A power flow prediction method for a new energy power system considering the uncertainties of the power source, grid, and load

Abstract

The present invention relates to a method for predicting the flow of a new energy power system that takes into account source-grid-load uncertainty. The specific steps are as follows: normalizing historical parameters of the power grid, extracting multidimensional trajectory features through variational mode decomposition, and constructing an operating trend set; combining a cellular automaton model to simulate spatiotemporal evolution and optimize the fault feature recognition mechanism; constructing a trend state classification database based on the residual transformation rate; dynamically assigning state contribution weights using a multi-constraint regression model to establish a prediction model; iteratively optimizing the model through error feedback, and constructing a vulnerability assessment system based on sensitivity and interference analysis. The present invention integrates multi-source data with dynamic weight allocation, and through a two-way feedback optimization framework, solves the problems of low traditional prediction accuracy and delayed fault response. It can improve the accuracy of power grid trend prediction and vulnerability assessment, providing support for security defense.

Description

Technical Field

[0001] The present invention relates to the field of power system control technology, and in particular to a new energy power system power flow prediction method taking into account source-grid-load uncertainty. Background Art

[0002] With the large-scale development of renewable clean energy, the commissioning of the "Three Hua" UHV synchronous power grid, and the commissioning of UHVDC transmission projects at large energy bases, the operation and dispatch of power grids are facing enormous challenges. The increasing number of interconnected large power grids and the strengthening of electrical connections have led to more complex operational characteristics and significantly increased the difficulty of grid dispatch and control. The integration of large-scale renewable energy sources such as wind and solar power, demand-side distributed generation, and the rapid development of electric vehicles have created significant uncertainty, significantly increasing the uncertainty of grid dispatch and control. Furthermore, extreme weather events such as ice storms, typhoons, earthquakes, and floods have severely impacted the safe and stable operation of power grids. Summary of the Invention

[0003] In order to cope with the uncertainty brought to the grid operation by the high penetration rate of renewable energy access and random demand-side response. With the help of artificial intelligence algorithms such as machine learning and deep learning, combined with the control of big data, by learning from massive training sample data, the future operation trend of the grid flow is predicted, which has important practical significance for ensuring the safe and stable operation of the grid. Based on this, the present invention provides a new energy power system flow prediction method taking into account the uncertainty of source-grid-load, including the following steps:

[0004] Preprocess the historical parameters of the power grid, extract multi-dimensional trajectory features through variational mode decomposition, and construct a set of power grid operation trends;

[0005] Combining the cellular automaton model to simulate the spatiotemporal evolution law, the grid operation trend model is iteratively optimized to obtain the grid operation trend evolution mechanism and mathematical criteria under different fault conditions;

[0006] Constructing a state recognition database for power grid operation trends based on residual transformation rates;

[0007] Using a multi-constraint regression model to dynamically assign state contribution weights, a power grid operation trend prediction model is established;

[0008] The model is iteratively optimized through error feedback, and a vulnerability evaluation system is constructed based on sensitivity and interference analysis to complete the quantitative analysis of power grid vulnerability.

[0009] Specifically, the preprocessing of historical parameters of the power grid, extracting multidimensional trajectory features through variational mode decomposition, and constructing a power grid operation trend set specifically include:

[0010] The historical operation parameter trajectories of the power grid are normalized, and combined with the uncertainty analysis model of the power grid operation parameter trajectory, the specificity between the power grid operation electrical quantities and the external parameters affecting the power grid operation status under different power grid operation states is analyzed. The mathematical statistics model and clustering algorithm are used to classify the power grid parameters, and the power grid operation parameter trajectory clusters composed of power grid operation electrical quantities and external parameters under different categories are determined; based on the variational mode decomposition algorithm, the frequency domain characteristics of the power grid operation parameter trajectories under different center frequencies are extracted to form a power grid operation trajectory feature information set, and the power grid operation parameter trajectory clusters and their multidimensional feature sets are defined as power grid operation trends. Data from different sources, different time scales, and different qualities are integrated to form a power grid operation trend set.

[0011] Specifically, the power grid operation trend set The mathematical expression is as follows:

[0012] ;

[0013] Where, refers to the cluster of power grid operation parameter trajectories, It refers to the n Nodes at time t The collective expression of the power grid operation parameter trajectory cluster at It refers to the q The grid operating parameters are n In the nodes t The value at the time, T Analyze the period of power grid operation trend; , refers to the cluster feature set of power grid operation trajectory, Refers to a power grid operation trend analysis cycle T Neidi n The grid operation parameter feature set of nodes, where Refers to a power grid operation trend analysis cycle T Neidi n The node q The first parameter of the power grid operation j Features, It is a collection of power grid operation trends.

[0014] Specifically, the cellular automaton model is combined to simulate the spatiotemporal evolution law, and the grid operation trend model is iteratively optimized to obtain the grid operation trend evolution mechanism and mathematical criteria under different fault states. Specifically, the following are included:

[0015] Based on the calculated power grid operation trend set and the basic spatiotemporal evolution law of power grid operation trends, the Agent-Cellular Automation model is adopted. With the spatial distribution and temporal evolution law of power grid operation trends as constraints, the dynamic spatiotemporal distribution process of power grid operation trends is simulated. The law of power grid operation trends changing with time in the same spatial region is characterized, the time domain variation characteristics of power grid operation trends are analyzed, and the spatial distribution law of power grid operation trends in different spatial regions at the same time is characterized. The spatial correlation characteristics of power grid operation trends are analyzed. Combined with the differences in the evolution models under normal and fault conditions, the power grid operation trend model is iteratively optimized, and the characteristic parameter set of power grid operation trends under fault conditions is accurately identified. Finally, the evolution mechanism and mathematical criteria of power grid operation trends under different fault conditions are obtained.

[0016] Specifically, the agent-cellular automaton model integrates the statistical laws of power grid operation trends at the macro level with the interactive behaviors between various agents at the micro level, discretizing the power grid space into a two-dimensional grid, forming the following spatial pattern:

[0017] ;

[0018] in, represents the cell in row ii and column jj, each cell represents a geographical area in the power grid, and the model uses a discrete time step Conduct dynamic evolution;

[0019] Each cell Contains one or more agents, defined as:

[0020] ;

[0021] in, Represents a cell The first k Agents represent key grid components such as loads, generators, and buses in the area;

[0022] Each Agent at time t Have state variables , which is used to represent the characteristic quantity of the power grid operation trend. The state transition rule is:

[0023] ;

[0024] in: Indicates external influencing factors; It is a state transfer function built based on the physical characteristics of the power grid and the agent behavior mechanism; for The set of states of a cell and its neighborhood is defined as the Moore neighborhood:

[0025] ;

[0026] The temporal evolution of the grid operation trend depends on the interaction with the surrounding neighborhood and external disturbances. The specific model expression is:

[0027] ;

[0028] in: α and β are the weight coefficients of neighborhood interaction and external disturbance respectively; For the neighborhood ( m , n ) cell to center cell The spatial impact weight of

[0029] At any time t The spatial distribution of power grid operation trends is constrained by spatial autocorrelation characteristics, and the following spatial variation function is defined: :

[0030] ;

[0031] in: is the normalized spatial weight, reflecting the influence intensity between different neighborhood units;

[0032] The model defines two types of state transfer functions: normal operating condition evolution function and fault condition evolution function ;

[0033] The difference between the two is expressed as:

[0034] ;

[0035] The model is iteratively optimized using historical operating data to minimize the deviation between the prediction and the observed data:

[0036] ;

[0037] in, is the historical observation data, is the set of parameters to be optimized;

[0038] After the optimization is completed, the characteristic parameter set under the fault condition is extracted Based on this, the evolution mechanism and mathematical criteria of the power grid under different fault types are formed:

[0039] ;

[0040] in, and They are the trend mutation threshold and spatial instability threshold, respectively, which are used to accurately identify power grid faults.

[0041] Specifically, the state identification database for power grid operation trends based on the residual transformation rate specifically includes: defining a trend-state space discrimination index based on the mathematical criterion of the power grid operation trend and combining a set of power grid operation trend characteristic parameters, comprehensively considering the state evolution characteristics of the power grid in the cluster, the power grid operation trend characteristics in different spatial regions, and the trend-state space discrimination index, and constructing a power grid operation trend state classification model; arranging in descending order according to the discrimination index, taking the minimum residual transformation rate of the power grid operation trend state division as the optimization goal, determining the power grid operation trend state set that minimizes the residual transformation rate, and finally forming a state identification database for the power grid operation trend.

[0042] Specifically, the trend-state space discrimination index is defined as as follows:

[0043] ;

[0044] in: : Trend characteristics of region i, ji, j at time tt; : the time-average value of the regional power grid trend; : No. i , j spatial dispersion of regional trends; : Abnormal deviation in the trend evolution of region i,ji,j; : is the weight factor of each characteristic index;

[0045] The set of discriminant indices in all cells On this basis, sort by discrimination index from large to small:

[0046] ;

[0047] in For the k Large discriminant index;

[0048] According to the sorting results, different power grid operation trend status clusters are divided ,in L The number of status categories that are initially divided;

[0049] Introducing residual transform rate As the optimization objective; it is defined as follows:

[0050] ;

[0051] in: For the lThe mean of the discriminant index within the trend state cluster; is the mean of the global discriminant index; the goal is to find a η The minimum trend state division scheme is recorded as:

[0052] ;

[0053] The partitioning process adopts dynamic partitioning method, combining hierarchical clustering and trend-state space characteristics to perform iterative partitioning to meet the requirements. ;

[0054] After completing the optimal trend state division, the trend state database is obtained:

[0055] ;

[0056] in, It is a feature data set that includes various power grid operation trend states.

[0057] Specifically, the use of a multi-constraint regression model to dynamically allocate state contribution weights and establish a power grid operation trend prediction model specifically includes: calculating the error sequence of the power grid operation trend prediction based on the power grid operation trend prediction results, defining a prediction error trend factor, and if the prediction error shows an increasing trend, updating the power grid operation trend evolution model; if the prediction error shows a decreasing or normal fluctuation trend, maintaining the current power grid operation trend evolution model unchanged at the next prediction moment; further, based on the error sequence, analyzing the response characteristics of the power grid to small disturbances under different operation trend states, proposing a sensitivity evaluation index for the power grid operation trend trajectory, and combining the influence of internal and external interference of the system to establish a power grid vulnerability evaluation index, using the entropy weight method to weight each trajectory, constructing a power grid vulnerability evaluation model, and finally completing the quantitative analysis of the power grid vulnerability.

[0058] Specifically, the grid vulnerability assessment model is defined as follows:

[0059] ;

[0060] in, is the grid operation parameter trajectory cluster Sensitivity, is the grid operation parameter trajectory cluster The sensitivity of the feature, is the actual value of the disturbance parameter trajectory, is the average value of the disturbance parameter trajectory;

[0061] ;

[0062] in, is the grid vulnerability index, is the weight assigned by the entropy weight method to the parameter, is the occurrence rate of instability events of this parameter in the face of disturbances, is the upper bound of safe operation of the trajectory, is the mean value of the trajectory when the power grid is operating smoothly.

[0063] Beneficial effects of the present invention:

[0064] The present invention integrates multi-source data with dynamic weight allocation, and solves the problems of low traditional prediction accuracy and delayed fault response through a two-way feedback optimization framework. It can improve the accuracy of power grid trend prediction and vulnerability assessment, and provide support for security defense. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The present invention has the following accompanying drawings:

[0066] Figure 1 It is a new energy power system power flow prediction method that takes into account the uncertainty of source-grid-load;

[0067] Figure 2 This is a schematic diagram of the power grid operation trend collection structure. DETAILED DESCRIPTION

[0068] In order to make the objectives, technical solutions and advantages of this application more clearly understood, the following description is given in conjunction with specific embodiments.

[0069] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should have the usual meanings understood by people with ordinary skills in the field to which this application belongs. The "first", "second" and similar words used in the embodiments of the present application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0070] In view of this, if Figure 1 As shown, the embodiment of the present invention proposes a new energy power system power flow prediction method taking into account the uncertainty of source-grid-load, and the implementation steps are as follows:

[0071] Step A. Normalize the historical grid operating parameter trajectories. Combined with the grid operating parameter trajectory uncertainty analysis model, analyze the specificity between grid operating electrical quantities and external parameters affecting grid operating states under different grid operating states (normal and fault states). Utilize mathematical statistics models and clustering algorithms to classify grid parameters and identify clusters of grid operating parameter trajectories composed of grid operating electrical quantities and external parameters under different categories. Based on the variational mode decomposition algorithm, extract the frequency domain characteristics of grid operating parameter trajectories at different center frequencies to form a grid operating trajectory feature information set. Define the grid operating parameter trajectory clusters and their multidimensional feature sets as grid operating trends. By integrating data from different sources, time scales, and quality, a grid operating trend set is formed, laying the foundation for grid operating trend prediction.

[0072] Step B. Based on the calculated grid operation trend set and the basic spatiotemporal evolution laws of grid operation trends, an agent-cellular automaton model is used. With the spatial distribution and temporal evolution laws of grid operation trends as constraints, the dynamic spatiotemporal distribution process of grid operation trends is simulated. The temporal variation of grid operation trends within the same spatial region is characterized, the temporal variation characteristics of grid operation trends are analyzed, and the spatial distribution laws of grid operation trends within different spatial regions at the same time are characterized. The spatial correlation characteristics of grid operation trends are analyzed. Combining the differences in the evolution models under normal and fault conditions, the grid operation trend model is iteratively optimized to accurately identify the characteristic parameter set of grid operation trends under fault conditions. Ultimately, the evolution mechanism and mathematical criteria for grid operation trends under different fault conditions are obtained.

[0073] Step C. Based on the mathematical criteria for grid operation trends and a set of grid operation trend characteristic parameters, a trend-state space discriminant index is defined. This model constructs a grid operation trend state classification model by comprehensively considering the state evolution characteristics of the grid within the cluster, the grid operation trend characteristics within different spatial regions, and the trend-state space discriminant index. The models are sorted in descending order by the discriminant index, and the optimization objective is to minimize the residual transformation rate of the grid operation trend state classification. The set of grid operation trend states that minimizes this residual transformation rate is determined, ultimately forming a grid operation trend state identification database.

[0074] Step D. Based on the power grid operation trend state identification database, a multivariate linear regression model is established with the trend state contribution weight constraint on the power grid as a condition. The residual sum of squares between the overall power grid operation trend and the weighted output characteristics of the power grid operation trend under each state is calculated. The optimal contribution weight is iteratively solved to minimize the residual sum of squares, and the power grid operation trend state contribution weight matrix is obtained. Combined with the dynamic change characteristics of the power grid operation trend, the contribution weight matrix is updated in real time. Finally, a power grid operation trend prediction model is established with the power grid operation trend state characteristics as input and the overall power grid operation trend as output.

[0075] Step E. Based on the grid operation trend forecast results, calculate the error sequence of the grid operation trend forecast and define the forecast error trend factor. If the forecast error shows an increasing trend, return to step B and update the grid operation trend evolution model. If the forecast error shows a decreasing or normal fluctuation trend, maintain the current grid operation trend evolution model unchanged at the next forecast time. Furthermore, based on the error sequence, analyze the grid's response characteristics to small disturbances under different operation trend states, propose a sensitivity evaluation index for the grid operation trend trajectory, and combine the influence of internal and external interference to establish a grid vulnerability evaluation index. Using the entropy weight method to assign weights to each trajectory, construct a grid vulnerability evaluation model, and ultimately complete the quantitative analysis of grid vulnerability.

[0076] like Figure 2 As shown, in a preferred embodiment, in step A, the power grid operation trend set The mathematical expression is as follows:

[0077] ;

[0078] Where, refers to the cluster of power grid operation parameter trajectories, It refers to the n Nodes at time t The collective expression of the power grid operation parameter trajectory cluster at It refers to the q The grid operating parameters are n In the nodes t The value at the time, T Analyze the period of power grid operation trend; It refers to the cluster feature set of power grid operation trajectory. Refers to a power grid operation trend analysis cycle T Neidi n The grid operation parameter feature set of nodes, where Refers to a power grid operation trend analysis cycle T Neidi n The node q The first parameter of the power grid operation j Features. It is a collection of power grid operation trends.

[0079] In a preferred embodiment, in step B, the Agent-Cellular Automation (Agent-CA) model integrates the statistical laws of grid operation trends at the macro level with the interactive behaviors between various agents at the micro level, discretizing the grid spatial area into a two-dimensional grid, forming the following spatial pattern:

[0080] ;

[0081] in, The cell in row ii and column jj is represented by , and each cell represents a geographical area in the power grid. The model uses a discrete time step. Dynamic evolution.

[0082] Each cell Contains one or more agents, defined as:

[0083] ;

[0084] in, Represents a cell The first k Agents represent key grid components such as loads, generators, and buses in the area.

[0085] Each Agent at time t Have state variables , used to represent the characteristic quantity of the power grid operation trend (such as frequency deviation, voltage level or load change, etc.). The state transition rule is:

[0086] ;

[0087] in: Represents external influencing factors (such as renewable energy output fluctuations, load disturbances, etc.); It is a state transfer function built based on the physical characteristics of the power grid and the agent behavior mechanism; for The set of states of a cell and its neighborhood is defined as the Moore neighborhood:

[0088] ;

[0089] The temporal evolution of the grid operation trend depends on the interaction with the surrounding neighborhood and external disturbances. The specific model expression is:

[0090] ;

[0091] in: α and β are the weight coefficients of neighborhood interaction and external disturbance, respectively. For the neighborhood ( m , n ) cell to center cell The spatial influence weight of the grid reflects the topological structure characteristics of the grid.

[0092] At any time tThe spatial distribution of power grid operation trends is constrained by spatial autocorrelation characteristics, and the following spatial variation function is defined: :

[0093] ;

[0094] in: is the normalized spatial weight, reflecting the influence intensity between different neighborhood units; The smaller the value, the stronger the spatial clustering of the trend, and the spatial distribution shows obvious local consistency.

[0095] In order to distinguish the evolution characteristics of the power grid under normal conditions and fault conditions, the model defines two types of state transfer functions: normal condition evolution function and fault condition evolution function

[0096] The difference between the two is expressed as:

[0097] ;

[0098] This difference can reflect the sudden change characteristics of the grid operation trend under fault conditions (such as a sharp drop in frequency, voltage collapse, etc.).

[0099] The model is iteratively optimized using historical operating data to minimize the deviation between the prediction and the observed data:

[0100] ;

[0101] in, is the historical observation data, is the set of parameters to be optimized. The optimization method can adopt swarm intelligence optimization methods such as genetic algorithm and particle swarm optimization.

[0102] After the optimization is completed, the characteristic parameter set under the fault condition is extracted Based on this, the evolution mechanism and mathematical criteria of the power grid under different fault types are formed:

[0103] ;

[0104] in, and They are the trend mutation threshold and spatial instability threshold, respectively, which are used to accurately identify power grid faults.

[0105] In a preferred embodiment, in step C, the trend-state space discrimination index is defined as as follows:

[0106] ;

[0107] in: : Trend characteristics of region i, ji, j at time tt; : the time-average value of the regional power grid trend; : No. i , j spatial dispersion of regional trends; : Abnormal deviation in the trend evolution of region i,ji,j; : is the weight factor of each characteristic index.

[0108] The set of discriminant indices in all cells On this basis, sort by discrimination index from large to small:

[0109] ;

[0110] in For the k Large discrimination index.

[0111] According to the sorting results, different power grid operation trend status clusters are divided ,in L is the number of state categories that are initially divided. In order to optimize the division effect of trend state, the residual transformation rate is introduced η As the optimization goal. It is defined as follows:

[0112] ;

[0113] in: For the l The mean of the discriminant index within the trend state cluster; is the mean of the global discriminant index; The smaller the value, the better the discrimination of the current state division and the more consistent the trend within the cluster. The goal is to find a η The minimum trend state division scheme is recorded as:

[0114] ;

[0115] The partitioning process adopts dynamic partitioning method, combining hierarchical clustering and trend-state space characteristics to perform iterative partitioning to meet the requirements. ( is the convergence threshold of the residual transformation rate).

[0116] After completing the optimal trend state division, the trend state database is obtained:

[0117] ;

[0118] in, It is a feature data set containing various power grid operation trend states, which is used for subsequent fault diagnosis, trend prediction, situation awareness and other applications.

[0119] In a preferred embodiment, in step D, based on the established power grid operation trend status identification database , assuming the overall operation trend of the power grid Y ( t ) is obtained by weighted combination of multiple trend state features, and the following weighted model is established:

[0120] ;

[0121] in: is the weighted output feature of the lth trend state cluster at time t; For the l The contribution weight of each state cluster to the overall trend; is the fitting residual; L is the number of trend states divided.

[0122] Construct a prediction model with trend state characteristics as input and overall operating trend as output:

[0123] ;

[0124] in: For the moment t forecast trends; Identify database via state Real-time extraction; weight Real-time updates ensure the model's ability to adapt to the dynamic characteristics of power grid operation trends.

[0125] In a preferred embodiment, in step E, the grid vulnerability assessment model is defined as follows:

[0126] ;

[0127] in, is the grid operation parameter trajectory cluster Sensitivity, is the grid operation parameter trajectory cluster The sensitivity of the feature, is the actual value of the disturbance parameter trajectory, is the average value of the disturbance parameter trajectory.

[0128] ;

[0129] in, is the grid vulnerability index, is the weight assigned by the entropy weight method to the parameter, is the occurrence rate of instability events of this parameter in the face of disturbances, is the upper bound of safe operation of the trajectory, is the mean value of the trajectory when the power grid is operating smoothly.

Claims

1. A new energy power system power flow prediction method taking into account source-grid-load uncertainty, characterized by: The steps include: Preprocess the historical parameters of the power grid, extract multi-dimensional trajectory features through variational mode decomposition, and construct a set of power grid operation trends; Combining the cellular automaton model to simulate the spatiotemporal evolution law, the grid operation trend model is iteratively optimized to obtain the grid operation trend evolution mechanism and mathematical criteria under different fault conditions; Constructing a state recognition database for power grid operation trends based on residual transformation rates; Using a multi-constraint regression model to dynamically assign state contribution weights, a power grid operation trend prediction model is established; The model is iteratively optimized through error feedback, and a vulnerability evaluation system is constructed based on sensitivity and interference analysis to complete the quantitative analysis of power grid vulnerability.

2. The method according to claim 1, characterized in that The preprocessing of the historical parameters of the power grid, extracting multi-dimensional trajectory features through variational mode decomposition, and constructing a power grid operation trend set specifically include: The historical operation parameter trajectories of the power grid are normalized, and combined with the uncertainty analysis model of the power grid operation parameter trajectory, the specificity between the power grid operation electrical quantities and the external parameters affecting the power grid operation status under different power grid operation states is analyzed. The mathematical statistics model and clustering algorithm are used to classify the power grid parameters, and the power grid operation parameter trajectory clusters composed of power grid operation electrical quantities and external parameters under different categories are determined; based on the variational mode decomposition algorithm, the frequency domain characteristics of the power grid operation parameter trajectories under different center frequencies are extracted to form a power grid operation trajectory feature information set, and the power grid operation parameter trajectory clusters and their multidimensional feature sets are defined as power grid operation trends. Data from different sources, different time scales, and different qualities are integrated to form a power grid operation trend set.

3. The method according to claim 2, characterized in that The power grid operation trend set The mathematical expression is as follows: ； Where, refers to the cluster of power grid operation parameter trajectories, It refers to the n Nodes at time t The collective expression of the grid operation parameter trajectory cluster at It refers to the q The grid operating parameters are n In the nodes t The value at the time, T Analyze the period of power grid operation trend; , refers to the cluster feature set of power grid operation trajectory, Refers to a power grid operation trend analysis cycle T Neidi n The grid operation parameter feature set of nodes, where Refers to a power grid operation trend analysis cycle T Neidi n The node q The first parameter of the power grid operation j Features, It is a collection of power grid operation trends.

4. The method according to claim 1, wherein The above-mentioned method of combining the cellular automaton model to simulate the spatiotemporal evolution law, iteratively optimize the power grid operation trend model, and obtain the power grid operation trend evolution mechanism and mathematical criteria under different fault states specifically includes: Based on the calculated power grid operation trend set and the basic spatiotemporal evolution law of power grid operation trends, the Agent-Cellular Automation model is adopted. With the spatial distribution and temporal evolution law of power grid operation trends as constraints, the dynamic spatiotemporal distribution process of power grid operation trends is simulated. The law of power grid operation trends changing with time in the same spatial region is characterized, the time domain variation characteristics of power grid operation trends are analyzed, and the spatial distribution law of power grid operation trends in different spatial regions at the same time is characterized. The spatial correlation characteristics of power grid operation trends are analyzed. Combined with the differences in the evolution models under normal and fault conditions, the power grid operation trend model is iteratively optimized, and the characteristic parameter set of power grid operation trends under fault conditions is accurately identified. Finally, the evolution mechanism and mathematical criteria of power grid operation trends under different fault conditions are obtained.

5. The method according to claim 4, characterized in that The agent-cellular automaton model combines the statistical laws of power grid operation trends at the macro level with the interactive behaviors between various agents at the micro level, discretizing the power grid space into a two-dimensional grid, forming the following spatial pattern: ； in, represents the cell in row ii and column jj, each cell represents a geographical area in the power grid, and the model uses a discrete time step Conduct dynamic evolution; Each cell Contains one or more agents, defined as: ； in, Represents a cell The first k Agents represent key grid components such as loads, generators, and buses in the area; Each Agent at time t Have state variables , which is used to represent the characteristic quantity of the power grid operation trend. The state transition rule is: ； in: Indicates external influencing factors; It is a state transfer function built based on the physical characteristics of the power grid and the agent behavior mechanism; for The set of states of a cell and its neighborhood is defined as the Moore neighborhood: ； The temporal evolution of the grid operation trend depends on the interaction with the surrounding neighborhood and external disturbances. The specific model expression is: ； in: α and β are the weight coefficients of neighborhood interaction and external disturbance respectively; For the neighborhood ( m , n ) cell to center cell The spatial impact weight of At any time t The spatial distribution of power grid operation trends is constrained by spatial autocorrelation characteristics, and the following spatial variation function is defined: : ； in: is the normalized spatial weight, reflecting the influence intensity between different neighborhood units; The model defines two types of state transfer functions: normal operating condition evolution function and fault condition evolution function ; The difference between the two is expressed as: ； The model is iteratively optimized using historical operating data to minimize the deviation between the prediction and the observed data: ； in, is the historical observation data, is the set of parameters to be optimized; After the optimization is completed, the characteristic parameter set under the fault condition is extracted Based on this, the evolution mechanism and mathematical criteria of the power grid under different fault types are formed: ； in, and They are the trend mutation threshold and spatial instability threshold, respectively, which are used to accurately identify power grid faults.

6. The method according to claim 1, characterized in that The state identification database for power grid operation trends based on the residual transformation rate is specifically constructed as follows: defining a trend-state space discrimination index based on a mathematical criterion for power grid operation trends and a set of power grid operation trend characteristic parameters, comprehensively considering the state evolution characteristics of the power grid in the cluster, the power grid operation trend characteristics in different spatial regions, and the trend-state space discrimination index, and constructing a power grid operation trend state classification model; arranging in descending order according to the discrimination index, taking the minimum residual transformation rate of the power grid operation trend state division as the optimization goal, determining the power grid operation trend state set that minimizes the residual transformation rate, and finally forming a state identification database for power grid operation trends.

7. The method according to claim 6, characterized in that The definition of trend-state space discrimination index as follows: ； in: : Trend characteristics of region i, ji, j at time tt; : the time-average value of the regional power grid trend; : No. i , j spatial dispersion of regional trends; : Abnormal deviation in the trend evolution of region i,ji,j; : is the weight factor of each characteristic index; The set of discriminant indices in all cells On this basis, sort by discrimination index from large to small: ； in For the k Large discriminant index; According to the sorting results, different power grid operation trend status clusters are divided ,in L The number of status categories that are initially divided; Introducing residual transform rate As the optimization objective, it is defined as follows: ； in: For the l The mean of the discriminant index within the trend state cluster; is the mean of the global discriminant index; the goal is to find a η The minimum trend state division scheme is recorded as: ； The partitioning process adopts dynamic partitioning method, combining hierarchical clustering and trend-state space characteristics to perform iterative partitioning to meet the requirements. ; After completing the optimal trend state division, the trend state database is obtained: ； in, It is a feature data set that includes various power grid operation trend states.

8. The method according to claim 1, characterized in that The method of dynamically allocating state contribution weights using a multi-constraint regression model to establish a power grid operation trend prediction model specifically includes: calculating the error sequence of the power grid operation trend prediction based on the power grid operation trend prediction result, defining a prediction error trend factor, and updating the power grid operation trend evolution model if the prediction error shows an increasing trend; if the prediction error shows a decreasing or normal fluctuation trend, maintaining the current power grid operation trend evolution model unchanged at the next prediction moment; further, based on the error sequence, analyzing the response characteristics of the power grid to small disturbances under different operation trend states, proposing a sensitivity evaluation index for the power grid operation trend trajectory, and establishing a power grid vulnerability evaluation index in combination with the influence of internal and external interference of the system, using the entropy weight method to weight each trajectory, constructing a power grid vulnerability evaluation model, and finally completing the quantitative analysis of the power grid vulnerability.

9. The method according to claim 8, characterized in that The grid vulnerability assessment model is defined as follows: ； in, is the grid operation parameter trajectory cluster Sensitivity, is the grid operation parameter trajectory cluster The sensitivity of the feature, is the actual value of the disturbance parameter trajectory, is the average value of the disturbance parameter trajectory; ； in, is the grid vulnerability index, is the weight assigned by the entropy weight method to the parameter, is the occurrence rate of instability events of this parameter in the face of disturbances, is the upper bound of safe operation of the trajectory, is the mean value of the trajectory when the power grid is operating smoothly.

Images