Distributed cooperative AGC method for heterogeneous frequency modulation units in power system
Patent Information
- Application Number
- CN202411484520.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-23
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
Smart Images

Figure CN119010222A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of operation control of power systems, and specifically relates to a distributed collaborative AGC method for heterogeneous frequency modulation units in power systems. Background Art
[0002] In the power system, the distributed coordinated automatic generation control (AGC) of heterogeneous frequency-regulated units is one of the key technologies to ensure the stable operation of the power grid. In the existing technology, the control strategy of the power system usually adopts a centralized control method, which has some limitations when facing large-scale power grids and complex operating conditions.
[0003] The current status of existing technologies includes: centralized control methods rely heavily on global information and require a large amount of data to be transmitted to the central controller for processing, which will cause communication delays and information congestion, affecting the real-time and accuracy of control. Insufficient consideration is given to the characteristics of heterogeneous frequency-modulated units, making it difficult to give full play to the advantages of different types of units, resulting in low resource utilization efficiency. When dealing with uncertainties and interference in the power grid, the adaptability and robustness of existing technologies need to be improved, and it is difficult to cope with rapid changes in load and various abnormal situations.
[0004] To this end, we propose a distributed collaborative AGC method for heterogeneous frequency-regulated units in power systems. Summary of the invention
[0005] In view of the above deficiencies in the prior art, the purpose of the present invention is to provide a distributed collaborative AGC method for heterogeneous frequency-modulation units in a power system, which improves the distributed collaborative capability of power grid control, fully considers the characteristics of heterogeneous frequency-modulation units to achieve optimized collaboration, and enhances the adaptability and robustness of the power grid to uncertainty and interference, thereby improving the real-time and accuracy of control, resource utilization efficiency, and ensuring stable operation of the power grid.
[0006] In order to achieve the above objectives, the present invention adopts the following technical solutions: A distributed collaborative AGC method for heterogeneous frequency regulation units in a power system comprises the following steps: S1. Select a directed graph in graph theory to describe the power grid topology, create a set of nodes and edges, the node set represents the substation, the edge set represents the transmission line, and use the node and edge sets to build a power grid topology model; S2. Load model establishment: Based on the power grid topology structure determined in S1, load models are established at different nodes of the power grid topology structure; S3. Design of distributed time deviation mining algorithm, using the grid structure constructed in S1 and the load model established in S2, to calculate the time deviation index of different nodes and lines under normal operation and abnormal conditions, and to monitor the time deviation index in real time; S4. Deviation compensation control strategy formulation: Based on the time deviation discovered in S3, formulate corresponding control strategies and constraints; S5. Distributed collaborative control strategy, based on the control strategy determined in S4, determines the information interaction and coordination relationship between nodes according to the power grid topology structure constructed in S1; S6. Real-time optimization and adjustment: dynamically adjust the parameters and weights of the control strategy according to the dynamic changes of the load model established in S2 and the deviation monitoring results of the distributed time deviation mining algorithm designed in S3; S7. Establishment of ACE state distributed estimator, using the data obtained in S1 to S6, to establish an ACE state estimator for each node or region; S8. Design a regulator and a controller based on the estimation results of the ACE state estimator established in S7.
[0007] Furthermore, in S1, the method of constructing a power grid topology model using a set of nodes and edges includes: S11. Choose a directed graph in graph theory to describe the topology of the power grid. A directed graph consists of nodes and directed edges. Nodes represent substations, and directed edges represent the direction of transmission lines. S12. Define the node set and edge set: Node Collection It is expressed as: ; Edge Set It is expressed as: ; in, and Represents different substations; is the total number of substations; Represents a substation To another substation transmission line; if there is a substation from To another substation The transmission line ,otherwise ; S13. Construct the adjacency matrix, which is expressed as , used to represent the connection relationship between nodes. If there is an edge ,but ,otherwise , in the adjacency matrix, Line Elements of a column From the substation To the substation Is there a connection? S14. Calculate the node admittance matrix: The admittance matrix is expressed as: ,The admittance matrix is used to describe the admittance characteristics between nodes in the power grid; the elements of the admittance matrix are By conductivity and electrosusceptor Composition, that is, ; in, is an imaginary unit; , Line resistance between two substations; , , is the grid frequency, is the line capacitance, is the line inductance.
[0008] Furthermore, in S2, the method for establishing load models at different nodes of the power grid structure includes: S21. Analyze the node load characteristics, based on the grid structure determined in S1, determine the user type connected to each node and the load characteristics of different user types; S22. Establish a load model function, define load characteristic variables for different time periods and seasons, and use the defined characteristic variables to reflect the load variation over time and season. The load is expressed as: ; in, Indicates load, which is used to describe the load situation of a node or area in the power system at a specific time; is the basic load, For the part that does not change with time and season; For the part that changes with time and season, is the coefficient of the characteristic variable, is the characteristic function, which is used to describe the influence of characteristic variables on load; S23. Select the load forecasting model and use the time series analysis model: ; in, For Node The load forecast value of the node At the moment Load forecasting situation; is the coefficient parameter; used to determine the white noise term The impact on the current load forecast value is different. Corresponding to different coefficients; is the white noise term, indicating random interference factors; It is a constant term in the model and is a parameter related to the base load part that does not change with time and season; The upper limit parameter used for the summation determines the time span of the historical load data involved in the calculation ( Items); is a coefficient parameter used to determine the historical load data The degree of influence on the current load forecast value is different. Corresponding to different coefficients; The upper limit parameter used for the summation determines the number of white noise terms involved in the calculation.
[0009] Furthermore, in said S3, the step of calculating the time deviation index of different nodes and lines under normal operation and abnormal conditions is as follows: S31. Define the time deviation index, which is defined as: ; ; ; in, Indicates the inflow node The set of adjacent nodes of Representation Node The time deviation, Representation Node At the moment The input power, Representation Node At the moment The output power, Representation Node Rated power; Indicates at time Slave Node Inflow Node Power; Indicates outflow node The set of adjacent nodes of Indicates at time Slave Node Flow out to the node Power; S32. Collect grid operation data, use the grid structure constructed in S1 to obtain real-time operation data of each node and line in the grid, and at the same time, combine the load model established in S2 to obtain load forecast data of each node; S33. Calculate the time deviation under normal operation, combine the time deviation index defined in S31 and the real-time operation data of each node and line in the power grid and the load forecast data of each node obtained in S32, and calculate the time deviation of each node: ; ; ; in, For normal operation of the node time deviation; Indicates the node under normal operation At the moment The input power, Indicates the node under normal operation At the moment The output power, Representation Node Rated power; Indicates the inflow node The set of adjacent nodes of Indicates outflow node The set of adjacent nodes of Indicates at time Slave Node Inflow Node Power; For Node The load forecast value; Indicates at time Slave Node Flow out to the node Power; S34. Detect abnormal conditions and set thresholds to determine whether abnormal conditions occur in the power grid; The power fluctuation threshold is set as: or When , it is considered that power fluctuation is abnormal, among which, and Node The average value of input power and output power; and Node Input power and output power; is the threshold of power fluctuation; The voltage abnormality threshold is set as: When the voltage is abnormal, For the moment node The voltage, For Node The average value of the voltage; Threshold indicating voltage anomaly; When an abnormal situation is detected, relevant abnormal information is recorded; S35. Calculate the time deviation under abnormal conditions, calculate the input power and output power of each node according to the real-time operation data and load forecast data, and the time deviation of each node under abnormal conditions: ; In abnormal situations, the node Input power; In abnormal situations, the node Output power; Nodes in abnormal situations Additional input power; Nodes in abnormal situations Additional output power; Indicates at time Slave Node Inflow Node Power; For Node The load forecast value; Indicates at time Slave Node Flow out to the node Power; Indicates the inflow node The set of adjacent nodes of Indicates outflow node The set of adjacent nodes of Representation Node Rated power.
[0010] Further, in said S4, the control strategy is expressed as: ; To control the amount; is the proportionality coefficient, is the integration coefficient, is the differential coefficient, For the moment time deviation; The constraints include: Determine the upper and lower limits of the control quantity according to the rated capacity of the power grid equipment and the stability requirements; The frequency of change of the controlled quantity shall not exceed the threshold; The rate of change of the controlled quantity shall not exceed the limit value; The constraints need to be determined based on the specific characteristics of the power grid, equipment parameters and operating requirements.
[0011] Furthermore, in S5, the information interaction and coordination relationship method between the nodes is determined: S51. Determine the information interaction and coordination relationship. The information interaction between nodes is expressed as: ; For Node At the moment The state of the node, which describes the state of the node after a time step; For Node At the moment Status; For Node Neighbor nodes At the moment Status; For Node The set of neighbor nodes of is the state update function, It is expressed as: ; For Node Its neighbor nodes The weights between them are used to adjust the intensity of information interaction. For Node The amount of control; S52. Control instruction transmission and execution ensures that the control instructions can be accurately and quickly transmitted and executed throughout the power grid. The consistency algorithm is used to make the status of each node gradually consistent, thereby achieving collaborative control. The distributed consensus algorithm is expressed as: ; Representation Node The rate of change of state; For Node Its neighbor nodes The connection weight between them; It is an external input used to guide the state of the node to reach the desired goal; For Node Neighbor nodes At the moment Status; For Node At the moment status.
[0012] Furthermore, in S6, the method for dynamically adjusting the parameters and weights of the control strategy is: S61. Determine the adjustment requirements. According to the dynamic changes of the load model established in S2 and the deviation monitoring results of the distributed time deviation mining algorithm designed in S3, determine whether the parameters and weights of the control strategy need to be adjusted. The adjusted control strategy is: ; is the adjusted control amount; is the adjusted proportional coefficient, is the adjusted integral coefficient; is the adjusted differential coefficient; For the moment time deviation; S62. Adaptive control algorithm, select fuzzy logic control algorithm to automatically adjust the control strategy according to real-time data, the input variable is the deviation and the rate of change of deviation ,right deviation sum The fuzzy set is defined as: "Positive big (PB)", "Positive small (PS)", "Zero (Z)", "Negative small (NS)", "Negative big (NB)"; the output control quantity definition fuzzy set is "Increase (IN)", "Stay (ST)", "Decrease (DE)"; S63. Formulate fuzzy control rules. According to the actual operation of the power grid and control requirements, formulate fuzzy control rules: “If the deviation is PB and the rate of change of the deviation is PB, then the output is IN”; "If the deviation is Z and the rate of change of the deviation is Z, then the output is ST"; Specific fuzzy control rules should be formulated according to the actual operation of the power grid and the control requirements to ensure the effectiveness and adaptability of the control strategy; S64. Perform fuzzy reasoning, and determine the fuzzy set to which the specific values of the deviation and the deviation change rate collected in real time according to the defined fuzzy set; perform fuzzy reasoning according to the established fuzzy control rules to obtain the fuzzy control quantity.
[0013] Furthermore, in S7, a method for establishing a state estimator for each node or region is as follows: S71. Establishment State Estimator, The status can be expressed as: ; ; is the regional power deviation; is the actual power; is the planned power; is the frequency deviation coefficient; is the regional frequency deviation; S72. Select a state estimation method and use a distributed Kalman filter state estimation method to achieve distributed estimation of the ACE state. The Kalman filter state update equation is: ; in, For the moment The estimated value of the state; For the moment The predicted state value of is the Kalman gain; is the observed value; is the observation matrix; The specific form of depends on the observed variables and the arrangement of the measuring equipment. It is expressed as: ; in, is the dimension of the state vector, is the number of observed variables, Indicates The state variable The degree of influence of the observed variable; S73. Use a distributed estimation method to evaluate the meter. When using distributed Kalman filtering, the nodes in the power grid are divided into multiple sub-areas. The nodes in each sub-area cooperate with each other to jointly estimate the ACE state of the sub-area. The distributed estimation includes: Initialization: Each node initializes its own state estimate based on local data and the covariance matrix ; Information interaction: nodes exchange information about state estimates and covariance matrices, such as Send to neighbor nodes and , and receive neighbor nodes Sent and ; State update, based on the received information, each node updates its own state estimate and covariance matrix, and the state update equation is: ; ; For Node At the moment The estimated value of the state; For Node At the moment The estimated value of the state; Neighbor node At the moment The estimated value of the state; For Node At the moment The covariance matrix of For Node At the moment The covariance matrix of For Node At the moment The Kalman gain, For Node The set of neighbor nodes of Neighbor node The observed value of is the observation matrix of neighbor nodes; is the identity matrix; Repeat information exchange and state update until the state estimate converges or reaches the set number of iterations.
[0014] Furthermore, in said S8, the method for designing a regulator and a controller is: A linear quadratic regulator (LQR) is used to achieve precise control and optimized operation of the power grid. The design goal is to minimize the performance indicators: ; To minimize performance indicators; The value after the state is updated; for The transpose of Obtain the fuzzy control quantity for S67; for The transpose of is the state weight matrix, is the control weight matrix; By solving the equation: , get the optimal control strategy, where is the system matrix, is the input matrix, is a matrix variable related to the system state, which is related to the system matrix , input matrix and other related matrices or parameters ( and ) interact with each other to meet the conditions set by the equation and obtain the optimal control strategy.
[0015] In summary, due to the adoption of the above technical solution, the beneficial technical effects of the invention are: By selecting directed graphs in graph theory to depict the topology of the power grid, creating a set of nodes and edges, constructing an adjacency matrix and calculating the node admittance matrix, the connection relationship and admittance characteristics between the nodes in the power grid are accurately described, which provides a basis for the subsequent load model establishment, time deviation discovery and control strategy formulation, and produces a technical effect of improving the accuracy of power grid topology modeling; By establishing load models at different nodes, analyzing node load characteristics, establishing load model functions and selecting load forecasting models, it is possible to accurately reflect the load variation patterns over time and seasons, as well as the load characteristics of different user types, thus producing a technical effect of improving the accuracy and adaptability of load models, which helps to better predict and manage grid loads. Through the design of distributed time deviation mining algorithm, the constructed power grid structure and established load model are used to calculate the time deviation indicators of different nodes and lines under normal operation and abnormal conditions, which can timely discover abnormal conditions in the power grid, produce the technical effect of improving the stability and reliability of power grid operation, and provide a basis for the formulation of deviation compensation control strategy; Through the formulation of deviation compensation control strategy, corresponding control strategies and constraints are formulated based on time deviation, and the deviation is controlled within the allowable range, ensuring that key indicators such as frequency, voltage and power balance of the power grid remain stable, ensuring the safe and reliable operation of the power grid, and producing technical effects of optimizing power grid control strategy and improving power grid stability; Through the distributed collaborative control strategy, the information interaction and coordination relationship between nodes is determined, and the consistency algorithm is used to make the status of each node gradually consistent, thus realizing collaborative control and producing the technical effect of improving the collaborative control capability and response speed of the power grid; Through real-time optimization and adjustment, the parameters and weights of the control strategy are dynamically adjusted according to the dynamic changes and deviation monitoring of the load model, and the fuzzy logic control algorithm is used to automatically adjust the control strategy according to real-time data, which improves the adaptability of the power grid to load changes and deviations, and produces the technical effect of optimizing the real-time control of the power grid and improving the operation efficiency of the power grid; By establishing a state distributed estimator, using the data obtained in the previous steps, a state estimator is established for each node or area, and a distributed Kalman filter state estimation method is adopted to achieve distributed estimation of the ACE state, which produces a technical effect of improving the accuracy and reliability of power grid state estimation, and provides more accurate information for the design of regulators and controllers; By designing the regulator and controller according to the ACE state estimation results and adopting the linear quadratic regulator (LQR), precise control and optimized operation of the power grid are achieved, producing the technical effect of improving the control accuracy of the power grid and optimizing the operation effect.
[0016] In summary, this application improves the distributed coordination capability of power grid control, reduces dependence on centralized control, reduces the impact of communication delays and information congestion, and improves the real-time and accuracy of control. Fully consider the characteristics of heterogeneous frequency-modulated units, achieve optimized coordination between units, and improve resource utilization efficiency. Enhance the adaptability and robustness of the power grid to uncertainty and interference, and be able to better cope with rapid changes in load and various abnormal situations, ensuring the stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 A method flow chart of a distributed cooperative AGC method for heterogeneous frequency regulation units in a power system; Figure 2 A flow chart of a method for constructing a power grid topology model using a set of nodes and edges; Figure 3 A flow chart of a method for establishing load models at different nodes of a power grid structure; Figure 4 A flow chart of the method for calculating the time deviation index of different nodes and lines under normal operation and abnormal conditions; Figure 5 A flowchart for determining the information interaction and coordination relationship between nodes; Figure 6 A flow chart of a method for dynamically adjusting parameters and weights of a control strategy; Figure 7 Flowchart of the method for establishing a state estimator for each node or region. DETAILED DESCRIPTION
[0018] In order to make the purpose, technical solution and advantages of the invention more clear, the invention is further described in detail below in combination with the embodiments. It should be understood that the specific embodiments described here are only used to explain the invention and are not used to limit the invention.
[0019] like Figure 1 As shown, a distributed collaborative AGC method for heterogeneous frequency regulation units in a power system includes the following steps: S1. Select a directed graph in graph theory to describe the power grid topology, create a set of nodes and edges, the node set represents the substation, the edge set represents the transmission line, and create a set of nodes and edges to construct a power grid topology model; like Figure 2 As shown in FIG. 1 , a method for constructing a power grid topology model using a set of nodes and edges is as follows: S11. Choose a directed graph in graph theory to describe the topology of the power grid. A directed graph consists of nodes and directed edges. Nodes represent substations, and directed edges represent the direction of transmission lines. S12. Define the node set and edge set: A node collection is represented as: ; The edge set is represented as: ; in, and Represents different substations; is the total number of substations; Represents a substation To another substation transmission line; if there is a substation from To another substation The transmission line ,otherwise ; S13. Construct the adjacency matrix, which is expressed as , used to represent the connection relationship between nodes. If there is an edge ,but ,otherwise ; In the adjacency matrix, Line Elements of a column From the substation To the substation Is there a connection; for example, Indicates that there is a connection from substation 1 to substation 2. It means there is no connection from substation 1 to substation 3; S14. Calculate the node admittance matrix: The admittance matrix is expressed as: ,The admittance matrix is used to describe the admittance characteristics between nodes in the power grid; the elements of the admittance matrix are By conductivity and electrosusceptor Composition, that is, ;in, is an imaginary unit; , Line resistance between two substations; , , is the grid frequency, is the line capacitance, is the line inductance.
[0020] S2. Load model establishment: Based on the power grid topology structure determined in S1, load models are established at different nodes of the power grid topology structure; like Figure 3 As shown, the method for establishing load models at different nodes includes: S21. Analyze the node load characteristics. Based on the grid structure determined in S1, determine the user type (such as residential, industrial, commercial, etc.) connected to each node, and determine the load characteristics of different user types, including the time distribution of the load, seasonal variation patterns, and the impact of special events on the load; S22. Establish a load model function and define load characteristic variables for different time periods and seasons, such as characteristic variables for summer and weekdays. The defined characteristic variables are used to reflect the load variation over time and season. The load can be expressed as: ; in, Indicates load, which is used to describe the load situation of a node or area in the power system at a specific time; is the basic load, It is the part that does not change with time and season; For the part that changes with time and season, is the coefficient of the characteristic variable, is the characteristic function, which is used to describe the influence of characteristic variables on load; S23. Select the load forecasting model and use the time series analysis model: ; in, For Node The load forecast value of the node At the moment Load forecasting situation; is the coefficient parameter; used to determine the white noise term The degree of influence on the current load forecast value is different. Corresponding to different coefficients; is the white noise term, indicating random interference factors; It is a constant term in the model and is a parameter related to the base load part that does not change with time and season; The upper limit parameter used for the summation determines the time span of the historical load data involved in the calculation ( Items); is a coefficient parameter used to determine the historical load data The degree of influence on the current load forecast value is different. Corresponding to different coefficients; The upper limit parameter used for the summation determines the number of white noise terms involved in the calculation.
[0021] S3. Design of distributed time deviation mining algorithm, using the grid structure constructed in S1 and the load model established in S2, to calculate the time deviation index of different nodes and lines under normal operation and abnormal conditions, and to monitor the time deviation index in real time; like Figure 4 As shown, S3 includes: S31. Define the time deviation index, which is defined as: ; ; ; in, Indicates the inflow node The set of adjacent nodes of Representation Node The time deviation, Representation Node At the moment The input power, Representation Node At the moment The output power, Representation Node Rated power; Indicates at time Slave Node Inflow Node Power; Indicates outflow node The set of adjacent nodes of Indicates at time Slave Node Flow out to the node Power; S32. Collect grid operation data, use the grid structure constructed in S1 to obtain real-time operation data of each node and line in the grid, including power, voltage, current and other information; at the same time, combine the load model established in S2 to obtain load forecast data of each node; S33. Calculate the time deviation under normal operation, combine the time deviation index defined in S31 and the real-time operation data of each node and line in the power grid and the load forecast data of each node obtained in S32, and calculate the time deviation of each node: ; ; ; in, For normal operation of the node time deviation; Indicates the node under normal operation At the moment The input power, Indicates that the node is in normal operation At the moment The output power, Representation Node Rated power; Indicates the inflow node The set of adjacent nodes of Indicates outflow node The set of adjacent nodes of Indicates at time Slave Node Inflow Node Power; For Node The load forecast value; Indicates at time Slave Node Flow out to the node Power; S34. Detect abnormal conditions and set thresholds to determine whether abnormal conditions occur in the power grid; The power fluctuation threshold can be set as: or When , it is considered that power fluctuation is abnormal, among which, and Node The average value of input power and output power; and Node Input power and output power; is the threshold of power fluctuation; The voltage abnormality threshold can be set as: When the voltage is abnormal, For the moment node The voltage, For Node The average value of the voltage; Threshold indicating voltage anomaly; When an abnormal situation is detected, relevant abnormal information is recorded, such as the time when the abnormality occurred, the node location, the abnormality type, etc. S35. Calculate the time deviation under abnormal conditions. Calculate the input power and output power of each node based on the real-time operation data and load forecast data. Since the power distribution changes due to abnormal conditions, the time deviation under abnormal conditions is different from that under normal conditions. The time deviation of each node under abnormal conditions is: ; In abnormal situations, the node Input power; In abnormal situations, the node Output power; Nodes in abnormal situations Additional input power; Nodes in abnormal situations Additional output power; Indicates at time Slave Node Inflow Node Power; For Node The load forecast value; Indicates at time Slave Node Flow out to the node Power; Indicates the inflow node The set of adjacent nodes of Indicates outflow node The set of adjacent nodes of Representation Node Rated power.
[0022] S4. Deviation compensation control strategy formulation: Based on the time deviation discovered in S3, corresponding control strategies and constraints are formulated; these strategies are designed to control the deviation within the allowable range to ensure that key indicators such as the frequency, voltage and power balance of the power grid remain stable and to ensure the safe and reliable operation of the power grid; S41. Formulate a control strategy, which is expressed as: ; To control the amount; is the proportionality coefficient, is the integration coefficient, is the differential coefficient, For the moment time deviation; S42. Set constraints, which include limiting the size and frequency of the controlled quantity and the frequency of change of the controlled quantity, so as to ensure the safe and reliable operation of the power grid.
[0023] S5. Distributed collaborative control strategy, based on the control strategy determined in S4, determines the information interaction and coordination relationship between nodes according to the power grid topology structure constructed in S1: like Figure 5 As shown, S5 includes: S51. Determine the information interaction and coordination relationship. The information interaction between nodes is expressed as: ; For Node At the moment Status; For Node At the moment Status; For Node Neighbor nodes At the moment Status; For Node The set of neighbor nodes of is the state update function, It is expressed as: ; For Node Its neighbor nodes The weights between them are used to adjust the intensity of information interaction. For Node The amount of control; S52. Control instruction transmission and execution ensures that the control instructions can be accurately and quickly transmitted and executed throughout the power grid. The consistency algorithm is used to make the status of each node gradually consistent, thereby achieving collaborative control. The distributed consensus algorithm is expressed as: ; Representation Node The rate of change of state; For Node Its neighbor nodes The connection weight between them; It is an external input used to guide the state of the node to reach the desired goal; For Node Neighbor nodes At the moment Status; For Node At the moment status.
[0024] S6. Real-time optimization and adjustment: According to the dynamic changes of the load model established in S2 and the deviation monitoring results of the distributed time deviation mining algorithm designed in S3, the parameters and weights of the control strategy are dynamically adjusted: like Figure 6 As shown, S6 includes: S61. Determine the adjustment requirements. According to the dynamic changes of the load model established in S2 and the deviation monitoring results of the distributed time deviation mining algorithm designed in S3, determine whether the parameters and weights of the control strategy need to be adjusted. If the load increases or decreases, or the power grid has an abnormal deviation, adjustment is required. The adjusted control strategy is: ; is the adjusted control quantity; is the adjusted proportional coefficient, is the adjusted differential coefficient; is the adjusted integral coefficient; For the moment time deviation; S62. Adaptive control algorithm, select fuzzy logic control algorithm to automatically adjust the control strategy according to real-time data, the input variable is the deviation and the rate of change of deviation , for the deviation and the rate of change of deviation The fuzzy set is defined as: "Positive big (PB)", "Positive small (PS)", "Zero (Z)", "Negative small (NS)", "Negative big (NB)"; the output control quantity definition fuzzy set is "Increase (IN)", "Stay (ST)", "Decrease (DE)"; S63. Formulate fuzzy control rules. According to the actual operation of the power grid and control requirements, formulate fuzzy control rules: “If the deviation is PB and the rate of change of the deviation is PB, then the output is IN”; "If the deviation is Z and the rate of change of the deviation is Z, then the output is ST" etc. Specific fuzzy control rules should be formulated according to the actual operation of the power grid and the control requirements to ensure the effectiveness and adaptability of the control strategy; S64. Perform fuzzy reasoning, and determine the fuzzy set to which the specific values of the deviation and the deviation change rate collected in real time according to the defined fuzzy set; perform fuzzy reasoning according to the established fuzzy control rules to obtain the fuzzy control quantity.
[0025] S7. The state distributed estimator is established, using the data obtained in S1 to S6 to establish a state distributed estimator for each node or area. State estimator; like Figure 7 As shown, S7 includes: S71. Establishment State Estimator, The status can be expressed as: ; ; is the regional power deviation; is the actual power; is the planned power; is the frequency deviation coefficient; is the regional frequency deviation; S72. Select a state estimation method and use a distributed Kalman filter state estimation method to achieve distributed estimation of the ACE state. The Kalman filter state update equation is: ; in, For the moment The estimated value of the state; For the moment The predicted state value of is the Kalman gain; is the observed value; is the observation matrix; The specific form of depends on the observed variables and the arrangement of the measuring equipment. It is expressed as: ; in, is the dimension of the state vector, is the number of observed variables, Indicates The state variable The degree of influence of the observed variable; S73. Use a distributed estimation method to evaluate the meter. When using distributed Kalman filtering, the nodes in the power grid are divided into multiple sub-areas. The nodes in each sub-area cooperate with each other to jointly estimate the ACE state of the sub-area. The distributed estimation includes: Initialization: Each node initializes its own state estimate based on local data and the covariance matrix ; Information interaction: nodes exchange information about state estimates and covariance matrices, such as Send to neighbor nodes and , and receive neighbor nodes Sent and ; State update, based on the received information, each node updates its own state estimate and covariance matrix, and the state update equation is: ; ; For Node At the moment The Kalman gain, For Node The set of neighbor nodes of Neighbor node The observed value of is the observation matrix of neighbor nodes; For Node At the moment The estimated value of the state; For Node At the moment The estimated value of the state; Neighbor node At the moment The estimated value of the state; Neighbor node At the moment The estimated value of the state; For Node At the moment The covariance matrix of For Node At the moment The covariance matrix of is the identity matrix; Repeat information exchange and state update until the state estimate converges or reaches the set number of iterations.
[0026] S8. Design a regulator and a controller. According to the estimation results of the ACE state estimator established in S7, design a regulator and a controller: A linear quadratic regulator (LQR) is used to achieve precise control and optimized operation of the power grid. The design goal is to minimize the performance indicators: ; To minimize performance indicators; The value after the state is updated; for The transpose of Obtain the fuzzy control quantity for S67; for The transpose of is the state weight matrix, is the control weight matrix; By solving the equation: , get the optimal control strategy, where is the system matrix, is the input matrix, is a matrix variable related to the system state, which is related to the system matrix , input matrix and other related matrices or parameters ( and ) interact with each other to meet the conditions set by the equation and obtain the optimal control strategy.
[0027] The above description is a preferred embodiment of the invention and is not intended to limit the invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the invention should be included in the protection scope of the invention.
Claims
1. A distributed collaborative AGC method for heterogeneous frequency regulation units in a power system. It is characterized in that The following steps are involved: S1. Establishment of the power grid topology model: Select a directed graph in graph theory to describe the power grid topology, create a set of nodes and edges, and use the set of nodes and edges to construct the power grid topology model; S2. Load model establishment: Based on the power grid topology structure determined in S1, load models are established at different nodes of the power grid topology structure; S3. Design of distributed time deviation mining algorithm, using the grid structure constructed in S1 and the load model established in S2, to calculate the time deviation index of different nodes and lines under normal operation and abnormal conditions, and to monitor the time deviation index in real time; S4. Deviation compensation control strategy formulation: Based on the time deviation indicators discovered in S3, formulate corresponding control strategies and constraints; S5. Distributed collaborative control strategy, based on the control strategy determined in S4, determines the information interaction and coordination relationship between nodes according to the power grid topology structure constructed in S1; S6. Real-time optimization and adjustment: dynamically adjust the parameters and weights of the control strategy according to the dynamic changes of the load model established in S2 and the deviation monitoring results of the distributed time deviation mining algorithm designed in S3; S7. Establishment of ACE state distributed estimator, using the data obtained in S1 to S6, to establish an ACE state estimator for each node or area; S8. Design a regulator and a controller based on the estimation results of the ACE state estimator established in S7.
2. According to claim 1, a distributed collaborative AGC method for heterogeneous frequency modulation units in a power system, It is characterized in that In S1, the method of constructing a power grid topology model using a set of nodes and edges includes: S11. Choose a directed graph in graph theory to describe the topology of the power grid. A directed graph consists of nodes and directed edges. Nodes represent substations, and directed edges represent the direction of transmission lines. S12. Define the node set and edge set: Node Collection It is expressed as: ; Edge Set It is expressed as: ; in, and Represents different substations; is the total number of substations; Represents a substation To another substation transmission line; if there is a substation from To another substation The transmission line ,otherwise ; S13. Construct the adjacency matrix, which is expressed as , used to represent the connection relationship between nodes. If there is an edge ,but ,otherwise ; S14. Calculate the node admittance matrix: The admittance matrix is expressed as: ,The admittance matrix is used to describe the admittance characteristics between nodes in the power grid; the elements of the admittance matrix are By conductivity and electrosusceptor Composition, that is, ; in, is an imaginary unit; , Line resistance between two substations; , , is the grid frequency, is the line capacitance, is the line inductance.
3. According to claim 1, a distributed collaborative AGC method for heterogeneous frequency modulation units in a power system, It is characterized in that In S2, the method of establishing load models on different nodes includes: S21. Analyze the node load characteristics, based on the grid structure determined in S1, determine the user type connected to each node and the load characteristics of different user types; S22. Establish a load model function, define load characteristic variables for different time periods and seasons, and use the defined characteristic variables to reflect the load variation over time and season. The load is expressed as: ; in, Indicates load; is the basic load, For the part that does not change with time and season; For the part that changes with time and season, is the coefficient of the characteristic variable, is the characteristic function, which is used to describe the influence of characteristic variables on load; S23. Select the load forecasting model and use the time series analysis model: ; in, For Node The load forecast value of the node At the moment Load forecasting situation; is the coefficient parameter used to determine the white noise term The impact on the current load forecast value is different. Corresponding to different coefficients; is a white noise term, indicating random interference factors; It is a constant term in the model and is a parameter related to the base load part that does not change with time and season; The upper limit parameter used for summation determines the time span of the historical load data involved in the calculation; is a coefficient parameter used to determine the historical load data The impact on the current load forecast value is different. Corresponding to different coefficients; The upper limit parameter used for the summation determines the number of white noise terms involved in the calculation.
4. According to claim 1, a distributed collaborative AGC method for heterogeneous frequency modulation units in a power system, It is characterized in that In S3, the steps of calculating the time deviation index of different nodes and lines under normal operation and abnormal conditions are as follows: S31. Define the time deviation index, which is defined as: ; ; ; in, Indicates the inflow node The set of adjacent nodes of Representation Node The time deviation, Representation Node At the moment The input power, Representation Node At the moment The output power, Representation Node Rated power; Indicates at time Slave Node Inflow Node Power; Indicates outflow node The set of adjacent nodes of Indicates at time Slave Node Flow out to the node Power; S32. Collect grid operation data, use the grid structure constructed in S1 to obtain real-time operation data of each node and line in the grid, and at the same time, combine the load model established in S2 to obtain load forecast data of each node; S33. Calculate the time deviation under normal operation, combine the time deviation index defined in S31 and the real-time operation data of each node and line in the power grid and the load forecast data of each node obtained in S32, and calculate the time deviation of each node: ; ; ; in, For normal operation of the node time deviation; Indicates that the node is in normal operation At the moment The input power, Indicates that the node is in normal operation At the moment The output power, Representation Node Rated power; Indicates the inflow node The set of adjacent nodes of Indicates outflow node The set of adjacent nodes of Indicates at time Slave Node Inflow Node Power; For Node The load forecast value; Indicates at time Slave Node Flow out to the node Power; S34. Detect abnormal conditions and set thresholds to determine whether abnormal conditions occur in the power grid; The power fluctuation threshold is set as: or When , it is considered that power fluctuation is abnormal, among which, and Node The average value of input power and output power; and Node Input power and output power; is the threshold of power fluctuation; The voltage abnormality threshold is set as: When the voltage is abnormal, For the moment node The voltage, For Node The average value of the voltage; Threshold indicating voltage anomaly; When an abnormal situation is detected, relevant abnormal information is recorded; S35. Calculate the time deviation under abnormal conditions, calculate the input power and output power of each node according to the real-time operation data and load forecast data, and the time deviation of each node under abnormal conditions: ; In abnormal situations, the node Input power; In abnormal situations, the node Output power; Nodes in abnormal situations Additional input power; Nodes in abnormal situations Additional output power; Indicates at time Slave Node Inflow Node Power; For Node The load forecast value; Indicates at time Slave Node Flow out to the node Power; Indicates the inflow node The set of adjacent nodes of Indicates outflow node The set of adjacent nodes of Representation Node Rated power.
5. According to claim 1, a distributed collaborative AGC method for heterogeneous frequency modulation units in a power system, It is characterized in that In S4, The control strategy is expressed as: ; To control the amount; is the proportionality coefficient, is the integration coefficient, is the differential coefficient, For the moment time deviation; The constraints include: Determine the upper and lower limits of the control quantity according to the rated capacity of the power grid equipment and the stability requirements; The frequency of change of the controlled quantity shall not exceed the threshold; The rate of change of the controlled quantity shall not exceed the limit value; The constraints need to be determined based on the specific characteristics of the power grid, equipment parameters and operating requirements.
6. A distributed collaborative AGC method for heterogeneous frequency modulation units in a power system according to claim 1, It is characterized in that In S5, the information interaction and coordination relationship method between the nodes is determined: S51. Determine the information interaction and coordination relationship. The information interaction between nodes is expressed as: ; For Node At the moment Status; For Node At the moment Status; For Node Neighbor nodes At the moment Status; For Node The set of neighbor nodes; is the state update function, It is expressed as: ; For Node Its neighbor nodes The weights between them are used to adjust the intensity of information interaction. For Node The amount of control; S52. Control instruction transmission and execution ensures that the control instructions can be accurately and quickly transmitted and executed throughout the power grid. The consistency algorithm is used to gradually make the status of each node consistent, thereby achieving collaborative control. The distributed consensus algorithm is expressed as: ; Representation Node The rate of change of state; For Node Its neighbor nodes The connection weight between them; It is an external input used to guide the node's state to reach the desired goal; For Node Neighbor nodes At the moment Status; For Node At the moment status.
7. A distributed collaborative AGC method for heterogeneous frequency modulation units in a power system according to claim 1, It is characterized in that In S6, the method for dynamically adjusting the parameters and weights of the control strategy is: S61. Determine the adjustment requirements. According to the dynamic changes of the load model established in S2 and the deviation monitoring results of the distributed time deviation mining algorithm designed in S3, determine whether the parameters and weights of the control strategy need to be adjusted. The adjusted control strategy is: ; is the adjusted control quantity; is the adjusted proportional coefficient, is the adjusted differential coefficient; is the adjusted integral coefficient; For the moment time deviation; S62. Adaptive control algorithm, select fuzzy logic control algorithm to automatically adjust the control strategy according to real-time data, the input variable is the deviation and the rate of change of deviation , for the deviation and the rate of change of deviation The fuzzy set is defined as: "Positive large PB", "positive small PS", "zero Z", "negative small NS", "negative large NB"; the output control quantity definition fuzzy set is "increase IN", "maintain ST", "decrease DE"; S63. Formulate fuzzy control rules. According to the actual operation of the power grid and control requirements, formulate fuzzy control rules: "If the deviation is PB and the rate of change of the deviation is PB, the output is IN"; "If the deviation is Z and the rate of change of the deviation is Z, then the output is ST"; S64. Perform fuzzy reasoning, and determine the fuzzy set to which the specific values of the deviation and the deviation change rate collected in real time according to the defined fuzzy set; perform fuzzy reasoning according to the established fuzzy control rules to obtain the fuzzy control quantity.
8. A distributed collaborative AGC method for heterogeneous frequency modulation units in a power system according to claim 1, It is characterized in that In S7, a Methods of the state estimator: S71. Establishment State Estimator, The status is represented by: ; ; is the regional power deviation; is the actual power; is the planned power; is the frequency deviation coefficient; is the regional frequency deviation; S72. Select a state estimation method and use a distributed Kalman filter state estimation method to achieve distributed estimation of the ACE state. The Kalman filter state update equation is: ; in, For the moment The estimated value of the state; For the moment The predicted state value of is the Kalman gain; is the observed value; is the observation matrix; The specific form of depends on the observed variables and the arrangement of the measuring equipment. It is expressed as: ; in, is the dimension of the state vector, is the number of observed variables, Indicates The state variable The degree of influence of the observed variable; S73. Evaluate using a distributed estimation method. When using distributed Kalman filtering, the nodes in the power grid are divided into multiple sub-areas. The nodes in each sub-area cooperate with each other to jointly estimate the ACE state of the sub-area. The distributed estimation includes: Initialization: Each node initializes its own state estimate based on local data and the covariance matrix ; Information interaction: nodes exchange information about state estimates and covariance matrices, such as Send to neighbor nodes and , and receive neighbor nodes Sent and ; State update, based on the received information, each node updates its own state estimate and covariance matrix, and the state update equation is: ; ; For Node At the moment The estimated value of the state; For Node At the moment The estimated value of the state; Neighbor node At the moment The estimated value of the state; For Node At the moment The covariance matrix of For Node At the moment The covariance matrix of For Node At the moment The Kalman gain, For Node The set of neighbor nodes of Neighbor node The observed value of is the observation matrix of neighbor nodes; is the identity matrix; Repeat information exchange and state update until the state estimate converges or reaches the set number of iterations.
9. A distributed collaborative AGC method for heterogeneous frequency modulation units in a power system according to claim 1, It is characterized in that In S8, the method for designing the regulator and the controller is: A linear quadratic regulator (LQR) is used to achieve precise control and optimized operation of the power grid. The design goal is to minimize the performance indicators: ; To minimize performance indicators; The value after the state is updated; for The transpose of To obtain the fuzzy control quantity; for The transpose of is the state weight matrix, is the control weight matrix; By solving the equation: , get the optimal control strategy, where is the system matrix, is the input matrix, is a matrix variable related to the system state.
Citation Information
Patent Citations
Method for correcting change of interconnection AGC frequency modulation unit in electricity market region
CN109861249A
Master station side frequency control method for AGC and primary frequency modulation coordination
CN110429669A
Optimal control method and system for energy storage system participating in power grid frequency modulation, and storage medium
CN112615399A
AGC control method based on multi-region interconnected power grid
CN113241778A
Multi-region control method and device for new energy and conventional energy
CN115085283A
Cited By
Intelligent power grid planning coordination evaluation method
CN119721853A
Distributed power supply secondary frequency regulation control method for communication misleading data
CN121395316A
Distributed monitoring and automatic response control system of network security equipment
CN121966957A
