix.

[0099] S5.4. Iterate the loop until the termination condition is reached and output the optimal simplified model and its parameters.

[0100] Specifically, the termination condition is that the number of loop iterations reaches .

[0101] The fitness (dynamic response similarity index) variation curve obtained by iteration in step S5 in the specific embodiment of the present invention is as follows: Figure 6 As shown in the figure, the best fitness starts from 0.385 and converges to 0.183 after the 30th iteration. Output the optimized simplified model parameters, compare the simplified model with the initial parameters and the simplified model with the optimized parameters with the original model, and measure the voltage and frequency curves under the same fault disturbance scenario and the same research node. For example, the fault disturbance is that the synchronous unit on bus No. 23 in the study area trips in the simulation time of 1 second, and the power loss is .

[0102] The voltage and frequency curves measured in three cases are as follows Figure 8 As shown, it is obvious that the optimized simplified model can better track the dynamic response of the original model in terms of both oscillation amplitude and oscillation frequency.

[0103] Reference Figure 2 The present invention provides a novel power grid dynamic equivalent system, comprising:

[0104] The regional division module is used to obtain the original model of the new power system and divide the study area and the external area;

[0105] A disturbance module, used to preset disturbance scenarios in the study area and establish a disturbance scenario data set;

[0106] A cluster number confirmation module, which determines the cluster number based on the electrical distance between the units in the external area and the nodes in the research area;

[0107] A clustering module, which performs clustering on the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model;

[0108] The cross optimization module performs cyclic iterative optimization on the parameters of the simplified model based on the improved vertical and horizontal cross algorithm until the termination condition is reached to obtain the optimal simplified model and its parameters.

[0109] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0110] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A new type of power grid dynamic equivalent method for power system, characterized in that: The following steps are involved: Obtain the original model of the new power system and divide the study area and external area; Presetting disturbance scenarios in the study area and establishing a disturbance scenario data set; Determining the number of clusters based on the electrical distances from the units in the external area to the nodes in the study area; Clustering the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model; The parameters of the simplified model are optimized iteratively based on the improved cross-cutting algorithm until the termination condition is reached, thereby obtaining the optimal simplified model and its parameters; The step of determining the number of clusters based on the electrical distances from the units in the external area to the nodes in the research area specifically includes: Calculate the transfer reactance from each synchronous unit and new energy station in the external area to the nearest node of the study area to obtain transfer reactance data; Calculate the mean and standard deviation of the transfer reactance data; Dividing the transfer reactance data based on the mean value and the standard deviation to obtain a divided data set; Determine the number of clusters based on the number of empty sets in the divided data set; The step of clustering the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model specifically includes: Based on the synchronous generator power angle time series data set in the disturbance scenario data set, calculating the swing mean and instability times of the power angle of each synchronous generator in the external area; Based on the swing mean value of the power angle of each synchronous unit, the number of instability, the rated capacity, the steady-state reactance and the transient reactance, the external area is clustered by using the K-means clustering algorithm and the number of cluster families to obtain a first clustering result; Based on the voltage time series data set of the grid connection points of the new energy stations in the disturbance scenario data set, the average voltage drop degree and the number of times of entering low voltage at the grid connection points of each new energy station in the external area are calculated; Based on the average voltage drop degree of the grid connection points of each new energy station, the number of times of entering low voltage, the new energy type, the rated capacity and the stator and rotor reactance, the external area is clustered by using the K-means clustering algorithm and the number of cluster families to obtain a second clustering result; The clusters in the external area are merged based on the first clustering result and the second clustering result to obtain a simplified model.

2. According to the novel power grid dynamic equivalent method of claim 1, it is characterized in that: The step of optimizing the parameters of the simplified model based on the improved cross-cutting algorithm in a loop iteration until the termination condition is reached to obtain the optimal simplified model and its parameters specifically includes: Initialize the parameters of the improved cross-cutting algorithm and determine the iterative optimization parameters of the simplified model; An index of dynamic response similarity is constructed based on the voltage-frequency deviation coefficient; Based on the index of the dynamic response similarity and the vertical crossover probability, respectively performing horizontal crossover and vertical crossover on the population, recording the optimal index value of the dynamic response similarity and updating the population; The loop is iterated until the termination condition is reached, and the optimal simplified model and its parameters are output.

3. According to the novel power grid dynamic equivalent method of claim 2, it is characterized in that: The index of the dynamic response similarity is expressed as follows: in, An indicator representing the similarity of dynamic responses; Represents a disturbance scenario Next research node The voltage-frequency deviation coefficient; represents the number of disturbance scenarios; Indicates the number of research nodes; Indicates the number of sampling measurements; Represents a disturbance scenario Next research node No. Sub-sampling measurement voltage; Represents a disturbance scenario Next research node The average voltage value of Represents a disturbance scenario Next research node No. Subsampling measurement frequency; Represents a disturbance scenario Next research node The frequency average of represents the voltage-frequency deviation coefficient matrix; Indicates the use of simplified models in disturbance scenarios Next research node No. Sub-sampling measurement voltage; Indicates the use of simplified models in disturbance scenarios Next research node No. Subsampling measurement frequency.

4. According to the novel power grid dynamic equivalent method of claim 3, it is characterized in that: The longitudinal intersection is expressed as follows: in, Indicates new particles generated by longitudinal crossing; Indicates that the selected dimension is Particles New particles generated by longitudinal crossing; Indicates that the randomly selected dimension is Particles ; Indicates that the randomly selected dimension is Particles ; Representation range Random numbers on ; Indicates the size of the population; represents a particle; represents the population dimension; Indicates a natural number within a range; Indicates the number of equivalent machines in the synchronous unit; and Represents particles Two different dimension numbers.

5. According to the novel power grid dynamic equivalent method of claim 4, it is characterized in that: The longitudinal crossing probability is determined by the degree of data confusion of the parameters contained in the particle, and its expression is as follows: in, represents the vertical crossover probability; represents the division coefficient; Indicates the size of the population; represents the population dimension; Represents particles The between-group covariance matrix containing the parameters; Represents particles The within-group covariance matrix containing the parameters; represents the trace of a matrix; Represents particles.

6. A new type of power grid dynamic equivalent system, characterized in that: A method for executing a power grid dynamic equivalent value of a novel power system as claimed in claim 1, comprising: The regional division module is used to obtain the original model of the new power system and divide the study area and the external area; A disturbance module, used for presetting disturbance scenarios in the study area and establishing a disturbance scenario data set; A cluster number confirmation module, which determines the cluster number based on the electrical distance between the units in the external area and the nodes in the research area; A clustering module, which performs clustering on the external area based on the number of clusters, the disturbance scenario data set and the unit characteristic parameters to obtain a simplified model; The cross optimization module performs cyclic iterative optimization on the parameters of the simplified model based on the improved vertical and horizontal cross algorithm until the termination condition is reached to obtain the optimal simplified model and its parameters.

Images