A practical aggregation method for power system frequency models that take into account the spatiotemporal distribution of frequency.

Abstract

This invention discloses a practical aggregation method for power system frequency models that considers the spatiotemporal distribution of frequencies. It addresses the current lack of a method to aggregate the entire frequency process of a power system without relying on electromechanical simulation, even when system parameters are incomplete. The invention first constructs a power disturbance set for the power system, then calculates the initial unbalanced power and initial steady-state frequency of each node under a specific disturbance. Based on the initial unbalanced power and initial steady-state frequency of each node, the frequency difference and unbalanced power of each node are calculated alternately until each node reaches a uniform steady-state frequency, obtaining the frequency difference trajectory between nodes under a specific disturbance. Finally, clustering is performed based on the calculated frequency difference trajectories of each node.

Description

Technical Field

[0001] This invention belongs to the field of novel power system modeling and control technology, specifically relating to a practical aggregation method for power system frequency models. Background Technology

[0002] In recent years, renewable energy has developed rapidly, and the new power system exhibits the "dual high" characteristics of high power electronics and high proportion of new energy sources. New energy power generation, such as wind and solar power, is characterized by randomness and time-varying nature. Furthermore, the long electrical distance between new energy power plants and the main grid, along with large-scale system interconnections, makes the spatiotemporal frequency distribution characteristics of the power system more pronounced. To ensure the safe and stable operation of the new power system, dividing the power system frequency into zones and aggregating and reducing the order of the frequency model are prerequisites for assessing the frequency stability of the power system.

[0003] Currently, there are numerous studies on the spatiotemporal frequency distribution characteristics of power systems, such as:

[0004] 1. Pei Ming et al. published "Inertia Estimation Method for New Energy Power Systems Considering Spatiotemporal Correlation of Frequency Response" Electric Power System Automation: 53-66 [2024-04-25]. The article studies the calculation method of inertia for synchronous generator units and new energy generator units. It uses the Granger causality test algorithm to dynamically analyze the frequency correlation between different nodes in the system, constructs a time-varying spatiotemporal causal correlation set of system frequency, establishes a system inertia constant-frequency state space model based on the swing equation of the frequency response process of each node in the system, designs an unscented Kalman filter and a fixed hysteresis smoother, and proposes an inertia estimation method for high-proportion new energy power systems. The study only considers the partitioning and estimation of inertia, without considering the whole process of system frequency response.

[0005] 2. Xu Taishan and Xue Yusheng published “Quantitative Analysis of Acceptability of Transient Frequency Deviation” in Automation of Electric Power Systems: 7-10 [2002, 26(19)]. The article proposes a binary table evaluation index based on transient frequency deviation threshold and deviation duration, which is used to evaluate the global frequency stability of the power system, but does not consider the node frequency security index of the power system.

[0006] 3. Larbi ES et al., in their paper "Evaluation of key performance indices for frequency quality: a method for evaluating frequency stability in the Nordic power system" (Stockholm, Sverige: KTH Royal Institute of Technology

[2023] ), using the Nordic power system as an example, employed correlation analysis to show that frequency, standard deviation, frequency area, frequency exceedance frequency, and R... max The standard deviation and four indicators can comprehensively measure the security of node frequencies, but they only consider node frequency indicators and lack analysis of global frequency security and stability indicators.

[0007] 4. “Analysis of the Spatiotemporal Distribution Characteristics of Frequency in New Energy Power Systems” by Ma Ningjia et al., High Voltage Engineering: 406-413 [2014-01-31]. This article establishes an analytical model of the frequency dynamic response of new energy power systems and analyzes the influencing factors of the spatiotemporal distribution characteristics of frequency and the influence mechanism of new energy grid connection on these characteristics. The article discusses the necessity of studying the spatiotemporal distribution characteristics of frequency and the quantitative indicators for describing these characteristics, and summarizes the influence law of new energy unit grid connection on these characteristics. However, it does not consider the aggregation and order reduction of the frequency response equation for the spatiotemporal distribution characteristics.

[0008] In summary, current research on the spatiotemporal distribution characteristics of power system frequencies focuses on qualitative analysis of numerical examples and research on global or nodal frequency security assessment indicators, lacking practical methods for aggregated equivalence of the entire frequency response process. Summary of the Invention

[0009] This invention aims to address the current lack of a method to aggregate the entire frequency process of a power system without relying on electromechanical simulation, even when system parameters are incomplete.

[0010] A practical aggregation method for power system frequency models that considers the spatiotemporal distribution of frequency includes the following steps:

[0011] S1. Construct a power disturbance set for the power system. The disturbance set includes disturbance scenarios consisting of single disturbances and disturbance scenarios consisting of composite disturbances.

[0012] S2. Calculate the initial unbalanced power and steady-state frequency of each node in the power system under a specific disturbance;

[0013] S3. Based on the initial unbalanced power and steady-state frequency of each node in the power system, the frequency difference and unbalanced power of each node are calculated alternately until each node reaches a uniform steady-state frequency, thereby obtaining the frequency difference trajectory between each node in the power system under a specific disturbance.

[0014] S4. Based on the calculated trajectory of the frequency difference of each node, perform clustering according to the trajectory of the frequency difference of each node.

[0015] Furthermore, the disturbance types in the disturbance scenario of S1 include impact loads, generator tripping, sudden changes in the operating conditions of wind power or photovoltaic power plants, and line tripping.

[0016] Furthermore, the calculation process for the initial unbalanced power in S2 includes:

[0017] For thermal power units, the node voltage equations are determined based on the power system admittance matrix before the impact load change; the voltage vector at the grid connection point of new energy power plants is represented by V. R This indicates that the new energy power station is represented as a controlled current source, with an equivalent current of I. R The node voltage equations corresponding to the extended thermal power unit are obtained as follows:

[0018]

[0019] In the formula, the block diagonal matrix represents the self-admittance matrix, and the off-diagonal block matrix represents the mutual admittance matrix, Y GG Y is the self-admittance matrix of the generator node of a thermal power unit. RR Y is the self-admittance matrix of the new energy power station node. BB Y is the self-admittance matrix of the load node. GR Y RG Y is the mutual admittance matrix of thermal power nodes and renewable energy plant nodes. GB Y BG Y is the mutual admittance matrix between the thermal power node and the load node. RB Y BR I is the mutual admittance matrix between new energy power station nodes and load nodes; G Generator injection current, V B It is the node voltage of the load node; V G V L These are the voltages of the generator busbars, respectively.

[0020] According to V G I R Solve I G V R V B Then according to V G I R and the solution obtained I G VR V B The initial unbalanced power of each power node is calculated.

[0021] Furthermore, the admittance matrix of the power system before the change in the impact load is:

[0022]

[0023] Among them, X d ' is the subsynchronous reactance of the generator, X T Y is the equivalent reactance of the line and transformer. T Y represents the equivalent admittance of the line and transformer; L1 This is the equivalent admittance before the sudden increase in load.

[0024] Furthermore, the nodal voltage equations determined based on the admittance matrix of the power system before the change in impact load are as follows:

[0025]

[0026] In the formula, V G V L These are the voltages of the generator bus and the load node, respectively; E is the electromotive force within the synchronous machine.

[0027] Furthermore, the calculation process for the steady-state frequency described in S2 includes:

[0028] First, determine the equivalent droop coefficient of the power system. Then, determine the offset of the steady-state frequency of the power system based on the equivalent droop coefficient. Finally, determine the steady-state frequency under a specific disturbance by using the offset of the steady-state frequency of the power system.

[0029] The deviation of the steady-state frequency of the power system is:

[0030]

[0031] In the formula, Δf ∞ The shift in the steady-state frequency of the power system after the disturbance; ΔP d It is the disturbance power of the power system.

[0032] Furthermore, the equivalent droop coefficient of the power system is:

[0033]

[0034] In the formula, R i R represents the droop factor of power supply i. eq This represents the equivalent droop coefficient of the power system.

[0035] Furthermore, the process of alternately calculating the frequency difference and unbalanced power of each node as described in S3 includes the following steps:

[0036] The formula for calculating the frequency difference between nodes is determined based on the DC power flow method:

[0037]

[0038] Where, in the formula, Δf ij B represents the difference in frequency between node i and node j. RN The system admittance matrix is ​​ΔP, where ΔP is the node shrinkage matrix. ij Δt represents the increase in electromagnetic power transmitted from node i to node j; Δt represents the time change and is used to represent the iteration step size.

[0039] Then, based on the frequency difference formula between nodes, the frequency difference of each power supply node is iteratively calculated, and the unbalanced power of each node is updated according to the known inertia and droop coefficient of each node and the inter-machine transmission power.

[0040] Determine whether each node has reached a uniform steady-state frequency; if so, stop the iteration.

[0041] Furthermore, the conditions for each node to reach a uniform steady-state frequency are as follows:

[0042]

[0043] Among them, f i f is the frequency of node i. ∞ The steady-state frequency of the system is δ; δ1 and δ2 represent the tolerances.

[0044] When formula (17) is true, it means that each node has reached a uniform steady-state frequency.

[0045] Furthermore, the process of clustering based on the calculated trajectories of frequency differences between each node, as described in S4, includes:

[0046] Select the clustering condition interval [-ε, ε], where ε is the clustering boundary point, and the clustering condition is:

[0047]

[0048] When formula (18) is satisfied, the frequency trajectories of node i and node j are similar, and the frequency models of node i and node j are aggregated.

[0049] The beneficial effects of this invention are:

[0050] For power system frequency models that consider the spatiotemporal distribution of frequencies, a method is proposed that does not rely on electromechanical simulation. Even with incomplete system parameters, it can cluster frequencies for specific disturbances and reduce the order of the frequency model through clustering. This method not only aggregates the entire frequency process of the power system without relying on electromechanical simulation, but also improves the speed of estimating the system frequency. It provides a practical method for reducing the order of power system frequency response models that consider the spatiotemporal distribution of frequencies, and provides theoretical and technical support for improving the frequency security and stable operation of new power systems. Attached Figure Description

[0051] Figure 1 The frequency response transfer function diagram of the single center of inertia system of the present invention is shown.

[0052] Figure 2 This is a graph of the system frequency response transfer function for calculating the spatiotemporal distribution of the present invention;

[0053] Figure 3 This is a flowchart of the algorithm of the present invention. Detailed Implementation

[0054] The invention will now be further described with reference to the accompanying drawings.

[0055] This invention presents a practical aggregation method based on a power system frequency model considering the spatiotemporal distribution of frequencies. This model includes frequency response models for thermal power units, wind power units, and photovoltaic units. The power system frequency model utilizes the DC power flow method to characterize the power flow. This method does not rely on large-scale electromechanical simulations and can perform partitioned aggregation of power system frequencies for specific disturbances, effectively simplifying the power system frequency model considering the spatiotemporal distribution of frequencies. This provides a theoretical basis for analyzing the spatiotemporal distribution of power system frequencies and technical support for improving the safe operation of the power grid. Detailed descriptions of specific implementation methods follow.

[0056] The frequency response of a power system is mainly related to the system's disturbances and the frequency response characteristics of each power source. A single center of inertia power system refers to a power system with a uniform frequency, lacking obvious spatiotemporal distribution characteristics. Its model is relatively simple and easy to analyze. The frequency response transfer function diagram of a single center of inertia power system is attached. Figure 1As shown, this invention establishes a frequency response model for a single-center-of-inertia power system, which can characterize the frequency response characteristics of thermal power units, wind farms, and photovoltaic power plants. The prime mover of the thermal power unit is characterized by the transfer function of a reheat turbine, and the synchronous generator is represented by a second-order model. The wind turbine model considers load shedding control and frequency regulation functions, including virtual inertia control and droop control. The photovoltaic unit model considers overvoltage load shedding and frequency regulation functions, including virtual inertia control and droop control. The model covers common new energy power plant types and frequency control methods in current power systems and has universal significance.

[0057] The DC power flow method assumes that the phase angle difference between nodes is small, and linearizes the active power flow equations. When studying frequency problems, the DC power flow method can incorporate the power oscillations between power systems with a single center of inertia into the frequency model. The resulting power system frequency model considering the spatiotemporal distribution is shown in the appendix. Figure 2 The spatiotemporal frequency model of the power system established in this invention considers the frequency characteristics of thermal power units and new energy power plants, and can effectively characterize the spatiotemporal frequency distribution characteristics of the power system. The power system model considering spatiotemporal frequency distribution is node-based and takes the form of state equations. The order of these equations is relatively high, and analyzing them requires first performing aggregation and order reduction on the model.

[0058] The core idea of ​​a practical aggregation method for power system frequency models that take into account the spatiotemporal distribution of frequency is as follows: calculate the initial unbalanced power of each power node for a specific disturbance, calculate the frequency difference between each node iteratively, cluster the power system according to the obtained frequency difference trajectory between each node, establish the mapping relationship between the disturbance and the system frequency partition, and thus aggregate and reduce the order of the power system frequency model that takes into account the spatiotemporal distribution of frequency.

[0059] A practical aggregation method for power system frequency models that take into account the spatiotemporal distribution of frequency includes the following steps:

[0060] S1. Construct a power disturbance set for the power system. The disturbances corresponding to the disturbance set include disturbance scenarios consisting of single disturbances and disturbance scenarios consisting of composite disturbances. Specific disturbance types include impact loads, generator tripping, sudden changes in the operating conditions of wind power or photovoltaic power stations, line tripping, etc.

[0061] S2. Calculate the initial unbalanced power and steady-state frequency of each node in the power system under a specific disturbance;

[0062] The calculation process for the initial unbalanced power is as follows:

[0063] Considering a single-machine-load system, the power system consists of synchronous generators and static loads, and the equivalent load admittance is Y. L The equivalent admittances before and after the sudden load increase are Y, respectively. L1 Y L2The Norton equivalent injection current of the synchronous generator is E / X. d ', E is the internal potential of the synchronous machine, X d ' is the subsynchronous reactance of the generator, X T Y is the equivalent reactance of the line and transformer. T The equivalent admittance for the line and transformer;

[0064] Equivalent reactance of lines and transformers Where X TR Let R be the equivalent reactance of the transformer, and X be the equivalent resistance and reactance of the line, respectively. j represents an imaginary number.

[0065] The admittance matrix of the power system before the change in impact load is:

[0066]

[0067] According to the node voltage equation:

[0068]

[0069] In the formula, V G V L These are the voltages of the generator bus and the load node, respectively.

[0070] The calculated generator bus voltage is:

[0071]

[0072] The electromagnetic power P output by the generator without an automatic voltage control device e for:

[0073]

[0074] The electromagnetic power P output by the generator when an automatic voltage control device is present e The expression is:

[0075]

[0076] Formula (4) is for the case without an automatic voltage control device, and Formula (5) is for the case with an automatic voltage control device.

[0077] Formula (5) needs to be derived from formula (4); from formula (4), it can be seen that in the absence of an automatic generator voltage control device, under power loss conditions such as power system impact load or power system generator tripping, the output electromagnetic power of the generator decreases. In a simple system with an automatic voltage control device, the impact load increases suddenly, the admittance corresponding to the load increases, and the automatic voltage control device maintains the generator voltage bus voltage V. G Unchanged, in VG Without changing the parameters, the electromagnetic power generated by the generator increases.

[0078] The time from the occurrence of a fault to the restoration of bus voltage for a typical automatic voltage control device is 0.1 to 0.3 seconds.

[0079] From the occurrence of the fault to the maintenance of the bus voltage by the automatic voltage control device, the electromagnetic power of the generator first decreases and then increases. Ignoring the dynamics of the automatic voltage control device, assuming that the bus voltage remains constant, the electromagnetic power output by the generator immediately increases after the impact load increases. Current thermal power units all have automatic voltage regulation devices. Therefore, assuming that the terminal voltage of the thermal power unit remains constant before and after the fault, the initial unbalanced power of the thermal power unit can be obtained through the nodal voltage equation.

[0080] The above only considers the case of thermal power units. Grid-connected renewable energy units do not have the ability to maintain terminal voltage similar to excitation control devices. The voltage vector at the grid connection point of renewable energy power plants is represented by V. R This indicates that the new energy power station is represented as a controlled current source, with an equivalent current of I. R The current value cannot change abruptly, extended formula (2):

[0081]

[0082] In the formula, the block diagonal matrix represents the self-admittance matrix, and the off-diagonal block matrix represents the mutual admittance matrix, Y GG Y is the self-admittance matrix of the generator node of a thermal power unit. RR Y is the self-admittance matrix of the new energy power station node. BB Y is the self-admittance matrix of the load node. GR Y RG Y is the mutual admittance matrix of thermal power nodes and renewable energy plant nodes. GB Y BG Y is the mutual admittance matrix between the thermal power node and the load node. RB Y BR This is the mutual admittance matrix between new energy power station nodes and load nodes.

[0083] V G I R Given a quantity, solve for I. G V R V B :

[0084]

[0085] σ1=Y BB Y GG Y RR -Y BB Y GR Y RG-Y BG Y GB Y RR +Y BG Y GR Y RB +Y BR Y GB Y RG -Y BR Y GG Y RB (7)

[0086] I G =Y GG V G +Y GR V R +Y GB V B

[0087] Among them, I G Generator injection current, V B It is the node voltage of the load node;

[0088] According to V G I R and the solution obtained I G V R V B The unbalanced power of each power node, i.e. the initial unbalanced power, can be calculated.

[0089] The steady-state frequency of a power system is calculated. The frequencies at all nodes in the system tend to be a uniform steady-state frequency, which depends on the droop coefficient of each power source. The equivalent droop coefficient of the power system is:

[0090]

[0091] In the formula, R i R represents the droop factor of power supply i. eq This represents the equivalent droop coefficient of the power system;

[0092] The deviation of the steady-state frequency of the power system is:

[0093]

[0094] In the formula, Δf ∞ The shift in the steady-state frequency of the power system after the disturbance; ΔP d It is the disturbance power of the power system;

[0095] Then, the steady-state frequency under a specific disturbance is determined by the deviation of the steady-state frequency of the power system.

[0096] S3. Determine the time iteration step size. Starting from the initial state, alternately calculate the frequency difference and unbalanced power of each node until each node reaches a uniform steady-state frequency. Stop the iteration to obtain the frequency difference trajectory between each node of the power system under a specific disturbance.

[0097] The basic idea for obtaining the frequency difference trajectory between nodes of a power system under a specific disturbance is to highlight the admittance matrix of the power system, i.e., the role of its spatiotemporal distribution characteristics in the power transfer between nodes after the disturbance occurs, and to simplify the frequency response of each power source; and to use the initial unbalanced power distribution after the disturbance as the initial condition to iteratively calculate the frequency difference trajectory of each node.

[0098] The dynamic process of power system frequency is determined not only by the system's unbalanced power but also by the frequency response characteristics of each power source. Since the parameters of each power source are incomplete, to highlight the influence of the spatiotemporal frequency distribution, each power source is simplified using a model containing only inertia and droop (a prerequisite known in the field). Numerical examples show that this simplification does not affect the frequency partitioning of the power system, and the inertia of each node can be estimated online using wide-area measurement devices. The power system eventually reaches a uniform steady-state frequency, determined by the droop coefficient of each power source. A practical aggregation method for calculating the spatiotemporal distribution of power system frequency models is presented, using an iterative approach.

[0099] Combined with appendix Figure 3 The principle and process of the iteration process are as follows:

[0100] Given the admittance matrix Y1 of the power system before and after the disturbance, and the equivalent inertia K of each node. ii The power of the disturbance ΔP d The droop coefficient K of each thermal power unit and new energy power station di The iterative calculation process for inter-node frequency differences and unbalanced power is as follows:

[0101] S301. Derive the formula for calculating the frequency difference between nodes based on the DC power flow method;

[0102] The inter-machine oscillation power determined by the DC power flow equation is:

[0103]

[0104] Where, ΔP ei B represents the change in electromagnetic power emitted by node i. RNij Let Δω be the contraction mutual admittance between node i and node j. i Let Δω be the change in electric angular velocity at node i. j Let n be the change in electric angular velocity at node j. g This represents the number of power supply nodes in the system.

[0105] Express the above equation in matrix form:

[0106]

[0107] In the formula, Δf ij B represents the difference in frequency between node i and node j. RN The system admittance matrix is ​​ΔP, where ΔP is the node shrinkage matrix. ij Δt represents the increase in electromagnetic power transmitted from node i to node j; Δt represents the time change, i.e., the iteration step size.

[0108] Discretizing the above equation yields the method for calculating the frequency difference between nodes:

[0109]

[0110] S302. Iteratively calculate the frequency difference of each power node, determine the iteration step size Δt, and begin the first iteration:

[0111]

[0112] Where, ΔP ij 0 For ΔP ij The initial value, with the superscript 0 indicating the number of iterations;

[0113] Update the unbalanced power of each node based on the known inertia, droop coefficient, and inter-machine transmission power:

[0114]

[0115] Where, ΔP i 1 , Δf i 1 The superscript 1 indicates the iteration number, that is, the ΔP corresponding to the first iteration. i , Δf i ΔP i Let Δf be the unbalanced power at node i. i K represents the frequency offset of node i; ii K is the equivalent inertia constant of node i or the virtual inertia coefficient of the new energy power station. di P is the equivalent droop coefficient of node i or the droop control coefficient of the virtual station; ji Let Ω represent the electromagnetic power transmitted by node i under the influence of connected node j, and let Ω represent the set of all node numbers connected to node i.

[0116] S303, In the k-th iteration, calculate the node frequency difference matrix in the (k+1)-th step:

[0117]

[0118] Update the unbalanced power of each node:

[0119]

[0120] S304. Based on the set iteration termination condition, determine whether the iteration has terminated; that is, stop the iteration when each node reaches a uniform steady-state frequency.

[0121] Selecting tolerances δ1 and δ2, calculate the 2-norm of the frequency difference matrix. The iteration termination condition is:

[0122]

[0123] Among them, f i f is the frequency of node i. ∞ This is the steady-state frequency of the system;

[0124] If all the above equations are true, it means that each node has reached a uniform steady-state frequency. If the condition is not met, continue iterating until the condition is met, then terminate the iteration.

[0125] S4. Based on the calculated trajectory of the frequency differences of each node, define the clustering conditions and perform clustering based on the trajectory of the frequency differences of each node.

[0126] Select the clustering condition interval [-ε, ε], where ε is the clustering boundary point, and the clustering condition is:

[0127]

[0128] Since the frequency trajectories of nodes i and j are similar, they can be classified into the same category, indicating that under this perturbation, the frequency models of nodes i and j can be aggregated.

[0129] Thus, a practical aggregation method for power system frequency models that takes into account the spatiotemporal distribution of frequency has been formed.

[0130] The above examples of the present invention are merely illustrative of the computational model and process of the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is impossible to exhaustively list all possible implementations here. Any obvious variations or modifications derived from the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A method for aggregating power system frequency models that take into account the spatiotemporal distribution of frequency, characterized in that, Includes the following steps: S1. Construct a power disturbance set for the power system. The disturbance set includes disturbance scenarios consisting of single disturbances and disturbance scenarios consisting of composite disturbances. S2. Calculate the initial unbalanced power and steady-state frequency of each node in the power system under a specific disturbance; S3. Based on the initial unbalanced power and steady-state frequency of each node in the power system, the frequency difference and unbalanced power of each node are calculated alternately until each node reaches a uniform steady-state frequency, thereby obtaining the frequency difference trajectory between nodes in the power system under a specific disturbance; the process of alternately calculating the frequency difference and unbalanced power of each node includes the following steps: The formula for calculating the frequency difference between nodes is determined based on the DC power flow method: （12） In the formula, Represents a node and nodes The difference in frequency, The system admittance matrix is ​​the result of node shrinkage. For nodes Towards The increase in transmitted electromagnetic power; It represents the change over time and is used to indicate the iteration step size; Then, based on the frequency difference formula between nodes, the frequency difference of each power supply node is iteratively calculated, and the unbalanced power of each node is updated according to the known inertia and droop coefficient of each node and the inter-machine transmission power. Determine whether each node has reached a uniform steady-state frequency; if so, stop iterating. S4. Based on the calculated trajectory of the frequency difference of each node, perform clustering according to the trajectory of the frequency difference of each node.

2. The aggregation method for power system frequency models considering the spatiotemporal distribution of frequency according to claim 1, characterized in that, The disturbance types in the S1 disturbance scenario include impact loads, generator tripping, sudden changes in the operating conditions of wind or photovoltaic power plants, and line tripping.

3. The aggregation method for power system frequency models considering the spatiotemporal distribution of frequency according to claim 1, characterized in that, The calculation process for the initial unbalanced power in S2 includes: For thermal power units, the node voltage equations are determined based on the admittance matrix of the power system before the impact load change; the voltage vector at the grid connection point of new energy power plants is then used... This indicates that the new energy power station is represented as a controlled current source, with an equivalent current of... The node voltage equations corresponding to the extended thermal power unit are obtained as follows: （6） In the formula, the block diagonal matrix represents the self-admittance matrix, and the off-diagonal block matrix represents the mutual admittance matrix. Let be the self-admittance matrix of the generator node of the thermal power unit. For the self-admittance matrix of the new energy power station nodes, Let be the self-admittance matrix of the load node. , For the mutual admittance matrix of thermal power nodes and new energy power plant nodes, , This is the mutual admittance matrix between the thermal power node and the load node. , This is the mutual admittance matrix between new energy power station nodes and load nodes; Generator injection current, It is the node voltage of the load node; This refers to the voltage of the generator bus. according to , Solve , , And then according to , and the solution obtained , , The initial unbalanced power of each power node is calculated.

4. The aggregation method for power system frequency models considering spatiotemporal frequency distribution according to claim 3, characterized in that, The admittance matrix of the power system before the change of the impact load is: （1） in, For the generator's subsynchronous reactance, The equivalent admittance for the line and transformer; This is the equivalent admittance before a sudden increase in load.

5. The aggregation method for power system frequency models considering spatiotemporal frequency distribution according to claim 4, characterized in that, The nodal voltage equations determined based on the admittance matrix of the power system before the change in impact load are as follows: （2） In the formula, , These are the voltages of the generator bus and the load node, respectively. This is the internal potential of the synchronizing machine.

6. The aggregation method for power system frequency models considering spatiotemporal frequency distribution according to claim 1, characterized in that, The calculation process of the steady-state frequency described in S2 includes: First, determine the equivalent droop coefficient of the power system. Then, determine the offset of the steady-state frequency of the power system based on the equivalent droop coefficient. Finally, determine the steady-state frequency under a specific disturbance by using the offset of the steady-state frequency of the power system. The deviation of the steady-state frequency of the power system is: （9） In the formula, The shift in the steady-state frequency of the power system after the disturbance; It is the disturbance power of the power system. This represents the equivalent droop coefficient of the power system.

7. The aggregation method for power system frequency models considering spatiotemporal frequency distribution according to claim 6, characterized in that, The equivalent droop coefficient of the power system is: （8） In the formula, Indicates power supply The droop coefficient.

8. The aggregation method for power system frequency models considering spatiotemporal frequency distribution according to claim 7, characterized in that, The conditions for determining that all nodes have reached a uniform steady-state frequency are as follows: （17） in, For the frequency of node i, This is the steady-state frequency of the system; Indicates tolerance; When all of the formulas (17) are true, it means that each node has reached a uniform steady-state frequency.

9. The aggregation method for power system frequency models considering spatiotemporal frequency distribution according to claim 8, characterized in that, The process described in S4, which involves clustering the trajectories based on the calculated frequency differences of each node, includes: Selecting clustering condition intervals , The boundary point for clustering is defined by the following clustering condition: （18） When formula (18) is satisfied, then the node and nodes Since the frequency trajectories are similar, the frequency models of nodes i and j are aggregated.

Images