Dynamic reactive power allocation method for high-proportion new energy power grid based on avc system

Abstract

The application discloses a high-proportion new energy power grid dynamic reactive power allocation method based on an AVC system and belongs to the technical field of power systems, which comprises the following steps: S1, constructing a dynamic time sequence scene set model; S2, determining an effective reactive power reserve constraint; S3, establishing a Wiener process degradation model; S4, building a two-stage robust distribution model; S5, processing the constraint; S6, model solving; S7, scheme execution; and S8, effect evaluation. The high-proportion new energy power grid dynamic reactive power allocation method based on the AVC system is adopted, the voltage deviation rate is greatly reduced, the reactive power dynamic allocation cost after optimization is reduced compared with a traditional robust model, the stable operation of the power system under the condition of large-scale new energy injection is ensured, and the economy of the reactive power dynamic allocation under the condition of large-scale new energy injection is optimized.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to a dynamic reactive power allocation method for high-proportion renewable energy power grids based on AVC systems. Background Technology

[0002] my country's energy structure is rapidly transitioning towards clean energy, with new energy sources gradually becoming a core component of the energy supply system. However, the large-scale integration of new energy sources into the grid presents inherent challenges to grid stability due to their high volatility and uncertainty, particularly in reactive power regulation and voltage control. Existing dynamic reactive power allocation schemes for the power grid have several shortcomings. On the one hand, traditional schemes struggle to balance effective reactive power reserves and voltage safety ranges when dealing with the dual uncertainties of new energy sources and loads. On the other hand, existing schemes do not adequately address the health status management of reactive power equipment, failing to link equipment health degradation with economic costs, leading to excessive equipment wear and increased maintenance costs.

[0003] From a technical perspective, traditional reactive power optimization models also have shortcomings. First, traditional models often employ static scenario assumptions, failing to accurately depict the dynamic temporal changes of renewable energy sources and loads, resulting in poor adaptability and robustness of the optimization results. Second, handling nonlinear power flow constraints in the models is difficult, leading to low solution efficiency. Third, the lack of coordination with automatic voltage control (AVC) systems results in an imperfect execution and feedback mechanism for optimization schemes, hindering the formation of a closed-loop management system encompassing "modeling-optimization-execution-evaluation," thus limiting the effective implementation of reactive power regulation. These technical pain points collectively constrain the safe and economical operation of power grids with a high proportion of renewable energy, necessitating innovative technical solutions to overcome these challenges. Summary of the Invention

[0004] The purpose of this invention is to provide a dynamic reactive power allocation method for high-proportion renewable energy power grids based on AVC systems, which can effectively cope with extreme scenarios and renewable energy fluctuations, and ensure voltage safety and controllability; reduce the cost of reactive power equipment throughout its entire life cycle and make precise and efficient use of regulation resources; improve the solution speed of complex models and be seamlessly compatible with existing AVC systems; and achieve accurate fault identification, hierarchical processing and robustness assurance.

[0005] To achieve the above objectives, this invention provides a dynamic reactive power allocation method for high-proportion renewable energy power grids based on an AVC system, comprising the following steps:

[0006] S1. Construct a dynamic time-series scenario set model that includes fluctuations in renewable energy output and uncertainties in load demand;

[0007] S2. Define the effective reactive power reserve of each node in a high-proportion renewable energy grid, and determine the constraints by combining the node voltage safety range and the reactive power output limit under all operating scenarios.

[0008] S3. Construct a data-physical fusion Wiener process degradation model, and update the model parameters online using Bayesian methods by combining the physical failure mechanism of the equipment with historical operating data.

[0009] S4. Establish a two-stage distributed bar optimization model. The first stage optimizes the dynamic allocation scheme of reactive power. The second stage considers the source load reduction, voltage deviation and health degradation of reactive power reserve equipment under the worst scenario. The health degradation is directly quantified into economic cost and embedded into the optimization target.

[0010] S5. By using the second-order cone relaxation method to handle nonlinear power flow constraints, the above model is transformed into a mixed-integer second-order cone programming problem.

[0011] S6. The column and constraint generation algorithm CCG is used in combination with the Gurobi solver to solve the model;

[0012] S7. Utilize the automatic voltage control system to execute the optimized reactive power distribution scheme in real time.

[0013] S8. Evaluate the optimization effect through voltage deviation rate and voltage fluctuation rate indicators.

[0014] Preferably, step S1 specifically includes:

[0015] Each individual scene is represented as:

[0016] ;

[0017] ;

[0018] in, Indicates inclusion A complete set of scenes for each scenario. Indicates inclusion The first energy station One scenario, Indicates the first The energy station at the first Time in each scene A complete set of scenes; Represented as:

[0019] ;

[0020] Each element includes a node number. Active power output and reactive power output As shown in the following formula:

[0021] ;

[0022] The complete set of dynamic time series uncertainty scenarios for the power grid is obtained, as shown in the following equation:

[0023] ;

[0024] The modeling of the dynamic load scenario is as follows:

[0025] ;

[0026] ;

[0027] in, Indicates the first In the first scenario One load, and Indicates the load point The uncertainty variable is obtained. The scenario is represented as follows:

[0028] ;

[0029] The complete set of scenarios for uncertain loads is mathematically expressed as follows:

[0030] ;

[0031] in, Indicates inclusion A complete load scenario set for each fluctuating load node, integrating new energy scenario sets. With load scenario set This yields a set of system-wide uncertainty scenarios for the entire power grid:

[0032] ;

[0033] in, This represents a complete set of scenarios that includes grid loads and renewable energy stations.

[0034] Preferably, step S2 specifically includes:

[0035] S21. Define the effective reactive power reserve as the smaller of the following two values: (1) the maximum / minimum reactive power capacity that can maintain the node voltage within the safe operating range within the effective range. , Secondly, under all operating scenarios, the maximum positive / negative reactive power output of the node. , ;

[0036] S22. Determine the core constraints, specifically including:

[0037] Current constraints:

[0038] ;

[0039] in, Indicates in the scene Next Time Node The active power adjustment amount, Indicates in the scene Next node In time The active power load demand, Indicates in the scene Next node In time voltage amplitude, Represents a node , The electrical conductivity between them Represents a node , The susceptance between them Indicates in the scene Next node In time voltage phase angle, Indicates in the scene Next Time Node reactive power adjustment amount Indicates in the scene Next node In time reactive load demand, Indicates in the scene Next node In time voltage amplitude, Indicates in the scene Next node In time voltage phase angle, Represents the set of all nodes;

[0040] Effective reactive power reserve boundary constraints:

[0041] ;

[0042] in, and They represent the scenes respectively. Next node The maximum effective positive and negative reactive power that affects voltage regulation;

[0043] Voltage upper and lower limit constraints:

[0044] ;

[0045] in, for Nodes in the scene Lower voltage limit for Nodes in the scene Voltage upper limit;

[0046] Normalization constraints:

[0047] ;

[0048] in, Represents the infinite norm, This represents the initial probability of a discrete scenario. Represents the probability of a scenario. This represents the range of probability variation allowed under the ∞-norm constraint, at a 95% confidence level. The calculation expression is:

[0049] ;

[0050] in, Indicates the confidence level. Indicates the number of scenes.

[0051] Preferably, step S3 specifically includes:

[0052] S31. Establish a Wiener process degradation model for data-physical fusion, and analyze the performance degradation process of reactive power equipment. Modeled as a Wiener process with drift, the expression is:

[0053] ;

[0054] in, For a moment The amount of equipment performance degradation, This is the initial degradation amount. The drift coefficient, The diffusion coefficient is... For standard Brownian motion, The cumulative action intensity function is expressed as follows:

[0055] ;

[0056] in, and Is the equipment in constant The reactive power output and current, and For the rated value, the index and These are the physical characteristic parameters of the equipment;

[0057] S32. Key parameters of the model based on the Bayesian update method Online learning includes:

[0058] S321. Set parameters based on group data of device models. prior distribution ;

[0059] S322, within the time window Internally, a series of degradation observation values ​​of the equipment are obtained through an online monitoring system. Based on the properties of the Wiener process, a likelihood function is constructed. ;

[0060] S323. According to Bayes' theorem, calculate the posterior distribution of the parameters:

[0061] ;

[0062] S324, From the posterior distribution The current optimal estimate of the parameters is obtained and used for optimization decision-making in the next cycle;

[0063] S33. Introducing surge early warning characteristic: kurtosis index The calculation formula is:

[0064] ;

[0065] in, Indicates kurtosis, Represent a random variable, express The expected value or average value;

[0066] Warning judgment criteria: Gaussian signal under normal operating conditions Resonance or short-circuit faults are characterized by spike signals. A capacitor open-circuit fault results in a flat signal. .

[0067] Preferably, step S4 specifically includes:

[0068] S41. The first stage of optimizing the dynamic reactive power allocation scheme has the following objective function:

[0069] ;

[0070] in, This represents the complete set of time steps in the voltage recovery process, while Indicates in the scene Next Time Node reactive power adjustment amount This represents the set of nodes in the first phase.

[0071] S42. The second stage considers the worst-case scenario of load reduction, voltage deviation, and the health degradation of reactive power reserve equipment. The goal of this sub-problem is to identify the worst-case probability distribution that maximizes the minimum load reduction cost under N-1 emergency conditions, expressed as:

[0072] ;

[0073] For each scenario , and They represent the scenes respectively. Mid-time node The active and reactive load demands, among which, Quantified nodes In time The load reduction ratio, For nodes Load priority weights, This represents the set of nodes in the second phase.

[0074] Calculate voltage deviation and reactive power reserve equipment health degradation:

[0075] ;

[0076] ;

[0077] in, This indicates that due to voltage changes, the node and nodes In time The reactive power loss generated between them; For equipment health costs, This is a collection of all reactive devices. For equipment The health cost coefficient, For the scene ,time ,equipment The predicted health index is calculated online by the data-physical fusion Wiener process model established in step S3.

[0078] ;

[0079] in, It is equipment In the scene The amount of degradation, It is the critical degradation level that leads to equipment failure.

[0080] Preferably, step S5 specifically includes:

[0081] By replacing the nonlinear power flow constraints with a strong second-order cone relaxation method, the mixed-integer nonlinear optimization problem is transformed into a linear optimization problem, avoiding local optima while satisfying the following constraints:

[0082] ;

[0083] in, Indicates in the scene Next node In time The active output, Indicates in the scene Lower reactive power absorption node In time The active output, For nodes and nodes The product of the real parts of the voltages, For nodes and nodes The product of the imaginary voltages between them Represents a node With nodes The square of the real part of the voltage. Represents a node With nodes The square of the imaginary part of the voltage. For nodes The square of the voltage modulus, Represents a node The square of the voltage modulus, Indicates in the scene Lower reactive power absorption node In time reactive power, Indicates in the scene Downstream power node In time reactive power, Indicates in the scene Next reactive power compensation node In time reactive power, Indicates time Located at node The status of the parallel reactive power compensation equipment in operation. Indicates whether or not to invest.

[0084] Preferably, step S6 specifically includes:

[0085] The objective function is decomposed to obtain the main problem and sub-problems. The min problems in each scenario of the sub-problems are independent of each other. The sub-problems are decomposed into multiple individual problems and solved independently in sequence. By solving the sub-problems, the probability distribution with the highest robust cost is obtained. This distribution is substituted into the main problem for iterative calculation, thereby realizing the solution of the two-stage sub-Bruker Reactive Power Optimization (DRO) model. After solving, the dynamic reactive power injection of each bus node of the system at different times is obtained.

[0086] Preferably, step S7 specifically includes:

[0087] The power grid condition monitoring system acquires real-time operating signals from the power grid. And perform analysis and diagnosis to detect the occurrence of faults and identify the fault type;

[0088] Based on the identified power grid conditions, the reactive power allocation optimization algorithm is adaptively adjusted through a closed-loop feedback mechanism.

[0089] The optimized allocation scheme is then transmitted to the Central Control Center (CCC) and the Renewable Energy Cluster Controller (RES-CC) for coordinated execution.

[0090] Reactive power regulation commands include nodes and its reactive power regulation As shown below:

[0091] ;

[0092] in, Indicates in the scene Next time interval internal nodes Dynamic reactive power distribution;

[0093] The set of reactive power regulation commands for all grid nodes is as follows:

[0094] ;

[0095] in, This represents the set of dynamic reactive power distributions across all grid nodes within different time intervals.

[0096] Preferably, step S8 specifically includes:

[0097] After obtaining the solution results, the voltage deviation rate (VDR) and voltage fluctuation rate (VFR) are calculated to evaluate the optimization effect. The calculation formula for the evaluation index is as follows:

[0098] ;

[0099] ;

[0100] in, For nodes In the scene Actual voltage at each point in time For nodes The reference voltage, The total number of nodes. The total number of moments. For nodes The maximum value of the actual voltage amplitude fluctuating within a certain time domain.

[0101] Therefore, this invention adopts the above-mentioned dynamic reactive power allocation method for high-proportion renewable energy power grids based on AVC systems. In terms of grid voltage stability, it effectively copes with extreme scenarios and renewable energy fluctuations, ensuring voltage safety and controllability; in terms of economy and equipment lifespan, it achieves reduced life-cycle costs of reactive power equipment and precise and efficient utilization of regulation resources; in terms of model solving and engineering application efficiency, it significantly improves the solving speed of complex models and is seamlessly compatible with existing AVC systems; in terms of fault response, it achieves accurate fault identification, hierarchical processing, and robustness assurance, comprehensively achieving the "stable, economical, efficient, and intelligent" operation goals of high-proportion renewable energy power grids.

[0102] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0103] Figure 1 This is a flowchart of a method according to an embodiment of the present invention;

[0104] Figure 2 This is a schematic diagram illustrating the effective reactive power reserve constraint definition in an embodiment of the present invention;

[0105] Figure 3 This is a flowchart illustrating the coordination process between the dynamic reactive power allocation optimization model and the AVC system according to an embodiment of the present invention.

[0106] Figure 4 This is a schematic diagram of the operation of the AVC system according to an embodiment of the present invention;

[0107] Figure 5 This is a diagram of the optimized reactive power dynamic allocation scheme for two representative nodes in an embodiment of the present invention, wherein (a) represents a high energy ratio and (b) represents a low energy ratio;

[0108] Figure 6 This is a comparison diagram of the reactive power dynamic allocation schemes for two representative nodes with high and low renewable energy ratios in an embodiment of the present invention, where (a) represents the high renewable energy ratio and (b) represents the low renewable energy ratio. Detailed Implementation

[0109] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0110] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0111] Example

[0112] The hardware environment for this embodiment is: an experimental platform built on a personal computer configured with an Intel Core i7-12500H CPU@2.50GHz, 32.0GB of memory, and an NVIDIA GTX 4050 GPU;

[0113] Software environment: The program is written in Python 3.10.9, and the optimization parts use the Gurobi 10.0.3 framework;

[0114] Example data: The example is built using a real power grid in a district of a prefecture-level city in an eastern province. The power grid has 38 nodes and 6 reactive power compensation nodes, and the penetration rate of new energy is 21.69%. Due to data confidentiality, the specific network topology is not disclosed.

[0115] Implementation as Figure 1 The steps shown include:

[0116] S1. Construct a dynamic time-series scenario set model that includes fluctuations in renewable energy output and uncertainties in load demand.

[0117] Each individual scene is represented as:

[0118] ;

[0119] ;

[0120] in, Indicates inclusion A complete set of scenes for each scenario. Indicates inclusion The first energy station One scenario, Indicates the first The energy station at the first Time in each scene A complete set of scenes; Represented as:

[0121] ;

[0122] Each element includes a node number. Active power output and reactive power output As shown in the following formula:

[0123] ;

[0124] By and Substituting into the above equation, we obtain the complete set of dynamic time series uncertainty scenarios for the power grid, as shown in the following equation:

[0125] ;

[0126] The modeling of the dynamic load scenario is as follows:

[0127] ;

[0128] ;

[0129] in, Indicates the first In the first scenario load ( ), and Indicates the load point The uncertainty variable is obtained. The scenario is represented as follows:

[0130] ;

[0131] The complete set of scenarios for uncertain loads is mathematically expressed as follows:

[0132] ;

[0133] in, Indicates inclusion A complete load scenario set for each fluctuating load node, integrating new energy scenario sets. With load scenario set This yields a set of system-wide uncertainty scenarios for the entire power grid:

[0134] ;

[0135] in, This represents a complete set of scenarios that includes grid loads and renewable energy stations.

[0136] S2. Define the effective reactive power reserve of each node in a high-proportion renewable energy grid, and determine the constraint terms by combining the node voltage safety range and the reactive power output limit under all possible operating scenarios.

[0137] S21. Define the effective reactive power reserve as the smaller of the following two values: (1) the maximum / minimum reactive power capacity that can maintain the node voltage within the safe operating range within the effective range. , Secondly, the maximum positive / negative reactive power that the node can output under all possible operating scenarios. , ;

[0138] S22. Determine the core constraints, specifically including:

[0139] Current constraints:

[0140] ;

[0141] in, Indicates in the scene Next Time Node The active power adjustment amount, Indicates in the scene Next node In time The active power load demand, Indicates in the scene Next node In time voltage amplitude, Represents a node , The electrical conductivity between them Represents a node , The susceptance between them Indicates in the scene Next node In time voltage phase angle, Indicates in the scene Next Time Node reactive power adjustment amount Indicates in the scene Next node In time reactive load demand, Indicates in the scene Next node In time voltage amplitude, Indicates in the scene Next node In time voltage phase angle, Represents the set of all nodes;

[0142] Effective reactive power reserve boundary constraints:

[0143] ;

[0144] in, and They represent the scenes respectively. Next node The maximum effective positive and negative reactive power that affects voltage regulation;

[0145] Voltage upper and lower limit constraints:

[0146] ;

[0147] in, for Nodes in the scene Lower voltage limit for Nodes in the scene Voltage upper limit;

[0148] Normalization constraints:

[0149] ;

[0150] in, Represents the infinite norm, This represents the initial probability of a discrete scenario. Represents the probability of a scenario. This represents the range of probability variation allowed under the ∞-norm constraint, at a 95% confidence level. The calculation expression is:

[0151] ;

[0152] in, Indicates the confidence level. Indicates the number of scenes.

[0153] S3. A Wiener process degradation model based on data-physical fusion is proposed. The model parameters are updated online using Bayesian methods by combining the physical failure mechanism of the equipment with historical operating data.

[0154] S31. Establish a Wiener process degradation model for data-physical fusion, and analyze the performance degradation process of reactive power equipment. It can be modeled as a Wiener process with drift, and the expression is:

[0155] ;

[0156] in, For a moment The amount of equipment performance degradation, This is the initial degradation amount, usually set to 0. Drift coefficient The average rate of performance degradation. Performance deteriorates linearly over time (e.g., increased vibration). Rare, indicating performance recovery after repair. The diffusion coefficient ( ), This is a standard Brownian motion (Wiener process). The cumulative action intensity function is expressed as follows:

[0157] ;

[0158] in, and Is the equipment in constant The reactive power output and current, and For the rated value, the index and These are the physical characteristic parameters of the equipment, determined through accelerated aging tests or manufacturer data, and are used to quantify the nonlinear impact of different intensity of action on the equipment's lifespan.

[0159] S32. To adapt the Wiener process model to individual differences and changes in the operating environment, key parameters of the model are updated based on the Bayesian update method. Online learning includes:

[0160] S321, Prior Distribution Settings: Based on group data of device models, set parameters. prior distribution ;

[0161] S322. Likelihood Function Construction: Within the Time Window Internally, a series of degradation observation values ​​of the equipment are obtained through an online monitoring system. Based on the properties of the Wiener process, a likelihood function is constructed. ;

[0162] S323, Posterior Distribution Update: According to Bayes' theorem, calculate the posterior distribution of the parameters:

[0163] ;

[0164] S324. Parameter point estimation: from the posterior distribution The current optimal estimate of the parameters (such as the maximum a posteriori estimate) is obtained and used for optimization decision-making in the next cycle;

[0165] S33. Introducing surge early warning characteristic: kurtosis index The calculation formula is:

[0166] ;

[0167] in, Indicates kurtosis, Represent a random variable, express The expected value or average value;

[0168] Warning judgment criteria: Gaussian signal under normal operating conditions Resonance or short-circuit faults are characterized by spike signals. A capacitor open-circuit fault results in a flat signal. .

[0169] S4. Establish a two-stage distributed bar optimization model. The first stage optimizes the dynamic allocation scheme of reactive power. The second stage considers the source load reduction, voltage deviation and health degradation of reactive power reserve equipment under the worst scenario. The health degradation is directly quantified into economic cost and embedded into the optimization objective.

[0170] S41. The first stage optimizes the dynamic reactive power allocation scheme. The objective function of the main problem is to calculate the reactive power regulation at each reactive power action point of the power system with large-scale renewable energy injection, i.e., the dynamic reactive power allocation scheme, expressed as:

[0171] ;

[0172] in, This represents the complete set of time steps in the voltage recovery process, while Indicates in the scene Next Time Node reactive power adjustment amount This represents the set of nodes in the first phase.

[0173] S42. The second stage considers the worst-case scenario of load reduction, voltage deviation, and the health degradation of reactive power reserve equipment. The goal of this sub-problem is to identify the worst-case probability distribution that maximizes the minimum load reduction cost under N-1 emergency conditions, expressed as:

[0174] ;

[0175] For each scenario , and They represent the scenes respectively. Mid-time node The active and reactive load demands, of which Quantified nodes In time The load reduction ratio, For nodes Load priority weights, This represents the set of nodes in the second phase.

[0176] Calculate voltage deviation and reactive power reserve equipment health degradation:

[0177] ;

[0178] ;

[0179] in, This indicates that due to voltage changes, the node and nodes In time The reactive power loss generated between them; For equipment health costs, This is a collection of all reactive devices. For equipment The health cost coefficient, For the scene ,time ,equipment The predicted health index is calculated online by the data-physical fusion Wiener process model established in step S3.

[0180] ;

[0181] in, It is equipment In the scene The amount of degradation, It is the critical degradation level that leads to equipment failure.

[0182] S5. The nonlinear power flow constraints are handled by the second-order cone relaxation method, and the model is transformed into a mixed-integer second-order cone programming problem.

[0183] In this approach, the strong second-order cone relaxation method is used to replace the nonlinear power flow constraints, transforming the mixed-integer nonlinear optimization problem into a linear optimization problem to avoid local optima, while satisfying the following constraints:

[0184] ;

[0185] in, Indicates in the scene Next node In time The active output, Indicates in the scene Lower reactive power absorption node In time The active output, For nodes and nodes The product of the real parts of the voltages, For nodes and nodes The product of the imaginary voltages between them Represents a node With nodes The square of the real part of the voltage. Represents a node With nodes The square of the imaginary part of the voltage. For nodes The square of the voltage modulus, Represents a node The square of the voltage modulus, Indicates in the scene Lower reactive power absorption node In time reactive power, Indicates in the scene Downstream power node In time reactive power, Indicates in the scene Next reactive power compensation node In time reactive power, Indicates time Located at node The status of the parallel reactive power compensation equipment in operation. Indicates whether or not to invest.

[0186] S6. The column and constraint generation algorithm CCG is combined with the Gurobi solver to solve the model efficiently;

[0187] Specifically, the objective function is decomposed to obtain the main problem and sub-problems. The min problems in each scenario of the sub-problems are independent of each other. The sub-problems are decomposed into multiple separate problems and solved independently in sequence. By solving the sub-problems, the probability distribution with the highest robust cost is obtained. This distribution is substituted into the main problem for iterative calculation, thereby realizing the solution of the two-stage sub-robust optimization DRO model. After solving, the dynamic reactive power injection amount of each bus node of the system at different times is obtained, that is, the dynamic reactive power allocation scheme.

[0188] Algorithm combined with AVC workflow such as Figure 4 As shown.

[0189] S7. Utilize the automatic voltage control system to execute the optimized reactive power distribution scheme in real time.

[0190] The power grid condition monitoring system acquires real-time operating signals from the power grid. And perform analysis and diagnosis to detect the occurrence of faults and identify the fault type;

[0191] Based on the identified power grid conditions, the reactive power allocation optimization algorithm is adaptively adjusted through a closed-loop feedback mechanism.

[0192] The optimized allocation scheme is then transmitted to the Central Control Center (CCC) and the Renewable Energy Cluster Controller (RES-CC) for coordinated execution; this integrated framework ensures dynamic voltage stability under various fault scenarios while minimizing operating costs.

[0193] Reactive power regulation commands include nodes and its reactive power regulation As shown below:

[0194] ;

[0195] in, Indicates in the scene Next time interval internal nodes Dynamic reactive power distribution;

[0196] The set of reactive power regulation commands for all grid nodes is as follows:

[0197] ;

[0198] in, This represents the set of dynamic reactive power distributions across all grid nodes within different time intervals.

[0199] S8. Evaluate the optimization effect using voltage deviation rate and voltage fluctuation rate indicators, and compare it with traditional methods.

[0200] After obtaining the solution results, the voltage deviation rate (VDR) and voltage fluctuation rate (VFR) are calculated to evaluate the optimization effect. The calculation formula for the evaluation index is as follows:

[0201] ;

[0202] ;

[0203] in, For nodes In the scene Actual voltage at each point in time For nodes The reference voltage, The total number of nodes. The total number of moments. For nodes The maximum value of the actual voltage amplitude fluctuating within a certain time domain.

[0204] First, based on S1, a dynamic time-series scenario set model is constructed, incorporating fluctuations in new energy output and uncertainties in load demand; combined with S2, and referring to... Figure 2 The diagram illustrates the effective reactive power reserve constraint, clearly defining the effective reactive power reserve of each node in a high-proportion renewable energy power grid. It also defines power flow constraints, effective reactive power reserve boundary constraints, voltage upper and lower limit constraints, and paradigm constraints, providing a foundation for subsequent optimization.

[0205] Following S3, a data-physical fusion Wiener process degradation model is adopted to quantify the health degradation process of reactive power-operated equipment. Combined with the two-stage sub-Bruker optimization model of S4, sub-problems, main problems, constraints, and scenario data are substituted into the model. By finding the reactive power regulation scheme with the highest robustness cost, the reactive power output allocation strategy for the six reactive power-operated nodes at different times in the entire scenario is determined, resulting in the optimized dynamic reactive power allocation scheme, as shown below. Figure 5 As shown.

[0206] A comparison of reactive power dynamic allocation schemes under scenarios with high and low proportions of renewable energy is shown below. Figure 6 As shown; based on S6 and S7, the optimized reactive power distribution scheme is executed in real time using the Automatic Voltage Control (AVC) system. Figure 3 This diagram illustrates the coordinated operation of the AVC system. The optimization effect is evaluated by calculating the voltage deviation rate (VDR) and voltage fluctuation rate (VFR) to verify the effectiveness of the scheme in the actual power grid.

[0207] The optimization effect of voltage deviation rate (VDR) and voltage fluctuation rate (VFR) is evaluated in Table 1. Judging from the voltage deviation rate (VDR) and voltage fluctuation rate (VFR), the optimization effect of the reactive power dynamic allocation scheme of this algorithm is significant and can effectively solve problems in practical engineering.

[0208] Table 1

[0209]

[0210] Therefore, the present invention adopts the above-mentioned dynamic reactive power allocation method for high-proportion renewable energy power grids based on the AVC system, which not only ensures the stable operation of the power system under large-scale renewable energy injection, but also achieves the optimal economic efficiency of dynamic reactive power allocation under large-scale renewable energy injection.

[0211] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A dynamic reactive power allocation method for high-proportion renewable energy power grids based on AVC systems, characterized in that, Includes the following steps: S1. Construct a dynamic time-series scenario set model that includes fluctuations in renewable energy output and uncertainties in load demand; S2. Define the effective reactive power reserve of each node in a high-proportion renewable energy grid, and determine the constraints by combining the node voltage safety range and the reactive power output limit under all operating scenarios. S3. Construct a data-physical fusion Wiener process degradation model, and update the model parameters online using Bayesian methods by combining the physical failure mechanism of the equipment with historical operating data. S4. Establish a two-stage distributed bar optimization model. The first stage optimizes the dynamic allocation scheme of reactive power. The second stage considers the source load reduction, voltage deviation and health degradation of reactive power reserve equipment under the worst scenario. The health degradation is directly quantified into economic cost and embedded into the optimization target. S5. By using the second-order cone relaxation method to handle nonlinear power flow constraints, the above model is transformed into a mixed-integer second-order cone programming problem. S6. The column and constraint generation algorithm CCG is used in combination with the Gurobi solver to solve the model; S7. Utilize the automatic voltage control system to execute the optimized reactive power distribution scheme in real time. S8. Evaluate the optimization effect through voltage deviation rate and voltage fluctuation rate indicators.

2. The method for dynamic reactive power allocation in a high-proportion renewable energy power grid based on an AVC system according to claim 1, characterized in that, Step S1 specifically includes: Each individual scene is represented as: ； ； in, Indicates inclusion A complete set of scenes for each scenario. Indicates inclusion The first energy station One scenario, Indicates the first The energy station at the first Time in each scene A complete set of scenes; Represented as: ； Each element includes a node number. Active power output and reactive power output As shown in the following formula: ； The complete set of dynamic time series uncertainty scenarios for the power grid is obtained, as shown in the following equation: ； The modeling of the dynamic load scenario is as follows: ； ； in, Indicates the first In the first scenario One load, and Indicates the load point The uncertainty variable is obtained as follows: The scenario is represented as follows: ； The complete set of scenarios for uncertain loads is mathematically expressed as follows: ； in, Indicates inclusion A complete load scenario set for each fluctuating load node, integrating new energy scenario sets. With load scenario set This yields a set of system-wide uncertainty scenarios for the entire power grid: ； in, This represents a complete set of scenarios that includes grid loads and renewable energy stations.

3. The high-proportion renewable energy grid dynamic reactive power allocation method based on an AVC system according to claim 2, characterized in that, Step S2 specifically includes: S21. Define the effective reactive power reserve as the smaller of the following two values: (1) the maximum / minimum reactive power capacity that can maintain the node voltage within the safe operating range within the effective range. , Secondly, under all operating scenarios, the maximum positive / negative reactive power output of the node. , ; S22. Determine the core constraints, specifically including: Current constraints: ； in, Indicates in the scene Next Time Node The active power adjustment amount, Indicates in the scene Next node In time The active power load demand, Indicates in the scene Next node In time voltage amplitude, Represents a node , The electrical conductivity between them Represents a node , The susceptance between them Indicates in the scene Next node In time voltage phase angle, Indicates in the scene Next Time Node reactive power adjustment amount Indicates in the scene Next node In time reactive load demand, Indicates in the scene Next node In time voltage amplitude, Indicates in the scene Next node In time voltage phase angle, Represents the set of all nodes; Effective reactive power reserve boundary constraints: ； in, and They represent the scenes respectively. Next node The maximum effective positive and negative reactive power that affects voltage regulation; Voltage upper and lower limit constraints: ； in, for Nodes in the scene Lower voltage limit for Nodes in the scene Voltage limit; Normalization constraints: ； in, Represents the infinite norm, This represents the initial probability of a discrete scenario. Represents the probability of a scenario. This represents the range of probability variation allowed under the ∞-norm constraint, at a 95% confidence level. The calculation expression is: ； in, Indicates the confidence level. Indicates the number of scenes.

4. The method for dynamic reactive power allocation in a high-proportion renewable energy power grid based on an AVC system according to claim 1, characterized in that, Step S3 specifically includes: S31. Establish a Wiener process degradation model for data-physical fusion, and analyze the performance degradation process of reactive power equipment. Modeled as a Wiener process with drift, the expression is: ； in, For a moment The amount of equipment performance degradation, This is the initial degradation amount. The drift coefficient, Where is the diffusion coefficient. For standard Brownian motion, The cumulative action intensity function is expressed as follows: ； in, and Is the equipment in constant The reactive power output and current, and For the rated value, the index and These are the physical characteristic parameters of the equipment; S32. Key parameters of the model based on the Bayesian update method Online learning includes: S321. Set parameters based on group data of device models. prior distribution ; S322, within the time window Internally, a series of degradation observation values ​​of the equipment are obtained through an online monitoring system. Based on the properties of the Wiener process, a likelihood function is constructed. ; S323. According to Bayes' theorem, calculate the posterior distribution of the parameters: ； S324, From the posterior distribution The current optimal estimate of the parameters is obtained and used for optimization decision-making in the next cycle; S33. Introducing surge early warning characteristic: kurtosis index The calculation formula is: ； in, Indicates kurtosis. Represent a random variable, express The expected value or average value; Warning judgment criteria: Gaussian signal under normal operating conditions Resonance or short-circuit faults are characterized by spike signals. A capacitor open-circuit fault results in a flat signal. .

5. The method for dynamic reactive power allocation in a high-proportion renewable energy power grid based on an AVC system according to claim 1, characterized in that, Step S4 specifically includes: S41. The first stage of optimizing the dynamic reactive power allocation scheme has the following objective function: ； in, This represents the complete set of time steps in the voltage recovery process, while Indicates in the scene Next Time Node reactive power adjustment amount This represents the set of nodes in the first phase. S42. The second stage considers the worst-case scenario of load reduction, voltage deviation, and the health degradation of reactive power reserve equipment. The goal of this sub-problem is to identify the worst-case probability distribution that maximizes the minimum load reduction cost under N-1 emergency conditions, expressed as: ； For each scenario , and They represent the scenes respectively. Mid-time node The active and reactive load demands, among which, Quantified nodes In time The load reduction ratio, For nodes Load priority weights, This represents the set of nodes in the second phase. Calculate voltage deviation and reactive power reserve equipment health degradation: ； ； in, This indicates that due to voltage changes, the node and nodes In time The reactive power loss generated between them; For equipment health costs, This is a collection of all reactive devices. For equipment The health cost coefficient, For the scene ,time ,equipment The predicted health index is calculated online by the data-physical fusion Wiener process model established in step S3. ； in, It is equipment In the scene The amount of degradation, It is the critical degradation level that leads to equipment failure.

6. The method for dynamic reactive power allocation in a high-proportion renewable energy power grid based on an AVC system according to claim 3, characterized in that, Step S5 specifically includes: By replacing the nonlinear power flow constraints with a strong second-order cone relaxation method, the mixed-integer nonlinear optimization problem is transformed into a linear optimization problem, avoiding local optima while satisfying the following constraints: ； in, Indicates in the scene Next node In time The active output, Indicates in the scene Lower reactive power absorption node In time The active output, For nodes and nodes The product of the real parts of the voltages, For nodes and nodes The product of the imaginary voltages between them Represents a node With nodes The square of the real part of the voltage. Represents a node With nodes The square of the imaginary part of the voltage. For nodes The square of the voltage modulus, Represents a node The square of the voltage modulus, Indicates in the scene Lower reactive power absorption node In time reactive power, Indicates in the scene Downstream power node In time reactive power, Indicates in the scene Next reactive power compensation node In time reactive power, Indicates time Located at node The status of the parallel reactive power compensation equipment in operation. Indicates whether or not to invest.

7. The method for dynamic reactive power allocation in a high-proportion renewable energy power grid based on an AVC system according to claim 1, characterized in that, Step S6 specifically includes: The objective function is decomposed to obtain the main problem and sub-problems. The min problems in each scenario of the sub-problems are independent of each other. The sub-problems are decomposed into multiple separate problems and solved independently in sequence. By solving the sub-problems, the probability distribution with the highest robust cost is obtained. This distribution is substituted into the main problem for iterative calculation, thereby realizing the solution of the two-stage sub-Bruker Reactive Power Optimization (DRO) model. After solving, the dynamic reactive power injection of each bus node of the system at different times is obtained.

8. The method for dynamic reactive power allocation in a high-proportion renewable energy power grid based on an AVC system according to claim 1, characterized in that, Step S7 specifically includes: The power grid condition monitoring system acquires real-time operating signals from the power grid. And perform analysis and diagnosis to detect the occurrence of faults and identify the fault type; Based on the identified power grid conditions, the reactive power allocation optimization algorithm is adaptively adjusted through a closed-loop feedback mechanism. The optimized allocation scheme is then transmitted to the Central Control Center (CCC) and the Renewable Energy Cluster Controller (RES-CC) for coordinated execution. Reactive power regulation commands include nodes and its reactive power regulation As shown below: ； in, Indicates in the scene Next time interval internal nodes Dynamic reactive power distribution; The set of reactive power regulation commands for all grid nodes is as follows: ； in, This represents the set of dynamic reactive power distributions across all grid nodes within different time intervals.

9. The method for dynamic reactive power allocation in a high-proportion renewable energy power grid based on an AVC system according to claim 1, characterized in that, Step S8 specifically includes: After obtaining the solution results, the voltage deviation rate (VDR) and voltage fluctuation rate (VFR) are calculated to evaluate the optimization effect. The calculation formula for the evaluation index is as follows: ； ； in, For nodes In the scene The actual voltage at each point in time. For nodes The reference voltage, The total number of nodes. The total number of moments. For nodes The maximum value of the actual voltage amplitude fluctuating within a certain time domain.

Images