A virtual power plant heterogeneous resource regulation capability evaluation method and system

Abstract

The present application belongs to the technical field of power system reliability evaluation and virtual power plant operation optimization, and provides a virtual power plant heterogeneous resource regulation capacity evaluation method and system. The technical scheme is that, for multiple source uncertainty variables, an empirical Copula function is obtained by fitting historical data, and a time series scenario set is generated based on the empirical Copula function and a trained stacked auto-encoder model. Based on the generated time series scenario set, the regulation capacity of all uncertainty scenarios is calculated, and a regulation capacity sample set is constructed based on the calculated regulation capacity of all uncertainty scenarios. Based on the constructed regulation capacity sample set, a multi-confidence level regulation curve library is constructed. The multi-confidence level regulation curve library is embedded as a constraint condition in the intraday rolling scheduling, and the power supply reliability criterion is set to calculate the power supply reliability, and the probabilistic evaluation result of the power supply reliability is obtained. It can adapt to the access demand of different types and different sizes of distributed energy resources.

Description

Technical Field

[0001] This invention belongs to the field of power system reliability assessment and virtual power plant operation optimization technology, and particularly relates to a method and system for assessing the heterogeneous resource regulation capability of a virtual power plant. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Virtual power plants utilize advanced information and communication technologies to aggregate and manage heterogeneous resources such as distributed photovoltaic, wind power, energy storage systems, gas turbine units, and controllable loads, achieving unified dispatch and optimized operation. With the deepening of power market reforms, virtual power plants' participation in capacity application and dispatch decisions in the electricity market requires accurate quantification of their available regulatory capacity. However, the regulatory capacity of virtual power plants is directly affected by multi-source uncertainties such as fluctuations in wind and solar power output and load changes. How to accurately assess the reliable regulatory capacity of virtual power plants under uncertain environments has become a key technological bottleneck restricting their safe and stable operation and market application.

[0004] Currently, the scale and types of heterogeneous resources aggregated by virtual power plants are continuously expanding. These heterogeneous resources differ significantly in terms of output characteristics, adjustability, and sources of uncertainty: wind power output is random and intermittent, photovoltaic output is highly time-varying due to diurnal and weather conditions, load demand fluctuates with user behavior, energy storage systems are constrained by state of charge (SOC), and gas turbine units are limited by ramp rate. More importantly, these heterogeneous resources exhibit complex temporal correlations and complementary characteristics. For example, wind power and photovoltaic output are negatively correlated (strong winds often lead to weak sunlight), while photovoltaic output and load demand are positively correlated (increasing synchronously during the day). This complex correlation structure results in the overall regulation capability of virtual power plants exhibiting significant time-varying, random, and correlational characteristics.

[0005] Traditional methods for assessing the regulation capacity of virtual power plants are mainly divided into two categories: deterministic assessment and simple probabilistic assessment. Deterministic assessment methods calculate the regulation capacity of virtual power plants based on historical data or predicted values. While computationally simple, they cannot quantify the impact of multi-source uncertainties such as wind and solar power output fluctuations and load changes on regulation capacity, making it difficult to meet the reliability requirements of power systems. Simple probabilistic assessment methods, although introducing probability distributions to describe uncertainties, often assume that resources are independent or only consider linear correlations (such as the Pearson correlation coefficient), ignoring the complex nonlinear correlation structures between heterogeneous resources (especially tail correlations in extreme scenarios). This leads to insufficient assessment accuracy and is prone to significant deviations during high-risk periods. Therefore, traditional deterministic or simple probabilistic assessment methods struggle to accurately quantify reliable regulation capacity.

[0006] In recent years, researchers have gradually recognized the advantages of data-driven methods in handling high-dimensional nonlinear uncertainty problems. Machine learning techniques can extract operational patterns and correlation characteristics of heterogeneous resources from historical operational data, providing a new technical approach for assessing the regulation capacity of virtual power plants. However, existing data-driven methods for assessing the regulation capacity of virtual power plants often focus on average regulation capacity or assessments at a single confidence level, lacking quantitative assessment methods with multiple confidence levels to cater to different risk preferences. However, the reliability requirements of power system operation are multi-layered: lower confidence levels are acceptable for daily operation, while higher confidence levels are needed during critical periods. The lack of multi-confidence level quantification methods makes it difficult to support differentiated risk management and dispatch decisions for virtual power plants. Summary of the Invention

[0007] To address at least one of the technical problems mentioned above, this invention provides a method and system for assessing the heterogeneous resource regulation capacity of a virtual power plant. This method generates a large number of uncertain scenario samples, quantifies the callable regulation capacity of the virtual power plant under different confidence levels using a probability distribution curve library, and embeds the regulation curve library as a constraint into intraday rolling scheduling to achieve a probabilistic assessment of power supply reliability, identifying critical risk periods and vulnerable equipment. It can adapt to the access needs of different types and scales of distributed energy resources.

[0008] To achieve the above objectives, the present invention adopts the following technical solution:

[0009] The first aspect of this invention provides a method for evaluating the heterogeneous resource regulation capability of a virtual power plant, comprising the following steps:

[0010] For multi-source uncertain variables, an empirical Copula function is obtained by fitting historical data, and a time series scene set is generated based on the empirical Copula function and the trained stacked autoencoder model.

[0011] Based on the generated time series scenario set, the adjustment capacity of all uncertain scenarios is calculated, and an adjustment capacity sample set is constructed based on the calculated adjustment capacity of all uncertain scenarios; based on the constructed adjustment capacity sample set, a multi-confidence level adjustment curve library is constructed.

[0012] By embedding the multi-confidence level adjustment curve library as a constraint into the intraday rolling scheduling, and combining it with the reliability calculation of power supply based on the set power supply reliability criteria, a probabilistic assessment result of power supply reliability is obtained.

[0013] Furthermore, the generation of time series scene sets based on the empirical Copula function and the trained stacked autoencoder model includes:

[0014] Based on the empirical Copula function, variables including wind power, solar power, and load uncertainties are generated. t A static scene that reflects the characteristics of a historical joint distribution at any given moment;

[0015] The generated t A static scene at a given time is input into a trained stacked autoencoder network, which predicts... t The reference scene at time +1 is used to calculate the difference between the reference scene and the sampled scene.

[0016] The difference between the reference scenario and the sampled scenario is compared. If the difference meets the set threshold condition, the generated time series scenario is reasonable and is retained. Otherwise, the scenario is resampled using the empirical Copula function until the data meets the requirements, and the final time series scenario set is obtained.

[0017] Furthermore, the calculation of the adaptability of all uncertain scenarios based on the generated time-series scenario set includes:

[0018] Calculation time Scene The power boundaries of the system include the maximum available power and the minimum required power;

[0019] According to time Scene Power boundary calculation time of the lower system Scene The upward and downward adjustment capabilities of the lower system;

[0020] N scenes and times generated by combining time series scene sets Scene The upregulation and downregulation capabilities of the system are used to construct a sample set of the regulation capabilities at time t.

[0021] Furthermore, based on the constructed set of adjustment capability samples, a multi-confidence level adjustment curve library is constructed, including:

[0022] For the set of samples with adjustment capabilities, construct an empirical cumulative distribution function;

[0023] The ability to regulate is calculated based on the empirical cumulative distribution function. p quantiles;

[0024] Based on regulatory capacity p Quantiles define a set of confidence levels, for each confidence level... p We construct upward and downward adjustment curves and assemble them to obtain a conditional adjustment curve library.

[0025] Furthermore, the reliability calculation of power supply based on the multi-confidence level adjustment curve library and the set power supply reliability criteria yields the power supply reliability for various scenarios, including:

[0026] The total adjustment requirements of the computing system include total upward adjustment requirements and total downward adjustment requirements;

[0027] Query the system's total regulation capacity limit from the regulation curve library, and determine the capacity adequacy based on the total regulation capacity limit and total regulation demand;

[0028] The first power supply reliability criterion is determined based on the capacity adequacy to determine whether the system power supply is reliable. If the power supply is reliable, the constructed cost optimization model is solved. The second power supply reliability criterion is used to determine whether the power supply is reliable. If so, the model optimization for this time step in this scenario is successful.

[0029] The power supply reliability is evaluated for all scenarios, resulting in a probabilistic evaluation of power supply reliability.

[0030] Furthermore, the cost optimization model includes an objective function and constraints; the objective function includes corrected maintenance costs, corrected fuel costs, energy storage degradation deviation, adjustment penalties, and energy storage SOC trajectory tracking deviation adjustment penalties; the constraints include power balance constraints, unit up-adjustment capability constraints, and unit down-adjustment capability constraints.

[0031] Furthermore, the first power supply reliability criterion is:

[0032] ,

[0033] The second power supply reliability criterion is:

[0034] ,

[0035] when and When both values ​​are 1, it indicates that the power supply is reliable; otherwise, the power supply is unreliable.

[0036] in, As a unit-level criterion indication, It is an indicative variable used to indicate whether there are excessive adjustments at the unit level. Indicates time Scene The capacity adequacy has been adjusted upwards. Indicates time Scene The capacity adequacy of the reduction.

[0037] A second aspect of the present invention provides a virtual power plant heterogeneous resource regulation capability assessment system, comprising:

[0038] The scene generation module is used to fit an empirical Copula function using historical data for multi-source uncertain variables, and to generate a time series scene set based on the empirical Copula function and the trained stacked autoencoder model.

[0039] The module for building a multi-confidence level adjustment curve library is used to calculate the adjustment capability of all uncertain scenarios based on the generated time series scenario set, construct an adjustment capability sample set based on the calculated adjustment capability sample set, and construct a multi-confidence level adjustment curve library based on the constructed adjustment capability sample set.

[0040] The capability assessment module is used to embed the multi-confidence level adjustment curve library as a constraint into the intraday rolling scheduling, and combine it with the set power supply reliability criteria to calculate the reliability of power supply, so as to obtain the probabilistic assessment result of power supply reliability.

[0041] A third aspect of the present invention provides a computer-readable storage medium.

[0042] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above-described method for assessing the heterogeneous resource regulation capacity of a virtual power plant.

[0043] A fourth aspect of the present invention provides a computer device.

[0044] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the above-described method for assessing the heterogeneous resource regulation capacity of a virtual power plant.

[0045] Compared with the prior art, the beneficial effects of the present invention are:

[0046] This invention constructs a CSML uncertainty modeling framework based on the fusion of Copula functions and stacked autoencoders, establishes a joint probability model of multi-source uncertainty, generates probability-regulation capacity distribution curves, and designs a time-varying and global reliability index system. This enables accurate quantification of callable regulation capacity under different confidence levels, effectively solving the problems of multi-source uncertainty correlation modeling and high-dimensional time series data feature extraction. It significantly improves the accuracy and reliability of virtual power plant regulation capacity assessment, provides a scientific quantitative basis for the safe and stable operation and scheduling decisions of virtual power plants, and supports the reliable operation of new power systems and the efficient consumption of distributed energy.

[0047] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0048] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0049] Figure 1 This is a flowchart of a method for evaluating the heterogeneous resource regulation capability of a virtual power plant, provided by an embodiment of the present invention.

[0050] Figure 2 This is a flowchart of the adjustment library curve construction provided in an embodiment of the present invention;

[0051] Figure 3 This is a flowchart of power supply reliability calculation provided in an embodiment of the present invention;

[0052] Figure 4 The results are correlation analysis results generated based on the CSML uncertainty model provided in this embodiment of the invention; wherein, (a) is a scatter plot of the joint distribution of wind power output and photovoltaic power output, (b) is a scatter plot of the joint distribution of wind power output and load demand, (c) is a scatter plot of the joint distribution of photovoltaic power output and load demand, and (d) is a comprehensive quantitative matrix of the correlation coefficients of wind power, photovoltaic power, and load.

[0053] Figure 5 This is the theoretical regulation capacity distribution provided in the embodiments of the present invention; wherein, (a) is the theoretical regulation capacity distribution for wind power output scenarios, (b) is the theoretical regulation capacity distribution for photovoltaic power output scenarios, and (c) is the theoretical regulation capacity distribution for load demand scenarios;

[0054] Figure 6 This is a schematic diagram of the construction of a typical probability-capacity adjustment curve library provided in the embodiments of the present invention, wherein (a) is an upward adjustment capability adjustment curve and (b) is a downward adjustment capability adjustment curve;

[0055] Figure 7 This is the adjustment capability decomposition provided in the embodiments of the present invention, wherein (a) is the upward adjustment capability decomposition and (b) is the downward adjustment capability decomposition;

[0056] Figure 8 The results of power supply reliability analysis under the constraints of the adjustment curve library at different confidence levels provided in the embodiments of the present invention are as follows: (a) is the power supply reliability rate at different confidence levels, and (b) is the comparison result of the overall reliability index.

[0057] Figure 9 This is a power supply reliability analysis under the constraints of the P95 adjustment curve library provided in the embodiments of the present invention, wherein (a) is the analysis of the adequacy of the upward adjustment capacity, and (b) is the analysis of the adequacy of the downward adjustment capacity.

[0058] Figure 10This is a visualization of the scheduling results at each time step under the constraints of the P95 adjustment curve library provided in this embodiment of the invention. Detailed Implementation

[0059] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0060] It should be noted that the following detailed description is illustrative and intended to provide further explanation of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0061] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0062] The existing methods for assessing the regulation capacity of virtual power plants face three major technical challenges: (1) a lack of precise quantitative methods for dispatchable capacity under multiple confidence levels, making it difficult to meet the differentiated reliability requirements of the power market; (2) deterministic assessment methods cannot quantify the impact of multi-source uncertainties such as wind and solar power output fluctuations and load changes on regulation capacity; existing methods mostly adopt independence assumptions or simple linear correlation models, making it difficult to accurately capture the complex nonlinear correlation between wind and solar power output and load changes, especially the tail correlation in extreme scenarios, resulting in inaccurate assessment of system risk. For example, traditional methods assume that wind power and photovoltaic power are independently distributed, which will underestimate the probability of both having low output at the same time, and thus overestimate the reliable regulation capacity of virtual power plants. (3) limited ability to extract features from high-dimensional time-series data: virtual power plants aggregate a large number of heterogeneous resources, and the operating data exhibits characteristics such as high dimension, nonlinearity, and time-series correlation. Traditional statistical methods can only extract linear features, and shallow neural networks are difficult to mine deep nonlinear laws, resulting in insufficient accuracy and generalization ability in the assessment of regulation capacity. (4) Existing methods mostly use daily-scale averaging for assessment, which makes it difficult to characterize the hourly variation characteristics of the virtual power plant's regulation capacity and effectively identify critical risk periods, affecting the pertinence and effectiveness of operational decisions. The inability to effectively identify critical risk periods in system operation leads to a lack of pertinence in dispatching decisions.

[0063] Meanwhile, the heterogeneous resources aggregated by virtual power plants, such as distributed photovoltaic, wind power, energy storage systems, gas turbines, and controllable loads, have significant differences in output characteristics, response speed, and sources of uncertainty. Furthermore, there are complex temporal correlations and complementary characteristics among these resources, making it difficult to accurately quantify the overall regulation capacity.

[0064] This invention proposes a method for assessing the regulation capacity of heterogeneous resources in a virtual power plant. This method targets heterogeneous resources such as distributed photovoltaic (PV), wind power, energy storage systems, gas turbines, and controllable loads aggregated in a virtual power plant. First, it utilizes a quasi-steady-state model that balances computational accuracy and efficiency and accurately captures dynamic changes in equipment, along with a multi-timescale virtual power plant scheduling optimization model, as the basic framework for calculating regulation capacity. Then, it captures the nonlinear correlation structure between wind power output, PV output, and load demand, performing deep feature extraction and dimensionality reduction on high-dimensional time-series operational data to establish a CSML joint uncertainty model, achieving joint probabilistic modeling of wind and solar power output fluctuations and load changes. Based on the CSML model, a large number of uncertainty scenario samples are generated through Monte Carlo simulation. Under each scenario, the scheduling optimization model is solved to obtain the up-regulation and down-regulation capacities. A probability distribution curve library is constructed through statistical analysis, and time-varying reliability and global reliability indices are designed to quantify the available regulation capacity of the virtual power plant under different confidence levels. Finally, the regulation curve library is embedded as a constraint condition into intraday rolling scheduling to achieve a probabilistic assessment of power supply reliability and identify critical risk periods and vulnerable equipment.

[0065] The method proposed in this invention enables probabilistic and accurate assessment of the regulation capacity of virtual power plants, effectively identifying operational weaknesses and risk periods, and providing a scientific quantitative basis for capacity application and operational optimization for virtual power plants participating in the power ancillary services market. Furthermore, the method exhibits good versatility and scalability, adapting to the access needs of different types and scales of distributed energy resources, supporting the safe and reliable operation and flexible regulation capabilities of virtual power plants in new power systems, and promoting the efficient consumption of distributed energy.

[0066] Example 1

[0067] like Figure 1 As shown, this embodiment provides a method for evaluating the heterogeneous resource regulation capability of a virtual power plant, including the following steps:

[0068] Step 1: For multi-source uncertain variables, generate a time series scenario set by combining the constructed CSML joint probability model;

[0069] In this embodiment, based on Copula theory and deep learning technology, a CSML joint probability model is constructed to accurately generate time series scenarios for multi-source uncertainties in wind power, photovoltaics, and load. The CSML joint probability model is proposed based on empirical Copula and a stacked autoencoder model. Empirical Copula is used to establish the joint probability distribution of uncertainties, while the stacked autoencoder model can simulate the time-varying relationship between each system variable. The implementation of empirical Copula at time t is related to its implementation at time t+1, and this relationship can be established using the stacked autoencoder model. Specifically, the following steps are included:

[0070] Step 101: Obtain the empirical Copula function by fitting historical data, and generate the sampling scene at the current moment based on the empirical Copula function;

[0071] First, based on historical operating data, empirical marginal distribution functions for three uncertain variables—wind power, photovoltaic power, and load—are constructed respectively. , , and the empirical Copula function corresponding to their joint distribution. ,in .

[0072] According to Sklar's theorem, any n The joint distribution function of dimension can be expressed as:

[0073] (1),

[0074] in, for A random variable, Let each variable be a marginal distribution function. for 3D real space, To and Related coupling functions, It is in The multivariate distribution function on the x-axis has margins that are standard uniform distributions. Therefore, the coupling function It is unique, and functions can be related through the following relationships:

[0075] (2),

[0076] in, For a uniformly distributed variable vector, For the theoretical Coupula function, express The generalized quantile function, and its generalized inverse, can be defined as:

[0077] (3),

[0078] in, For the first A uniformly distributed variable.

[0079] Based on this theory, a single time point containing wind power, photovoltaic power, and load is generated. t The specific steps for operating a static scene are as follows:

[0080] Sampling: from a three-dimensional uniform distribution Draw a random vector from This vector implies the empirical Copula function The correlation between variables is described.

[0081] Inverse Transformation: The uniform random numbers are transformed using empirical quantile functions of each variable to obtain the random demand values ​​of the three uncertain variables—wind power, photovoltaic power, and load—at that moment, i.e., the sampling scenario at the current moment.

[0082] (4),

[0083] (5),

[0084] (6),

[0085] This is how we obtain ( , , ) is a t A static scene that is present in a given moment and conforms to the characteristics of a historical joint distribution is denoted as a scene set. , This represents the vector consisting of the values ​​of all uncertain variables (wind power, solar power, and load) at time t in the s-th scenario.

[0086] Step 102: Generate t A static scene at a given time is input into a trained stacked autoencoder network, which predicts... t The reference scene at time +1 is used to calculate the difference between the reference scene and the sampled scene.

[0087] Since the scene sequence generated in step 101 is independent in the time dimension and does not consider the dynamic characteristics between adjacent time points, in order to generate scenes with reasonable temporal evolution patterns, the generated scenes will be... t A static scene at a given time is input into a trained stacked autoencoder network, which predicts... t The reference scene at time +1 is used to calculate the difference between the reference scene and the sampled scene.

[0088] In this embodiment, the stacked autoencoder (SAE) network consists of two parts: an encoder and a decoder.

[0089] (7),

[0090] (8),

[0091] in, For encoder parameters, These are decoder parameters. and These are the input data and the reconstructed output. n For the input data dimensions, It is the raw data. Feature representation in low-dimensional space and These represent parameter-dependent activation functions, including the weight matrix and the bias vector.

[0092] The training objective of an autoencoder is to minimize the reconstruction error, and the loss function is defined as:

[0093] (9),

[0094] in, The loss function, which is the square of the Euclidean norm, takes the form of mean squared error (MSE) and measures the Euclidean distance between the original input and the reconstructed output.

[0095] As a further implementation, the stacked autoencoder (SAE) architecture: Unlike autoencoders, SAE models have the same number of input and output layers, with the number of input layers greater than the number of hidden layers. Therefore, this model can generate new information by eliminating noise and yield effective properties with complex relationships.

[0096] In this embodiment, the following SAE structure is designed for three-dimensional time series data of wind power, photovoltaics, and load:

[0097] Input layer dimensions: 3 (wind, light, load); hidden layer dimensions: [64, 32, 16]; output layer dimensions: 3; training epochs: 100; batch size: 32; learning rate: α = 0.001. The SAE network was built using the PyTorch framework and trained using the Adam optimizer. For a generated... t Moment Scene The input is given to a pre-trained SAE, which predicts the sequence dynamics patterns it has learned. t A reference scenario at time +1.

[0098] Step 103: Compare the difference between the reference scene and the sampled scene. If the difference meets the set threshold condition, the generated time series scene is reasonable and is retained. Otherwise, the scene is resampled using the empirical copula function until the data meets the requirements.

[0099] In this embodiment, the set threshold condition is: If the generated time series scenario is deemed reasonable, it is retained. Otherwise, the scenario is resampled using an empirical copula function until the data meets the requirements. Then, the process moves to the next time step and continues generating time series data. In this way, a complete time series that conforms to both statistical characteristics and time dependence can be constructed step by step.

[0100] Step 2: Based on the generated time series scene set By combining uncertainty scenario analysis and probability statistics theory, a probabilistic evaluation model for the power adjustment capability of virtual power plants under different operating scenarios is established to achieve accurate quantification of callable capacity under multiple confidence levels.

[0101] like Figure 2 As shown, the specific steps include the following:

[0102] Step 201: Calculate the adjustment capability for all uncertainty scenarios, and construct a sample set of adjustment capabilities based on the calculated adjustment capabilities for all uncertainty scenarios;

[0103] Specifically, the steps include the following:

[0104] Step 2011: Calculate the power boundaries of the system, including the maximum available power and the minimum required power;

[0105] Maximum available power:

[0106] (10)

[0107] in, , For a moment Scene Reduce wind and solar power output; This is the maximum output of the gas turbine unit; This represents the maximum discharge power of the energy storage. This represents the maximum power purchase capacity of the power grid.

[0108] Minimum required power:

[0109] (11),

[0110] in, For a moment Scene Under the load demand, This is the maximum charging power for energy storage. This represents the maximum power output of the power grid.

[0111] Step 2012: Calculate the up-adjustment and down-adjustment capabilities based on the system's power boundaries;

[0112] Regulation capacity (RC) refers to the power adjustment capability of an integrated energy system in response to uncertain disturbances under a given dispatch plan.

[0113] Increase ability:

[0114] (12),

[0115] in, For a moment Scene Upward adjustment capability (MW) For a moment Scene The maximum power (MW) that the system can provide. For a moment The planned power (MW) scheduled for the day.

[0116] Reduce ability:

[0117] (13)

[0118] in, For a moment Scene Downward adjustment capability (MW) For a moment Scene The minimum power (MW) that the system must provide.

[0119] when At times, the system has the capability to increase output to cope with sudden load increases or decreases in renewable energy output. When the load drops suddenly or the output of new energy sources increases, the system can reduce its output to cope with the sudden drop in load or the increase in output of new energy sources. Conversely, it will operate at full load with no margin for adjustment.

[0120] Step 2013: Combine the generated N scenarios to construct a sample set of adjustment capabilities at time t:

[0121] (14)

[0122] (15)

[0123] in, To adjust the capability sample set, To reduce the capacity sample set, Let represent a set containing N elements, where the s-th element is the up-adjustment capability of scene s at time t. The upscaling capabilities of scenarios s from 1 to N are all included in this set. Similarly, this represents the down-adjustment capability of scenario s at time t. .

[0124] Step 202: Based on probability quantile theory and multidimensional data structure design, construct a multi-confidence level adjustment curve library for the constructed adjustment capability sample set, targeting the up-adjustment capability and down-adjustment capability at different confidence levels.

[0125] Specifically, the steps include the following:

[0126] Step 2021: For the tolerance capacity sample set, construct the empirical cumulative distribution function:

[0127] (16)

[0128] in, Adjustment capability (including upward adjustment capability) under scenario s at time t. With reduced ability All calculations are performed using this formula.

[0129] Step 2022: Calculate the regulatory capacity p Quantile:

[0130] (17)

[0131] in, For confidence level, It is the empirical cumulative distribution function of the regulatory capacity defined in formula (17). The generalized inverse function (i.e., the quantile function) means that at a given time... t This function returns a regulation capability value. x In the generated scenario, the regulation capacity of the virtual power plant is less than or equal to... x The probability is at least p This represents in t At least p The probability can ensure the boundary of the provided adjustment capability. This indicates the median adjustment capability (achievable in 50% of scenarios); This indicates high confidence adjustment capability (achievable in 95% of scenarios), and the same applies to the others.

[0132] In this embodiment, to improve the accuracy of quantile estimation, a linear interpolation method is used to optimize the quantile calculation, as shown below:

[0133] (18)

[0134] in For the sorted sample set, and The results after sorting in ascending order. This is for rounding down.

[0135] Step 2023: Based on regulatory capacity p Quantiles define the set of confidence levels;

[0136] Confidence Level p This represents the probability guarantee that the adjustment capability will meet the demand. Define the set of confidence levels:

[0137] (19)

[0138] in, In this embodiment, the number of confidence levels is set to... This corresponds to five confidence levels.

[0139] Step 2024: For the confidence level p Construct upward and downward adjustment curves, and assemble them to obtain a conditional adjustment curve library;

[0140] Upward adjustment curve: for confidence level p The upward curve for the time period [1,T] is defined as follows:

[0141] (20)

[0142] The downward curve is defined as:

[0143] (twenty one),

[0144] Note: Superscript here T It is the transpose symbol. T This represents the total number of time periods.

[0145] The adjustment curve library is a three-dimensional tensor structure:

[0146] (twenty two),

[0147] Wherein: Dimension 1: Adjustment direction (upward / downward) Dimension 2: Confidence level ( K The third dimension (at each level): Time ( T (a moment).

[0148] In this embodiment, K =5, T =24, therefore the adjustment curve library contains 2×5×24=240 data points.

[0149] Step 3: Embed the multi-confidence level adjustment curve library as a constraint condition into the intraday rolling scheduling, and combine it with the set power supply reliability criterion to calculate the reliability of power supply, so as to obtain the probabilistic evaluation result of power supply reliability.

[0150] This embodiment, based on Monte Carlo stochastic simulation theory and a multi-scenario mapping mechanism using a regulation curve library, establishes a quantitative assessment model for power supply reliability to address various stochastic disturbance factors such as load fluctuations and uncertainties in renewable energy output. Figure 3 As shown, the specific steps include the following:

[0151] Step 301: Calculate the total adjustment demand of the system, including the total upward adjustment demand and the total downward adjustment demand;

[0152] In this embodiment, the system's total upward adjustment demand... :

[0153] (twenty three),

[0154] Total system demand reduction :

[0155] (twenty four),

[0156] in, For the day-ahead scheduling plan power under scenario s at time t, This represents the actual output demand under scenario s at time t.

[0157] Therefore, the scheduling plan can be expressed as , which is the set of scheduling plans from 1 to T under scenario s.

[0158] Step 302: Query the total regulation capacity limit of the system from the regulation curve library, and determine the capacity adequacy based on the total regulation capacity limit and the total regulation demand;

[0159] System capacity increased:

[0160] (25),

[0161] System downgrade capability:

[0162] (26)

[0163] Increase capacity adequacy:

[0164] (27)

[0165] Reduce capacity adequacy:

[0166] (28)

[0167] when <0 or A value less than 0 is defined as insufficient capacity, and scheduling is considered a failure.

[0168] Step 303: Determine the first power supply reliability criterion based on the capacity adequacy to determine whether the system power supply is reliable. If the power supply is reliable, solve the constructed cost optimization model.

[0169] In this embodiment, the first power supply reliability criterion is expressed as:

[0170] (29)

[0171] Under the condition of satisfying the first power supply reliability criterion, the constructed cost optimization model is solved. The cost optimization model includes an objective function and constraints, specifically:

[0172] Intraday scheduling objective function:

[0173] (30)

[0174] Intraday scheduling is optimized with the goal of minimizing penalty costs and operating costs while meeting the physical constraints of each unit.

[0175] Adjusted maintenance costs:

[0176] (31),

[0177] Adjusted fuel costs:

[0178] (32),

[0179] in, This represents the cost of adjusting the intraday scheduling layer relative to the day-ahead plan. For the i-th gas turbine unit The amount of maintenance cost adjustment at any time. The maintenance cost of the i-th gas turbine unit, , These are the optimized intraday and day-ahead scheduling methods, respectively. arrive Unit output at any given time. for Fuel cost adjustment at any given time. For unit fuel cost, for The scheduled output of the gas turbine unit is [not specified].

[0180] Energy storage degradation bias:

[0181] (33),

[0182] Adjusting penalties:

[0183] (34),

[0184] Energy storage SOC trajectory tracking deviation adjustment penalty:

[0185] (35),

[0186] in, Adjusting the penalty coefficient for gas turbine units , These are the units under intraday and day-ahead scheduling, respectively. i exist Electric power during a period of time Adjusting the penalty coefficient for energy storage and Energy storage in intraday and day-ahead dispatch, respectively. i During the period The charging power, , For intraday and day-ahead dispatch of energy storage i During the period The discharge power, This is the penalty coefficient for SOC trajectory tracking. , Energy storage in intraday and day-ahead dispatch, respectively. i exist SOC at the end of the time period.

[0187] Power balance constraints:

[0188] (36)

[0189] Unit upscaling capacity constraints:

[0190] (37)

[0191] Unit capacity reduction constraints:

[0192] (38),

[0193] in, For units in intraday scheduling i At any moment t Scene s The plan for the next phase of effort For the units currently under dispatch i At any moment t Scene s The following is a plan for the output of resources;

[0194] (39)

[0195] (40),

[0196] , The units i At any moment t Scene s The up and down adjustment volume , The units i At confidence level p Next moment t The upward and downward adjustment amounts.

[0197] Step 304: Determine whether the power supply is reliable based on the second power supply reliability criterion. If so, the model optimization for this time step in this scenario is successful.

[0198] The second power supply reliability criterion is:

[0199] (41),

[0200] in, This is an indicative variable. A value of 1 indicates that at time t and scenario s, the actual regulation behavior of all units does not exceed their promised regulation capacity boundary at the specified confidence level p, meaning that the regulation constraints at the unit level are met and the power supply is reliable. A value of 0 indicates that at least one unit's regulation exceeds the limit (including curve library constraints and physical regulation capacity constraints), and the power supply is unreliable from the perspective of unit regulation capacity. As a unit-level criterion indication, when A value of 1 indicates reliable power supply. When the value is 0, the power supply is unreliable.

[0201] The optimization of this time step in this scenario is considered successful only if both the first power supply reliability criterion and the second power supply reliability criterion are met simultaneously.

[0202] Step 304: Evaluate the power supply reliability for all scenarios and obtain the power supply reliability calculation results;

[0203] The reliability at time t is defined as the proportion of scenarios in N evaluation scenarios where intraday rolling scheduling optimization is successful:

[0204] (42),

[0205] in, For a moment t Power supply reliability, This is an indicator function, which takes the value 1 when the intraday rolling optimization is successful in scenario s at time t, and 0 otherwise.

[0206] Power supply reliability in various scenarios:

[0207] (43),

[0208] in, Let T represent the power supply reliability of scenario s, and T be the total number of time periods throughout the day.

[0209] To verify the effectiveness of the method of this invention, a virtual power plant simulation control system covering day-ahead, intraday, and real-time scenarios was designed. The system was tested using an IEEE 30-node network, and the Gurobi solver was used for optimized scheduling. Day-ahead scheduling used a one-hour time step, intraday scheduling used a 15-minute time step, and real-time scheduling used a 15-second time step.

[0210] Figure 4 The correlation analysis revealed a complex interdependent structure of resources within the virtual power plant. Wind power and photovoltaic power showed a weak positive correlation (0.124), exhibiting certain synergistic characteristics; wind power and load showed a weak negative correlation (-0.193), while photovoltaic power and load showed a weak negative correlation (-0.271). This result validates the necessity of using the Copula function to characterize the correlation. The correlation coefficient matrix shows a non-linear dependency between the variables, indicating that the variables cannot be simply assumed to be independent and unaffected by each other, laying the foundation for subsequent probability assessment.

[0211] Figure 5 It showcases 500 typical operational scenarios generated based on 365 days of historical data. Among them: according to Figure 5 In (a), the wind power output scenario aggregates the total output of wind power 1 and wind power 2. The scenario average matches historical data well, but due to the uncertainty of wind force, many extreme scenarios also occur. According to Figure 5 In section (b), the photovoltaic power output scenario presents a typical "mountain peak" shape, according to... Figure 5 In scenario (c), the load demand scenario exhibits a "bi-peak" characteristic, with significant uncertainty in the morning and evening peaks, reflecting the effective learning of temporal dependencies by the SAE model. Scenario generation results demonstrate that the CSML model can accurately capture the spatiotemporal coupling relationship between wind, solar, and load while preserving the statistical characteristics of historical data. This forms the basis for scenario generation in the subsequent probabilistic assessment of the virtual power plant's regulation capacity.

[0212] from Figure 6As can be seen from Figs. (a) and (b), under different confidence levels (p50, p75, p90, p95, p99), the upward and downward regulation capabilities of the system exhibit significant spatio-temporal distribution characteristics and confidence level sensitivities. The upward regulation capability curve shows that obvious peaks of regulation demand appear in the 8th - 12th period and the 20th - 24th period, which is closely related to the superposition effect of the uncertainty of new energy output and load fluctuations; while the downward regulation capability curve shows greater regulation pressure in the 0th - 6th period and the 14th - 18th period. As the confidence level decreases from p50 to p99, the regulation capacity that the system can provide shows a significant decreasing trend, indicating that when considering extremely adverse scenarios, the reliable regulation margin of the system will be greatly narrowed. The construction of the typical probability-capacity regulation curve library provides a scientific calculation method for quantifying the system regulation capacity boundary under different confidence levels.

[0213] Figure 7 Figs. (a) and (b) are the decomposition diagrams of the upward regulation capacity and the downward regulation capacity. It can be seen that there are significant differences in the composition of the flexible regulation capabilities of the system at different times. GT1, GT2, and GT3, as base load units, maintain relatively stable contributions at each time period. GT4, GT5, and GT6 have smaller capacities but are more flexible in regulation. They mainly provide additional upward regulation capacity during peak load periods to cope with the fluctuations in renewable energy output and the sharp increase in load demand. The energy storage system shows obvious peak-valley response characteristics. It absorbs excess power during the low valley period at night in terms of downward regulation capacity, and quickly releases electricity during the peak power consumption period in terms of upward regulation capacity. This reflects the response capabilities and response preferences of different energy sources at different times, forming a hierarchical complementary flexible resource allocation pattern.

[0214] In the reliability analysis part, 1000 scenarios were generated using different seeds for analysis. From Figure 8 it can be seen that the confidence level P99 > P95 > P90 > P75 > P50, and the reliability rate P99 < P95 < P90 < P75 < P50. Because the regulation margin corresponding to a high confidence level is smaller, in some extreme scenarios, the power supply reliability will decrease due to insufficient regulation capacity. When actually scheduling, it is necessary to comprehensively consider the power supply reliability and the scheduling reliability to select the scheduling plan.

[0215] Figure 9 Analysis is carried out on the power supply reliability of specific time periods: The average required upward and downward regulation capacities and the available upward and downward regulation capacities under 1000 test scenarios are shown here. It can be seen that the adequacy of the upward regulation capacity decreases significantly at noon because at this time the photovoltaic output increases greatly, and the probability and amplitude of the prediction deviation of the renewable energy part also increase accordingly, resulting in a decrease in the adjustable adequacy. At this time, it is also more likely to cause the optimization to fail due to insufficient regulation capacity. The reason for the increase in the adequacy of the downward regulation capacity at noon is the same. Figure 10 Show the optimization situation of the first 400 scenarios for 24 time periods.

[0216] Analysis of scheduling failure cases:

[0217] Table 1. Scheduling Failure Cases

[0218]

[0219] Table 2. Statistical analysis of structural characteristics of scheduling failures.

[0220]

[0221] Table 1 lists several representative dispatch failure cases and specific reasons for GT1-GT6 and energy storage device ESS1. Table 2, in conjunction with these, helps identify vulnerable devices and vulnerable periods. The statistical results show that ESS1, GT4, GT5, and GT6 have significantly more defaults than GT1, GT2, and GT3, indicating that the flexibility of the adjustable devices is insufficient, constrained by ramp-up and their own output, resulting in limited adjustable capacity.

[0222] The time period and the amount of defaults can be combined. Figure 5 Analysis shows that the default rate for upward capacity adjustments reached as high as 30.0% during periods 0-4, with an average default amount of 38.3 MW. This was mainly due to the unpredictable nature of wind power generation at night. The default rate for periods 10-14 reached 28.6%, caused by random fluctuations due to high photovoltaic output. During periods 16-20, the default rate for upward capacity adjustments was significantly higher than that for downward adjustments due to reduced renewable energy output and increased load demand. In terms of the distribution of default amounts, small fluctuations accounted for a large proportion, which can be significantly improved by optimizing day-ahead and intraday scheduling logic. Large fluctuations and extreme events accounted for a smaller proportion, but require systemic countermeasures such as demand response, energy storage expansion, or grid interconnection. These findings fully demonstrate the necessity of the dynamic regulation pool method proposed in this paper—by considering the ramp-up constraints and regulation demands of future periods through rolling optimization, it can identify risks in vulnerable areas such as periods 10-14 and 16-20 in advance, optimize day-ahead scheduling output arrangements and energy storage charging and discharging strategies, thereby significantly reducing the overall system default rate.

[0223] Example 2

[0224] This embodiment provides a virtual power plant heterogeneous resource regulation capability assessment system, including:

[0225] The scene generation module is used to fit an empirical Copula function using historical data for multi-source uncertain variables, and to generate a time series scene set based on the empirical Copula function and the trained stacked autoencoder model.

[0226] The module for building a multi-confidence level adjustment curve library is used to calculate the adjustment capability of all uncertain scenarios based on the generated time series scenario set, construct an adjustment capability sample set based on the calculated adjustment capability sample set, and construct a multi-confidence level adjustment curve library based on the constructed adjustment capability sample set.

[0227] The capability assessment module is used to embed the multi-confidence level adjustment curve library as a constraint into the intraday rolling scheduling, and combine it with the set power supply reliability criteria to calculate the reliability of power supply, so as to obtain the probabilistic assessment result of power supply reliability.

[0228] It should be noted that the specific implementation of the virtual power plant heterogeneous resource regulation capability assessment method in this embodiment of the invention is similar to the specific implementation of the virtual power plant heterogeneous resource regulation capability assessment method in this embodiment of the invention. For details, please refer to the description in the method section. In order to reduce redundancy, it will not be repeated here.

[0229] Example 3

[0230] This embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps in the above-described method for evaluating the heterogeneous resource regulation capability of a virtual power plant.

[0231] Example 4

[0232] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the steps in the virtual power plant heterogeneous resource regulation capability assessment method described above.

[0233] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage and optical storage) containing computer-usable program code.

[0234] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0235] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0236] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0237] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0238] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for evaluating the heterogeneous resource regulation capacity of a virtual power plant, characterized in that, Includes the following steps: For multi-source uncertain variables, an empirical Copula function is obtained by fitting historical data. Based on the empirical Copula function and the trained stacked autoencoder model, a time series scene set is generated, including: Based on the empirical Copula function, a static scenario is generated at time t that conforms to the historical joint distribution characteristics, including the uncertainties of wind power, photovoltaics, and load. The generated static scene at time t is input into the trained stacked autoencoder network to predict the reference scene at time t+1, and the difference between the reference scene and the sampled scene is calculated. Compare the difference between the reference scenario and the sampled scenario. If the difference meets the set threshold condition, the generated time series scenario is reasonable and is retained. Otherwise, the scenario is resampled using the empirical Copula function until the data meets the requirements, and the final time series scenario set is obtained. Based on the generated time-series scenario set, the adaptability of all uncertain scenarios is calculated, and a adaptability sample set is constructed based on the calculated adaptability of all uncertain scenarios; the calculation of the adaptability of all uncertain scenarios based on the generated time-series scenario set includes: Calculation time Scene The power boundaries of the system include the maximum available power and the minimum required power; According to time Scene Power boundary calculation time of the lower system Scene The upward and downward adjustment capabilities of the lower system; N scenes and times generated by combining time series scene sets Scene The upregulation and downregulation capabilities of the system are used to construct a sample set of the regulation capabilities at time t. Based on the constructed regulation capacity sample set, a multi-confidence level regulation curve library is constructed, including: For the set of samples with adjustment capabilities, construct an empirical cumulative distribution function; The ability to regulate is calculated based on the empirical cumulative distribution function. p quantiles; Based on regulatory capacity p Quantiles define a set of confidence levels, for each confidence level... p Construct upward and downward adjustment curves, and assemble them to obtain a conditional adjustment curve library; By embedding the multi-confidence level adjustment curve library as a constraint into the intraday rolling scheduling, and combining it with the reliability calculation of power supply based on the set power supply reliability criteria, a probabilistic assessment result of power supply reliability is obtained.

2. The method for evaluating the heterogeneous resource regulation capacity of a virtual power plant as described in claim 1, characterized in that, Power supply reliability is calculated based on a multi-confidence level adjustment curve library and established power supply reliability criteria to obtain power supply reliability for various scenarios, including: The total adjustment requirements of the computing system include total upward adjustment requirements and total downward adjustment requirements; Query the system's total regulation capacity limit from the regulation curve library, and determine the capacity adequacy based on the total regulation capacity limit and total regulation demand; The first power supply reliability criterion is determined based on the capacity adequacy to determine whether the system power supply is reliable. If the power supply is reliable, the constructed cost optimization model is solved. The second power supply reliability criterion is used to determine whether the power supply is reliable. If so, the model optimization for this time step in this scenario is successful. The power supply reliability is evaluated for all scenarios, resulting in a probabilistic evaluation of power supply reliability.

3. The method for evaluating the heterogeneous resource regulation capacity of a virtual power plant as described in claim 2, characterized in that, The cost optimization model includes an objective function and constraints. The objective function includes corrected maintenance costs, corrected fuel costs, energy storage degradation deviation, adjustment penalties, and energy storage SOC trajectory tracking deviation adjustment penalties. The constraints include power balance constraints, unit up-up capacity constraints, and unit down-down capacity constraints.

4. The method for evaluating the heterogeneous resource regulation capacity of a virtual power plant as described in claim 2, characterized in that, The first power supply reliability criterion is: ， The second power supply reliability criterion is: ， when and When both values ​​are 1, it indicates that the power supply is reliable; otherwise, the power supply is unreliable. in, As a unit-level criterion indication, It is an indicative variable used to indicate whether there are excessive adjustments at the unit level. Indicates time Scene The capacity adequacy has been adjusted upwards. Indicates time Scene The capacity adequacy of the reduction.

5. A virtual power plant heterogeneous resource regulation capacity assessment system, characterized in that, include: The scene generation module is used to fit an empirical Copula function to historical data for multi-source uncertain variables, and to generate a time-series scene set based on the empirical Copula function and the trained stacked autoencoder model, including: Based on the empirical Copula function, a static scenario is generated at time t that conforms to the historical joint distribution characteristics, including the uncertainties of wind power, photovoltaics, and load. The generated static scene at time t is input into the trained stacked autoencoder network to predict the reference scene at time t+1, and the difference between the reference scene and the sampled scene is calculated. Compare the difference between the reference scenario and the sampled scenario. If the difference meets the set threshold condition, the generated time series scenario is reasonable and is retained. Otherwise, the scenario is resampled using the empirical Copula function until the data meets the requirements, and the final time series scenario set is obtained. A multi-confidence level adjustment curve library construction module is used to calculate the adjustment capability of all uncertainty scenarios based on a generated time series scenario set, and to construct an adjustment capability sample set based on the calculated adjustment capabilities of all uncertainty scenarios; the calculation of the adjustment capability of all uncertainty scenarios based on the generated time series scenario set includes: Calculation time Scene The power boundaries of the system include the maximum available power and the minimum required power; According to time Scene Power boundary calculation time of the lower system Scene The upward and downward adjustment capabilities of the lower system; N scenes and times generated by combining time series scene sets Scene The upregulation and downregulation capabilities of the system are used to construct a sample set of the regulation capabilities at time t. Based on the constructed regulation capacity sample set, a multi-confidence level regulation curve library is constructed, including: For the set of samples with adjustment capabilities, construct an empirical cumulative distribution function; The ability to regulate is calculated based on the empirical cumulative distribution function. p quantiles; Based on regulatory capacity p Quantiles define a set of confidence levels, for each confidence level... p Construct upward and downward adjustment curves, and assemble them to obtain a conditional adjustment curve library; The capability assessment module is used to embed the multi-confidence level adjustment curve library as a constraint into the intraday rolling scheduling, and combine it with the set power supply reliability criteria to calculate the reliability of power supply, so as to obtain the probabilistic assessment result of power supply reliability.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps in the method for evaluating the heterogeneous resource regulation capability of a virtual power plant as described in any one of claims 1-4.

7. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps in the method for evaluating the heterogeneous resource regulation capability of a virtual power plant as described in any one of claims 1-4.

Images