An avc closed-loop simulation evaluation method and system considering new energy characteristics

Abstract

The application discloses an AVC closed-loop simulation evaluation method and system considering new energy characteristics, and the method comprises the following steps: a closed-loop simulation environment integrating a plurality of power element models is constructed, and interacts with an AVC control strategy to be verified; historical output data of a new energy station are preprocessed and cluster analyzed, at least two groups of typical output scenes containing output time sequence curves are generated, and the probabilities of the scenes are determined; the output time sequence curves of the scenes are assembled into simulation input data, and are sequentially input into the closed-loop simulation environment to perform batch time sequence simulation, and grid operation parameters of the scenes are obtained; quantitative evaluation indexes are calculated for each scene; weighted mean and weighted variance of the quantitative evaluation indexes are calculated according to the probabilities of the scenes, and a probabilistic performance evaluation report is generated. The application realizes multi-scene automatic simulation verification and probabilistic evaluation of the AVC control strategy, and improves test efficiency and evaluation comprehensiveness.

Description

Technical Field

[0001] This invention relates to an AVC simulation evaluation method, and more particularly to an AVC closed-loop simulation evaluation method and system that considers the characteristics of new energy sources. Background Technology

[0002] With the large-scale grid connection of new energy sources such as wind power and photovoltaics, new energy units have initially acquired reactive power regulation capabilities based on grid-connected inverters. Against this backdrop, automatic voltage control (AVC) reactive power optimization strategies that consider the reactive power regulation characteristics of new energy sources have become a research hotspot in the field of new power systems. Correspondingly, the need to conduct realistic and effective simulation verification of AVC control strategies adapted to this scenario to ensure their reliable operation under complex dynamic conditions is becoming increasingly urgent.

[0003] However, existing technologies still have the following shortcomings: 1. Most existing simulation verification technologies for AVC strategies are based on traditional power grid architecture designs, failing to fully incorporate the randomness of renewable energy output and its reactive power regulation characteristics, making it difficult to adapt to the actual operating requirements of new power systems. 2. Existing technologies generally adopt simple index comparison methods, which cannot comprehensively and deeply verify the effectiveness and robustness of AVC control strategies under different operating conditions. 3. When comprehensively considering the performance of AVC control strategies under various renewable energy output scenarios, existing technologies rely on manually setting scenario parameters one by one, manually executing multiple simulations, and compiling results, a process that is cumbersome, time-consuming, and prone to errors.

[0004] Specifically, patent CN105893714B proposes an AVC detection and evaluation method based on a large power grid simulation system, which evaluates control effectiveness through preset cases. However, this method only uses static operating condition simulation and cannot adapt to the real-world scenario of dynamic fluctuations in reactive power output from new energy sources. Patents CN113821016B and CN119690040A design closed-loop test schemes for AVC substations and master stations in new energy power plants, respectively. However, they mainly focus on verifying the communication and control logic of the master station and do not systematically evaluate the effectiveness of the AVC control strategy itself. All of the above patents focus on calculating indicators such as voltage deviation and network loss under a single typical operating scenario, and cannot reflect the substantial impact of the randomness and volatility of new energy output on the strategy's effectiveness. Summary of the Invention

[0005] Objectives of the Invention: The first objective of this invention is to provide an AVC closed-loop simulation evaluation method that achieves fully automated evaluation across the entire chain, significantly improves testing efficiency and the comprehensiveness of evaluation results, and exhibits good robustness. The second objective of this invention is to provide an AVC closed-loop simulation evaluation system that implements the aforementioned method.

[0006] Technical solution: The AVC closed-loop simulation evaluation method considering the characteristics of new energy sources described in this invention includes the following steps:

[0007] (1) Construct a closed-loop simulation environment. The closed-loop simulation environment integrates the power grid model, the new energy power source model, the load model and the traditional voltage regulating equipment model, and sets up an interactive interface with the AVC control strategy to be verified, so that the AVC control strategy issues voltage regulation commands based on the real-time operating data of the simulation environment to control the reactive power output of the traditional voltage regulating equipment and the new energy power source.

[0008] (2) Preprocess the historical power output data of the new energy power station, and divide the historical daily power output data into at least two typical power output modes through cluster analysis. Use the cluster center vector of each typical power output mode as the corresponding power output time series curve to generate at least two sets of typical new energy power output scenarios and determine the probability of occurrence of each scenario; wherein, each set of typical new energy power output scenarios includes its corresponding power output time series curve.

[0009] (3) Assemble the output time-series curves of the typical output scenarios of the new energy as simulation input data, and sequentially input the simulation input data into the closed-loop simulation environment to perform batch time-series simulation to obtain the power grid operation parameters under each scenario;

[0010] (4) For each scenario, calculate the quantitative evaluation index based on the power grid operation parameters corresponding to the scenario. The quantitative evaluation index includes at least the voltage qualification rate, the active power loss reduction rate, the renewable energy reactive power utilization rate, and the total number of actions of traditional voltage regulating equipment.

[0011] (5) Based on the probability of occurrence of each scenario, perform probabilistic statistical processing on the quantitative evaluation indicators, calculate the weighted mean and weighted variance, and generate a probabilistic performance evaluation report.

[0012] Preferably, in step (2), generating at least two sets of typical new energy power output scenarios includes:

[0013] The historical daily power output data of photovoltaic and wind power were clustered separately to obtain several typical power output patterns of photovoltaic and wind power. The cluster center curve of each pattern was used as the corresponding power output time series curve.

[0014] Typical power output modes of photovoltaics and wind power are combined and paired to generate wind-solar joint power output scenarios, and the probability of occurrence of each joint scenario is determined according to the proportion of days of each mode in historical data.

[0015] Preferably, in step (2), the preprocessing includes filling missing values, removing abnormal data, and performing time-series alignment and normalization on the historical output data.

[0016] Preferably, in step (3), when assembling the output time-series curves of the typical new energy output scenarios into simulation input data, the method further includes: superimposing the load time-series curves of typical load fluctuation scenarios; the typical load fluctuation scenarios are derived from the original historical load curves of typical days of the power grid, which are divided into three categories after preprocessing: peak load, valley load, and stable base load; after superposition, the probability of each scenario follows the original probability of the superimposed typical new energy output scenario; during superposition, the load time-series curves and the output time-series curves adopt the same duration and the same sampling interval to achieve full-time point-by-point alignment and assembly.

[0017] Preferably, in step (3), when performing batch timing simulation, a simulation timeout is set for each scenario, and when a scenario simulation times out, it is automatically skipped and the exception is recorded.

[0018] Preferably, in step (4):

[0019] The voltage pass rate is obtained by statistically analyzing the percentage of time points when the voltage passes at each node and taking the average value for all nodes, or by weighting the key nodes according to preset rules.

[0020] The active power loss reduction rate is obtained by calculating the ratio of the total network loss difference between the strategy to be evaluated and the benchmark strategy;

[0021] The reactive power utilization rate of the new energy source is obtained by calculating the time integral ratio of the actual reactive power output to the maximum reactive power output of each new energy power station and taking the average value of all stations.

[0022] The total number of operations of the traditional voltage regulating equipment is obtained by summing the number of operations of the parallel capacitor bank, the parallel reactor bank, and the tap changer of the on-load tap-changing transformer.

[0023] Preferably, in step (5), the probabilistic statistical processing further includes:

[0024] After normalizing each evaluation indicator, the comprehensive score is obtained by weighting and summing the results according to the preset weights.

[0025] The test information, probabilistic statistical results, and detailed results for each scenario are integrated into an evaluation report.

[0026] Preferably, in step (2), the clustering analysis uses the K-Means++ algorithm.

[0027] The AVC closed-loop simulation evaluation system considering the characteristics of new energy sources described in this invention is used to implement the method described in this invention, including:

[0028] Memory, used to store computer programs and power grid model data;

[0029] At least one processor, coupled to the memory, is configured to execute the computer program to work in conjunction with the following modules:

[0030] The closed-loop simulation module is configured to build a closed-loop simulation environment. The closed-loop simulation environment integrates a power grid model, a new energy power source model, a load model, and a traditional voltage regulating equipment model. It also sets up an interactive interface with the AVC control strategy to be verified, so that the AVC control strategy issues voltage regulation commands based on real-time operating data, controls the reactive power output of traditional voltage regulating equipment and new energy power sources, and outputs the power grid operating parameters for each scenario.

[0031] The scenario generation and management module is configured to preprocess the historical output data of new energy power plants, divide the historical daily output data into at least two typical output modes through cluster analysis, use the cluster center vector of each typical output mode as the corresponding output time series curve, generate at least two sets of typical new energy output scenarios and determine the occurrence probability of each scenario; and assemble the output time series curve of each set of typical new energy output scenarios into simulation input data, and sequentially input the simulation input data into the closed-loop simulation module to perform batch time series simulation.

[0032] The indicator calculation module is configured to receive the power grid operation parameters and automatically calculate quantitative evaluation indicators for each scenario. The quantitative evaluation indicators include at least voltage qualification rate, active power loss reduction rate, renewable energy reactive power utilization rate, and total number of actions of traditional voltage regulating equipment.

[0033] The probabilistic statistics and evaluation output module is configured to receive the quantitative evaluation indicators and the occurrence probability of each scenario, perform probabilistic statistical processing on each indicator, calculate the weighted mean and weighted variance, and generate a probabilistic performance evaluation report.

[0034] Preferably, the scenario generation and management module is further configured to: when assembling the output time-series curve as simulation input data, superimpose the load time-series curve of a typical load fluctuation scenario; the typical load fluctuation scenario originates from the original historical load curve of a typical day of the power grid, and is divided into three categories after preprocessing: peak load, valley load, and stable base load; after superposition, the probability of each scenario follows the original probability of the superimposed typical output scenario of new energy; during superposition, the load time-series curve and the output time-series curve adopt the same duration and the same sampling interval to achieve full-time point-by-point alignment and assembly.

[0035] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages: (1) By constructing a closed-loop simulation environment that integrates the reactive power regulation characteristics of new energy sources and generating typical power output scenarios based on historical data clustering, it can realistically simulate the randomness and volatility of new energy power output, overcoming the shortcomings of existing static simulations that cannot adapt to the operation requirements of new power systems; (2) By automatically generating multiple scenarios, batch simulation scheduling and multi-dimensional index calculation, it realizes a comprehensive quantitative evaluation of AVC control strategies in terms of voltage regulation effect, economy, new energy friendliness and equipment loss degree, avoiding the one-sidedness of single index comparison; (3) Through the full-chain automated simulation process and probabilistic statistical output, there is no need to manually set scenarios one by one and manually sort out results, which greatly improves the testing efficiency and accuracy, and provides quantitative technical support for strategy access testing, parameter optimization and robustness certification. Attached Figure Description

[0036] Figure 1 This is a flowchart of the present invention;

[0037] Figure 2 This is a flowchart of the scene generation and management module of the present invention. Detailed Implementation

[0038] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0039] As shown in the attached figure, the present invention provides an AVC closed-loop simulation evaluation method and system that considers the characteristics of new energy sources, including a closed-loop simulation module, a scenario generation and management module, an index calculation module, and a probabilistic statistics and evaluation output module. Each module works together to execute the closed-loop test process.

[0040] I. Closed-loop simulation module

[0041] The closed-loop simulation module is used to establish and run a closed-loop simulation environment, realizing the closed-loop linkage between the AVC control strategy and the simulation environment, and providing realistic and reliable simulation data support for subsequent evaluation index calculations. The module mainly includes:

[0042] 1. Model Library: Power grid topology models, new energy power source models (photovoltaic power plants / wind farms), load models, and traditional voltage regulation equipment models.

[0043] 2. AVC Strategy Interface: Used to send real-time operating data to the AVC control strategy to be verified, and to receive voltage regulation commands returned by the AVC control strategy.

[0044] 3. Simulation Unit: Executes time-series simulations under specified scenarios, outputting various power grid operating data, including node voltages, grid active power losses, real-time generator output, and the operational status of traditional voltage regulating equipment. Simultaneously, it calls the AVC strategy interface to control the switching of traditional voltage regulating equipment (capacitors, reactors, etc.) and the reactive power output regulation of photovoltaic and wind power inverters / converters.

[0045] II. Scene Generation and Management Module

[0046] The scenario generation and management module is responsible for generating multiple sets of typical new energy power output scenarios from historical data and driving the closed-loop simulation module to perform batch simulations. This module mainly includes:

[0047] 1. Historical Data Acquisition and Preprocessing Unit

[0048] Historical power output data and corresponding meteorological data of new energy power plants for at least one consecutive year are obtained as samples; the sample duration is set to 24 hours, with a sampling interval of 15 minutes, forming a daily standard sample of 96 power output time series.

[0049] Perform missing value imputation, outlier removal, time-series alignment, and normalization on the original data.

[0050] 2. Cluster Analysis Unit

[0051] The clustering analysis unit uses the K-Means++ algorithm to cluster the historical daily curves of photovoltaic and wind power respectively, obtaining several typical power output patterns. The number of photovoltaic clusters K1 and wind power clusters K2 are customized by the user according to actual needs. The specific clustering process is as follows:

[0052] (1) Initialize cluster centers: K1 historical days of photovoltaic power and K2 historical days of wind power are selected as initial "cluster centers" using the K-Means++ algorithm;

[0053] (2) Sample allocation: Calculate the distance from each historical day to the center of each cluster and assign it to the nearest cluster;

[0054] (3) Update cluster center: Recalculate the average of all historical days in each cluster as the new cluster center;

[0055] (4) Iterative convergence: Repeat steps (2) and (3) until the cluster center moves less than a preset threshold (preferably 0.001 in this embodiment) or reaches the maximum number of iterations (preferably 50 in this embodiment).

[0056] (5) Output typical output curves: The final cluster center vector of each cluster is used as the typical output curve of that class.

[0057] After clustering, the percentage of days in the historical data for each pattern is counted, which is taken as the probability of occurrence of that pattern.

[0058] 3. Scene Combination and Probability Assignment Unit

[0059] By pairing the output curves of various photovoltaic modes with those of various wind power modes, basic wind-solar combined output scenarios are generated. The probability of each combined scenario is calculated based on the percentage of days each typical curve represents in historical data. The calculation formula is as follows:

[0060]

[0061] Where, N PV,i and N WD,j These represent the historical days of the i-th type of photovoltaic and the j-th type of wind power curves, respectively; N PVtotal and N WDtotal These represent the total number of historical days for photovoltaic and wind power, respectively.

[0062] 4. Simulation input assembly unit

[0063] For each output typical renewable energy scenario, the renewable energy output time-series curve for that scenario is automatically extracted. Optionally, typical load fluctuation scenarios can be superimposed to form an independent simulation task. In this embodiment, the typical load fluctuation scenarios are selected from the original historical load curves of typical days of the power grid. After basic processing such as missing value imputation and abnormal data removal, they are mainly divided into three categories: peak load scenarios, valley load scenarios, and stable base load scenarios. During testing, a single typical load fluctuation scenario can be superimposed as needed. At this time, the overall probability of each scenario directly adopts the original probability of the superimposed typical renewable energy output scenario. During superposition, the load time-series curve and the output time-series curve of the typical renewable energy output scenario use the same duration and the same sampling interval to achieve full-time point-by-point alignment and ensure that the two curves can be combined one-to-one at every time point.

[0064] Each simulation task is assembled into a simulation input data set, which specifically includes: new energy scenario types and their output time series data, load scenario types and their time series data, and new energy scenario probability weights.

[0065] 5. Batch scheduling unit

[0066] For each typical new energy scenario output, the closed-loop simulation module is called sequentially to execute the simulation.

[0067] Set a timeout for single-scenario simulation. If the simulation times out, it will be automatically skipped and the exception will be recorded to ensure the robustness of batch simulation. After each scenario simulation is completed, the system receives the running data file (containing the voltage timing of each node, network loss, number of actions of traditional voltage regulating equipment, etc.) and the received AVC control instructions from the closed-loop simulation module, and then passes them to the index calculation module.

[0068] III. Indicator Calculation Module

[0069] The index calculation module is the core calculation unit of the system. It receives the full amount of operating data under each scenario output by the closed-loop simulation module and automatically calculates the four quantitative evaluation indicators shown in Table 1 for each scenario. The four indicators reflect the performance of the AVC control strategy from four dimensions: voltage regulation accuracy, economy, new energy friendliness and equipment operation frequency.

[0070] Table 1. Four quantitative indicators used to reflect the performance of the AVC control strategy.

[0071]

[0072] 1. Voltage qualification rate

[0073] Voltage compliance rate measures the extent to which the voltage at each node of the power grid remains within the allowable range throughout the entire simulation period. The specific calculation method is as follows:

[0074] First, the proportion of simulation time points where the voltage amplitude of a single node is within the allowable range is calculated out of the total number of simulation time points. Then, the arithmetic mean of the above proportions for all nodes is taken to obtain the overall network voltage qualification rate. In this embodiment, the allowable voltage range is set to [0.95pu, 1.05pu].

[0075] In addition, weighted calculations can be performed on key nodes (such as new energy grid connection points, load center busbars, etc.) according to user specifications. The calculation formula is as follows:

[0076]

[0077] Among them, R k The voltage qualification rate is denoted as N; N is the number of nodes being evaluated; T is the voltage qualification rate in the k-th scenario. j T represents the number of time points in which the voltage at node j is within the acceptable range. total This represents the total number of simulation time points.

[0078] 2. Active power loss reduction rate

[0079] The active power loss reduction rate is used to characterize the effectiveness of the AVC strategy in reducing network losses compared to the baseline strategy (or no AVC strategy). The specific calculation method is as follows:

[0080] First, the total active power loss under the evaluated strategy in this scenario is calculated by integrating (or discretely summing) the active power losses of all lines over time. Second, the total active power loss corresponding to the baseline strategy is calculated under the same simulation scenario. The formula for calculating the active power loss reduction rate is:

[0081]

[0082] Where, ΔP loss,k P represents the active power loss reduction rate in the k-th scenario. loss0,k To optimize network loss; P loss1,k To optimize network loss.

[0083] 3. Reactive power utilization rate of new energy sources

[0084] The reactive power utilization rate of new energy sources measures the proportion of reactive power actually generated by new energy power plants such as photovoltaic and wind turbines within their reactive power capacity range. It reflects whether relevant strategies prioritize the use of new energy sources for voltage regulation. The specific calculation method is as follows:

[0085] For each renewable energy power station, calculate the ratio of its actual reactive power output curve to the time integral of its maximum reactive power output curve; then average the above ratios for all stations. The formula for calculating the reactive power utilization rate of renewable energy is:

[0086]

[0087] Where, η k N represents the reactive power utilization rate of new energy sources in the k-th scenario; new Q represents the total number of new energy power stations; i For the actual reactive power output of the i-th power station; Q max,i Its maximum non-functional capacity.

[0088] 4. Total number of operations of traditional pressure regulating equipment

[0089] Within the closed-loop simulation cycle of Automatic Voltage Control (AVC), the sum of the number of operations of three types of traditional voltage regulating equipment—parallel capacitor banks, parallel reactor banks, and on-load tap changer (OLTC) taps—is counted to measure the frequency of operation of these traditional voltage regulating devices, thereby verifying the optimization effect of the "new energy priority voltage regulation" strategy. The formula for calculating the total number of operations of traditional voltage regulating equipment is as follows:

[0090]

[0091] Where, N total,k N represents the total number of actions of the traditional voltage regulator in the k-th scenario; C,k N represents the total number of switching operations on the parallel capacitor bank. L,k N represents the total number of switching operations of the parallel reactor group;T,k This represents the total number of tap changer operations of an on-load tap-changing transformer.

[0092] The calculation results of the indicators for each scenario are integrated into a K×4 evaluation indicator matrix (where K represents the number of scenarios and 4 represents the number of indicator items), and this matrix, along with the probability weights corresponding to each scenario, is passed to the probabilistic statistics and evaluation output module.

[0093] IV. Probabilistic Statistics and Evaluation Output

[0094] The probabilistic statistics and evaluation output module is the system's terminal module. It receives the evaluation index matrices for each scenario from the index calculation module and the occurrence probabilities of each scenario from the scenario generation and management module. It then performs probabilistic statistical processing and automatically generates a probabilistic performance evaluation report. This module mainly includes:

[0095] 1. Weighted Statistic Calculation Unit

[0096] For each evaluation index X(R) k ΔP loss,k η k N total,k The weighted statistic is calculated using the following formula.

[0097] (1) Calculate the weighted mean, which reflects the average expected performance of the indicator under uncertainty. The calculation formula is as follows:

[0098]

[0099] in, X is the weighted average of the evaluation metric X; K is the sequence number of the scenario. Let be the probability of the k-th scenario; Let X be the value of the evaluation metric X in the k-th scenario;

[0100] (2) Calculate the weighted variance to reflect the degree of fluctuation of the indicator under different scenarios. The larger the variance, the greater the impact of new energy and load fluctuations on the strategy performance. The calculation formula is as follows:

[0101]

[0102] in, The weighted variance of the evaluation index X; Let be the probability of the k-th scenario; Let X be the value of the evaluation metric X in the k-th scenario; This is the weighted mean of the evaluation index X.

[0103] 2. Comprehensive scoring calculation unit (optional)

[0104] This unit is used to perform a single numerical evaluation of the AVC strategy.

[0105] (1) Normalization

[0106] Given the differences in the dimensions and optimization directions of the various evaluation indicators, the indicators are first normalized, where:

[0107] Positive indicators (such as voltage qualification rate, active power loss reduction rate, and renewable energy reactive power utilization rate): Since higher values ​​are better, the range normalization method is used for normalization. The calculation formula is as follows:

[0108]

[0109] Where X max and X min These are the maximum and minimum values ​​(or preset ideal and lower limit values) of the indicator in all scenarios.

[0110] Inverse indicators (such as the total number of operations of traditional voltage regulators): Since lower values ​​are better, they are normalized using the following formula:

[0111]

[0112] Where N represents the total number of operations of the traditional voltage regulating equipment; N max and N min These represent the maximum and minimum number of actions performed by traditional voltage regulators in all scenarios (or the preset upper limit and ideal value).

[0113] (2) Weighted summation

[0114] After normalizing each evaluation index, the values ​​of each index are all within the range of [0, 1], with larger values ​​indicating better performance. Based on this, a weighted summation method is used to calculate the comprehensive score. This involves assigning weights to each normalized evaluation index and then summing the results. The resulting comprehensive score also falls within the range of [0, 1]. A higher score indicates better overall performance of the AVC strategy. The calculation formula is as follows:

[0115]

[0116] Where S represents the overall score; , , , These are the weights corresponding to voltage qualification rate, active power loss reduction rate, renewable energy reactive power utilization rate, and the total number of operations of traditional voltage regulating equipment, respectively. , , , These are the normalized values ​​of the corresponding indicators. The weights of each evaluation indicator are shown in Table 2. These values ​​can be adjusted according to the actual application scenario; for example, in areas with a high proportion of new energy sources, ω can be increased. η While prioritizing economic efficiency, increase ω ΔP .

[0117] Table 2 Weights of Quantitative Evaluation Indicators

[0118]

[0119] 3. Report Generation Unit

[0120] This unit integrates all calculation results and automatically generates an AVC test evaluation report, which includes the following:

[0121] (1) Basic test information: AVC strategy name to be evaluated, total number of scenarios, simulation parameters, scenario composition, etc.;

[0122] (2) Probabilistic statistical results: weighted mean, weighted variance, maximum value, minimum value, etc. for each indicator;

[0123] (3) Overall score (if enabled);

[0124] (4) Detailed results table for each scenario: scenario name, probability, original values ​​of 4 indicators, and whether timeout occurred.

[0125] V. System Workflow

[0126] The scenario generation and management module generates at least two sets of typical new energy output time series scenarios and probabilities from historical data and assembles them as simulation input data.

[0127] The scenario generation and management module sequentially sends the simulation input files to the closed-loop simulation module, which then performs closed-loop timing simulation and outputs the running data for each scenario.

[0128] The indicator calculation module reads the running data, calculates the four evaluation indicators for each scenario, and outputs the indicator matrix.

[0129] The probabilistic statistics and evaluation output module receives the indicator matrix and scenario probabilities, calculates the weighted mean, variance and comprehensive score, and generates an evaluation report.

Claims

1. An AVC closed-loop simulation evaluation method considering the characteristics of new energy sources, characterized in that, Includes the following steps: (1) Construct a closed-loop simulation environment. The closed-loop simulation environment integrates the power grid model, the new energy power source model, the load model and the traditional voltage regulating equipment model, and sets up an interactive interface with the AVC control strategy to be verified, so that the AVC control strategy issues voltage regulation commands based on the real-time operating data of the simulation environment to control the reactive power output of the traditional voltage regulating equipment and the new energy power source. (2) Preprocess the historical power output data of the new energy power station, and divide the historical daily power output data into at least two typical power output modes through cluster analysis. Use the cluster center vector of each typical power output mode as the corresponding power output time series curve to generate at least two sets of typical new energy power output scenarios and determine the probability of occurrence of each scenario; wherein, each set of typical new energy power output scenarios includes its corresponding power output time series curve. (3) Assemble the output time-series curves of the typical output scenarios of the new energy as simulation input data, and sequentially input the simulation input data into the closed-loop simulation environment to perform batch time-series simulation to obtain the power grid operation parameters under each scenario; (4) For each scenario, calculate the quantitative evaluation index based on the power grid operation parameters corresponding to the scenario. The quantitative evaluation index includes at least the voltage qualification rate, the active power loss reduction rate, the renewable energy reactive power utilization rate, and the total number of actions of traditional voltage regulating equipment. (5) Based on the probability of occurrence of each scenario, perform probabilistic statistical processing on the quantitative evaluation indicators, calculate the weighted mean and weighted variance, and generate a probabilistic performance evaluation report.

2. The method according to claim 1, characterized in that, In step (2), generating at least two sets of typical new energy power output scenarios includes: The historical daily power output data of photovoltaic and wind power are clustered separately to obtain several typical power output modes of photovoltaic and wind power. The cluster center vector of each mode is used as the corresponding power output time series curve. Typical power output modes of photovoltaics and wind power are combined and paired to generate wind-solar joint power output scenarios, and the probability of occurrence of each joint scenario is determined according to the proportion of days of each mode in historical data.

3. The method according to claim 1, characterized in that, In step (2), the preprocessing includes filling missing values, removing abnormal data, and performing time-series alignment and normalization on the historical output data.

4. The method according to claim 1, characterized in that, In step (3), when assembling the output time-series curves of the typical new energy output scenarios into simulation input data, it also includes: superimposing the load time-series curves of typical load fluctuation scenarios; the typical load fluctuation scenarios are derived from the original historical load curves of typical days of the power grid, which are divided into three categories after preprocessing: peak load, valley load and stable base load; after superposition, the probability of each scenario follows the original probability of the superimposed typical new energy output scenario; during superposition, the load time-series curve and the output time-series curve adopt the same duration and the same sampling interval to achieve full time-series point-by-point alignment and assembly.

5. The method according to claim 1, characterized in that, In step (3), when performing batch timing simulation, a simulation timeout is set for each scenario. When a scenario simulation times out, it is automatically skipped and the exception is recorded.

6. The method according to claim 1, characterized in that, In step (4): The voltage pass rate is obtained by statistically analyzing the percentage of time points when the voltage passes at each node and taking the average value for all nodes, or by weighting the key nodes according to preset rules. The active power loss reduction rate is obtained by calculating the ratio of the total network loss difference between the strategy to be evaluated and the benchmark strategy; The reactive power utilization rate of the new energy source is obtained by calculating the time integral ratio of the actual reactive power output to the maximum reactive power output of each new energy power station and taking the average value of all stations. The total number of operations of the traditional voltage regulating equipment is obtained by summing the number of operations of the parallel capacitor bank, the parallel reactor bank, and the tap changer of the on-load tap-changing transformer.

7. The method according to claim 1, characterized in that, In step (5), the probabilistic statistical processing further includes: After normalizing each evaluation indicator, the comprehensive score is obtained by weighting and summing the results according to the preset weights. The test information, probabilistic statistical results, and detailed results for each scenario are integrated into an evaluation report.

8. The method according to claim 1, characterized in that, In step (2), the clustering analysis uses the K-Means++ algorithm.

9. An AVC closed-loop simulation evaluation system considering the characteristics of new energy sources, used to implement the method described in any one of claims 1 to 8, characterized in that, include: Memory, used to store computer programs and power grid model data; At least one processor, coupled to the memory, is configured to execute the computer program to work in conjunction with the following modules: The closed-loop simulation module is configured to build a closed-loop simulation environment. The closed-loop simulation environment integrates a power grid model, a new energy power source model, a load model, and a traditional voltage regulating equipment model. It also sets up an interactive interface with the AVC control strategy to be verified, so that the AVC control strategy issues voltage regulation commands based on real-time operating data, controls the reactive power output of traditional voltage regulating equipment and new energy power sources, and outputs the power grid operating parameters for each scenario. The scenario generation and management module is configured to preprocess the historical output data of new energy power plants, and divide the historical daily output data into at least two typical output modes through cluster analysis. The cluster center vector of each typical output mode is used as the corresponding output time series curve to generate at least two sets of typical new energy output scenarios and determine the probability of occurrence of each scenario. The output time-series curves of each typical new energy output scenario are assembled into simulation input data, and the simulation input data is sequentially input into the closed-loop simulation module to perform batch time-series simulation. The indicator calculation module is configured to receive the power grid operation parameters and automatically calculate quantitative evaluation indicators for each scenario. The quantitative evaluation indicators include at least voltage qualification rate, active power loss reduction rate, renewable energy reactive power utilization rate, and total number of actions of traditional voltage regulating equipment. The probabilistic statistics and evaluation output module is configured to receive the quantitative evaluation indicators and the occurrence probability of each scenario, perform probabilistic statistical processing on each indicator, calculate the weighted mean and weighted variance, and generate a probabilistic performance evaluation report.

10. The system according to claim 9, characterized in that, The scenario generation and management module is further configured to: when assembling the output time-series curve as simulation input data, superimpose the load time-series curve of a typical load fluctuation scenario; the typical load fluctuation scenario originates from the original historical load curve of a typical day of the power grid, and is divided into three categories after preprocessing: peak load, valley load, and stable base load; after superposition, the probability of each scenario follows the original probability of the superimposed typical output scenario of new energy; during superposition, the load time-series curve and the output time-series curve adopt the same duration and the same sampling interval to achieve full time-series point-by-point alignment and assembly.

Images