Virtual power plant cluster collaborative regulation method and system based on dynamic aggregation

Abstract

The application discloses a virtual power plant cluster cooperative regulation method and system based on dynamic aggregation, belongs to the technical field of virtual power plant resource regulation, and through double analysis on distributed resource basic characteristics and response characteristics, accurately captures the heterogeneity difference of the distributed resources, constructs a dynamic feature library to homogenize and classify the heterogeneous resources, improves the adaptability of resource aggregation, avoids the limitation that static aggregation is difficult to cope with resource dynamic changes, covers the fluctuation scene of the resources through dynamic aggregation, enhances the robustness of the virtual power plant cluster in response to uncertainty, thereby reduces the scheduling deviation, takes the optimal comprehensive benefit as the target, improves the new energy consumption rate, realizes the cooperative optimization of cluster operation economy, safety and low carbon, and reduces the comprehensive operation cost of the system.

Description

Technical Field

[0001] This invention relates to the field of virtual power plant resource regulation technology, specifically to a method and system for collaborative regulation of virtual power plant clusters based on dynamic aggregation. Background Technology

[0002] Guided by dual-carbon goals, distributed resources such as wind power, photovoltaics, electric vehicles, and flexible loads are being connected to the grid on a large scale, driving the power system's transformation into a new type of power system. However, distributed resources across a wide area exhibit significant heterogeneity. Wind / photovoltaic output is intermittent and seasonally fluctuating, electric vehicle charging behavior is constrained by travel patterns, and the adjustability potential, response speed, and operational constraints of industrial and commercial flexible loads differ significantly. This leads to difficulties in resource aggregation and low efficiency in coordinated regulation. Traditional virtual power plants mostly target single-type resources and lack precise characterization and classification adaptation for heterogeneous resources. Furthermore, static aggregation models struggle to cope with dynamic grid connection / off-grid changes and are insufficient in handling uncertainties such as wind and solar output fluctuations and random load fluctuations, easily resulting in large dispatch deviations, low renewable energy absorption rates, and high system operating costs.

[0003] Chinese Patent Publication No. CN117614030A discloses a hierarchical collaborative scheduling method and apparatus for virtual power plants. In response to a received collaborative scheduling request, a corresponding target master node is selected from the virtual power plant system, and the collaborative scheduling request is sent to slave nodes through the target master node. Scheduling analysis data corresponding to the collaborative scheduling request is extracted from the slave nodes. Based on the scheduling analysis data and a preset optimization model, corresponding scheduling analysis parameters are generated and sent to the target master node. The target master node then performs collaborative processing on the scheduling analysis parameters to generate a corresponding collaborative scheduling scheme. However, due to the differences in distributed resources over a wide area, resource scheduling suffers from high aggregation difficulty and low collaborative control efficiency. Summary of the Invention

[0004] This invention addresses the problems of large resource scheduling deviations, low renewable energy absorption rates, and high system operating costs in existing virtual power plant clusters. It provides a collaborative control method and system for virtual power plant clusters based on dynamic aggregation. Through dual analysis of the basic characteristics and response features of distributed resources, it accurately captures the heterogeneity differences of distributed resources. Simultaneously, it constructs a dynamic feature library to homogenize heterogeneous resources, improving the adaptability of resource aggregation and avoiding the limitations of static aggregation in dealing with dynamic resource changes. Furthermore, dynamic aggregation covers resource fluctuation scenarios, enhancing the robustness of the virtual power plant cluster in the face of uncertainty, thereby reducing scheduling deviations. With the goal of optimal comprehensive benefits, it improves the renewable energy absorption rate, achieving synergistic optimization of cluster operation economy, safety, and low-carbon performance, and reducing the overall system operating cost.

[0005] In a first aspect, one technical solution provided in this embodiment of the invention is: a method for collaborative control of a virtual power plant cluster based on dynamic aggregation, comprising the following steps: S1. Perform basic characteristic analysis on the distributed resources of the power plant to obtain heterogeneity characteristics; perform response analysis on the distributed resources to obtain resource response characteristics; S2. Based on the real-time operation data of user-side distributed resources, the heterogeneity characteristics and resource response characteristics are quantified to obtain a dynamic feature library. The features in the dynamic feature library are classified to obtain resource clusters. S3. Perform uncertainty analysis on various resource clusters to obtain fluctuation scenarios, and aggregate resources in the fluctuation scenarios based on a dynamic selection mechanism to obtain virtual power plant clusters. S4. Construct a corresponding multi-resource coupling model for each virtual power plant cluster, build an objective function with the goal of maximizing the overall operating efficiency, and solve the objective function under the constraints of the multi-resource coupling model to obtain the collaborative control strategy of the virtual power plant cluster.

[0006] This solution accurately captures the heterogeneous differences of wide-area resources such as wind power, photovoltaics, and flexible loads through dual analysis of the basic characteristics and response features of distributed resources, providing data support for subsequent targeted aggregation and scheduling, and solving the problem of vague characterization of resource characteristics in traditional methods. By constructing a dynamic feature library based on real-time operational data and dividing resource clusters, it achieves homogeneous classification of heterogeneous resources, thereby improving the adaptability of resource aggregation and avoiding the limitations of static aggregation in dealing with dynamic changes in resources. Through uncertainty analysis of resource clusters and dynamic aggregation mechanisms, it comprehensively covers scenarios such as fluctuations in wind and solar power output and random changes in load, thereby enhancing the robustness of virtual power plant clusters in dealing with uncertainties and reducing scheduling deviations. By constructing a multi-resource coupling model and an optimal objective function for comprehensive benefits, it fully explores the potential for multi-energy complementarity of electricity, heat, and cooling, thereby improving the renewable energy absorption rate, achieving synergistic optimization of the economic efficiency, safety, and low carbon emissions of cluster operation, and reducing the overall operating cost of the system.

[0007] Optionally, in S1, a basic characteristic analysis of the power plant's distributed resources is performed to obtain heterogeneity characteristics, including the following steps: The distributed resources include distributed power sources, energy storage resources, and flexible loads; the distributed power sources include wind power output and photovoltaic power output. Analysis of distributed power generation output yields the periodic characteristics of wind power output, the diurnal characteristics of photovoltaic power output, seasonal fluctuations, and climate-sensitive fluctuations. The load characteristics and response potential of energy storage resources are obtained by analyzing the load. The scheduling potential is obtained by analyzing the load adjustability time, adjustable load ratio, and response speed of flexible loads. The cyclical characteristics of wind power output, the diurnal characteristics of photovoltaic power output, seasonal fluctuations, climate-sensitive fluctuations, load characteristics, response potential, and dispatch potential are considered as heterogeneous characteristics.

[0008] This solution analyzes the key characteristics of each core resource by subdividing it into three categories: distributed power sources, energy storage resources, and flexible loads. This achieves comprehensive coverage of heterogeneous characteristics and avoids the limitations of single-dimensional analysis. By accurately capturing the power characteristics of wind power cycles, photovoltaic day-night cycles, and climate-sensitive fluctuations, the solution clarifies the characteristics and response potential of energy storage loads and the scheduling potential of flexible loads, providing a precise basis for subsequent resource classification and aggregation.

[0009] Optionally, in S1, response analysis of distributed resources is performed to obtain resource response characteristics, including the following steps: The influence of wind speed on wind power output is quantified to obtain wind power response characteristics, and the influence of sunlight and temperature on photovoltaic power output is quantified to obtain photovoltaic response characteristics. The energy storage response characteristics are obtained by quantifying the changes in charging and discharging power and capacity of energy storage resources; The load response characteristics are obtained by quantifying the load changes of flexible loads.

[0010] In this scheme, by quantifying the impact of wind speed on wind power, and the impact of sunlight and temperature on photovoltaics, the response law of distributed power sources is accurately captured, solving the problem of ambiguity in the correlation between environmental factors and output in traditional analysis. By quantifying the charging and discharging power, capacity changes and flexible load fluctuations of energy storage, the response boundaries and control potential of various resources are clarified, providing a quantitative basis for the construction of subsequent response models.

[0011] Optionally, in S2, a dynamic feature library is obtained by quantifying heterogeneity characteristics and resource response characteristics based on real-time operational data of user-side distributed resources, including the following steps: Based on the time scale, heterogeneous features and resource response features are initially divided into intuitive descriptive features and ratio descriptive features; Using real-time operational data from user-side distributed resources as the data source, and assigning intuitive descriptive features and ratio descriptive features to mathematical calculation formulas, various numerical features are obtained through calculation. A dynamic feature library is obtained by organizing various numerical features and updating them in real time.

[0012] This solution constructs a standardized feature system by dividing intuitive descriptive features into intuitive descriptive features and ratio descriptive features according to time scale, avoiding the confusion of multi-dimensional features and providing a clear framework for subsequent quantitative calculations and resource classification. By using real-time operational data as the data source and combining mathematical formulas to quantify features, qualitative heterogeneity and response features are transformed into quantitative values, solving the problem of fuzzy characterization in traditional features and thus improving feature accuracy. By updating the feature library in real time, it dynamically adapts to changes such as distributed resource network entry / exit and output fluctuations, ensuring the timeliness of feature data and providing reliable support for subsequent resource cluster partitioning and uncertainty analysis, thus ensuring the adaptability and robustness of aggregated scheduling.

[0013] Optionally, in S3, the resource cluster includes a new energy cluster, a load cluster, and a multi-energy coupling cluster.

[0014] In this scheme, by dividing new energy clusters, distributed power sources such as wind power and photovoltaics, which are intermittent and fluctuating, can be centrally managed, thereby enabling the development of targeted strategies to cope with uncertainties and improving the renewable energy absorption rate. By dividing load clusters, the scheme can accurately focus on the differences in the scheduling potential of flexible loads, thereby adapting to differentiated scheduling needs such as peak shaving and valley filling, and maximizing the value of load-side regulation. By dividing multi-energy coupling clusters, the scheme can integrate multiple energy resources such as electricity, heat, and cooling with coupling equipment, giving full play to the advantages of multi-energy complementarity, providing a clear carrier for subsequent construction of coupling models and optimization of comprehensive benefits, and improving the economy and flexibility of cluster operation.

[0015] Optionally, in S3, uncertainty analysis is performed on various resource clusters to obtain fluctuation scenarios, including the following steps: Latin hypercube sampling method is used to sample the wind and solar power output fluctuations in various resource clusters to generate the original power output scenarios. The Monte Carlo simulation method is used to simulate load fluctuations in various resource clusters to generate original load scenarios; Time-series sampling of historical electricity prices in various resource clusters generates the original electricity price scenario for a time period. The original power output scenario, original load scenario, and original electricity price scenario are reduced and scenario weights are assigned by using the probabilistic distance method to obtain the fluctuation scenario.

[0016] This solution employs differentiated methods—Latin hypercube sampling, Monte Carlo simulation, and time-series sampling—for wind and solar power output, load, and electricity price, respectively. This approach accurately adapts to the fluctuation characteristics of various resources, comprehensively covers sources of uncertainty, and solves the problem that traditional single sampling cannot simultaneously ensure accuracy across multiple scenarios. By reducing scenarios and assigning weights using the probabilistic distance method, the solution retains key fluctuation scenarios (such as extreme power output and electricity price scenarios) while reducing computational complexity, thus providing efficient risk input for subsequent dynamic aggregation.

[0017] Optionally, in S3, a virtual power plant cluster is obtained by aggregating resources for fluctuating scenarios based on a dynamic selection mechanism, including the following steps: In fluctuating scenarios, the aggregation characteristics of resource clusters are calculated by dynamically matching the corresponding aggregation methods to the resource clusters. Associating functional tags with resource clusters based on their aggregation characteristics; Physical boundaries are defined based on the geographical location of resource clusters, and functional boundaries are defined based on the functional labels of resource clusters. The sub-clusters formed by the intersection of physical boundaries and functional boundaries are used as virtual power plant clusters.

[0018] In this solution, by dynamically matching aggregation methods under fluctuating scenarios, the heterogeneity and fluctuation characteristics of new energy, load and multi-energy coupled clusters are adapted to accurately calculate aggregation characteristics, avoiding the limitations of a single aggregation method. By associating functional tags with resource clusters, the scheduling positioning of each cluster such as peak shaving, valley filling and backup is clarified, providing a clear basis for subsequent coordinated control, thereby maximizing the functional value of different clusters.

[0019] Optionally, the aggregation method includes direct aggregation, Monte Carlo simulation, dynamic programming, and geometric computation. The aggregation characteristics include cluster response time, resource adjustment range, and cluster adjustable capacity limit; The functional labels include peak shaving cluster, valley filling cluster, and backup cluster.

[0020] This solution provides a multi-faceted aggregation method to accurately adapt to the heterogeneity and fluctuation characteristics of new energy sources, loads, and multi-energy coupled clusters, avoiding the limitations of a single method and improving aggregation suitability. By clearly defining labels such as peak shaving, valley filling, and backup functions, the cluster can accurately match differentiated scheduling needs, thereby maximizing the control value of heterogeneous resources and improving the control effect.

[0021] Optionally, in S4, a corresponding multi-resource coupling model is constructed for each virtual power plant cluster. An objective function is built with the goal of maximizing overall operational efficiency. The objective function is solved under the constraints of the multi-resource coupling model to obtain the collaborative control method for the virtual power plant cluster, including the following steps: An objective function is constructed with the sum of the costs of purchasing natural gas, energy storage, unit maintenance, and charging / discharging compensation for the virtual power plant cluster as the objective. A two-stage robust optimization method is used to solve the objective function to obtain the distributed output, load curves, and optimal electricity prices for each time period of each virtual power plant cluster; The distributed output, load curves, and optimal electricity prices for each time period of each virtual power plant cluster are used as the collaborative control strategy for virtual power plant clusters.

[0022] This solution constructs a full-cost objective function covering gas purchase, energy storage, maintenance, and charge / discharge compensation, precisely anchoring the core objective of maximizing overall operational efficiency, thus ensuring that the control strategy balances economy and practicality. By employing a two-stage robust optimization method, it effectively addresses uncertainties such as wind and solar power output and load fluctuations, thereby enhancing the anti-interference capability and robustness of the dispatch strategy. By outputting cluster distributed power output, load curves, and optimal electricity prices, it forms a feasible collaborative control strategy, supporting the precise execution of dispatch instructions by the virtual power plant cluster and maximizing the value of multi-energy coupling.

[0023] Secondly, one technical solution provided in this embodiment of the invention is: a virtual power plant cluster collaborative control system based on dynamic aggregation, including a data processing module, a feature library generation module, a virtual power plant generation module, and a collaborative control module; The data processing module performs basic characteristic analysis on the power plant's distributed resources to obtain heterogeneity characteristics, and performs response analysis on the distributed resources to obtain resource response characteristics. The feature library generation module quantifies heterogeneous features and resource response features based on real-time operational data of user-side distributed resources to obtain a dynamic feature library, and classifies the features in the dynamic feature library to obtain resource clusters. The virtual power plant generation module performs uncertainty analysis on various resource clusters to obtain fluctuating scenarios, and aggregates resources from the fluctuating scenarios based on a dynamic selection mechanism to obtain virtual power plant clusters. The collaborative control module constructs a corresponding multi-resource coupling model for each virtual power plant cluster, builds an objective function with the goal of maximizing comprehensive operational efficiency, and solves the objective function under the constraints of the multi-resource coupling model to obtain the collaborative control method for virtual power plant clusters.

[0024] In this solution, a corresponding system is built to implement the virtual power plant cluster collaborative control method, thereby achieving human-computer interaction and improving the user experience.

[0025] The beneficial effects of this invention are as follows: By analyzing both the basic characteristics and response features of distributed resources, this invention accurately captures the heterogeneity differences of distributed resources. At the same time, it constructs a dynamic feature library to classify heterogeneous resources in a homogeneous way, improving the adaptability of resource aggregation and avoiding the limitations of static aggregation in dealing with dynamic changes in resources. Furthermore, dynamic aggregation covers resource fluctuation scenarios, enhancing the robustness of the virtual power plant cluster in dealing with uncertainties, thereby reducing scheduling deviations. With the goal of optimal comprehensive benefits, it improves the renewable energy consumption rate, achieves synergistic optimization of the cluster's economic efficiency, safety, and low carbon emissions, and reduces the overall operating cost of the system.

[0026] The above description of the invention is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0027] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings. The drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings.

[0028] Figure 1 This is a flowchart of the collaborative control method for virtual power plant clusters based on dynamic aggregation, as described in this invention. Figure 2 This is a schematic diagram of the day-ahead stage power supply and demand scheduling results of the virtual power plant in this embodiment; Figure 3 This embodiment presents a schematic diagram of the day-ahead stage virtual power plant thermal energy supply and demand scheduling results. Figure 4 This is a schematic diagram of the virtual power plant cluster collaborative control system based on dynamic aggregation according to the present invention. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only one preferred embodiment of this invention and are only used to explain this invention. They do not limit the scope of protection of this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0030] Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of the operations (or steps) can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it may also have additional steps not included in the figures; the process may correspond to a method, function, procedure, subroutine, subroutine, etc.

[0031] Example 1: To address the issues of large resource scheduling deviations, low renewable energy absorption rates, and high system operating costs in existing virtual power plant clusters, this example provides a collaborative control method for virtual power plant clusters based on dynamic aggregation, such as... Figure 1 As shown, it includes the following steps: S1: Perform basic characteristic analysis on the distributed resources of the power plant to obtain heterogeneity characteristics; perform response analysis on the distributed resources to obtain resource response characteristics.

[0032] In this embodiment, the heterogeneity characteristics of the distributed resources of the power plant are obtained through basic characteristic analysis, including the following steps: The distributed resources include distributed power sources, energy storage resources, and flexible loads; the distributed power sources include wind power output and photovoltaic power output. Analysis of distributed power generation output yields the periodic characteristics of wind power output, the diurnal characteristics of photovoltaic power output, seasonal fluctuations, and climate-sensitive fluctuations. The load characteristics and response potential of energy storage resources are obtained by analyzing the load. The scheduling potential is obtained by analyzing the load adjustability time, adjustable load ratio, and response speed of flexible loads. The cyclical characteristics of wind power output, the diurnal characteristics of photovoltaic power output, seasonal fluctuations, climate-sensitive fluctuations, load characteristics, response potential, and dispatch potential are considered as heterogeneous characteristics.

[0033] This embodiment subdivides core resources into three categories: distributed power sources, energy storage resources, and flexible loads, and analyzes the key characteristics of each resource in a targeted manner, thereby achieving comprehensive coverage of heterogeneous characteristics and avoiding the limitations of single-dimensional analysis. By accurately capturing power characteristics such as wind power cycle characteristics, photovoltaic day-night cycles, and climate-sensitive fluctuations, it clarifies the characteristics and response potential of energy storage loads and the scheduling potential of flexible loads, providing a precise basis for subsequent resource classification and aggregation.

[0034] In this embodiment, the response analysis of distributed resources to obtain resource response characteristics includes the following steps: The influence of wind speed on wind power output is quantified to obtain wind power response characteristics, and the influence of sunlight and temperature on photovoltaic power output is quantified to obtain photovoltaic response characteristics. The energy storage response characteristics are obtained by quantifying the changes in charging and discharging power and capacity of energy storage resources; The load response characteristics are obtained by quantifying the load changes of flexible loads.

[0035] This embodiment accurately captures the response patterns of distributed power sources by quantifying the impact of wind speed on wind power, and the impact of sunlight and temperature on photovoltaics, thus solving the problem of ambiguity in the correlation between environmental factors and output in traditional analysis. By quantifying the charging and discharging power, capacity changes, and flexible load fluctuations of energy storage, it clarifies the response boundaries and control potential of various resources, providing a quantitative basis for the construction of subsequent response models.

[0036] S2: Based on the real-time operation data of user-side distributed resources, the heterogeneity characteristics and resource response characteristics are quantified to obtain a dynamic feature library. The features in the dynamic feature library are then classified to obtain resource clusters.

[0037] In this embodiment, a dynamic feature library is obtained by quantifying heterogeneity characteristics and resource response characteristics based on real-time operational data of user-side distributed resources, including the following steps: Based on the time scale, heterogeneous features and resource response features are initially divided into intuitive descriptive features and ratio descriptive features; Using real-time operational data from user-side distributed resources as the data source, and assigning intuitive descriptive features and ratio descriptive features to mathematical calculation formulas, various numerical features are obtained through calculation. A dynamic feature library is obtained by organizing various numerical features and updating them in real time.

[0038] Specifically, intuitive descriptive features focus on the resource regulation behavior itself, directly reflecting the regulation potential, including regulation direction, response time, duration, ramp rate, regulation amplitude (upward / downward), response rate, response duration, response capacity, etc.; ratio descriptive features focus on resource operation efficiency and patterns, quantifying long-term characteristics, including response capability, load shifting capability, daily electricity consumption, daily maximum / minimum / average load, daily peak-to-valley difference, valley electricity coefficient, daily load factor, peak-to-valley difference rate, peak / normal / valley electricity consumption rate, weekend electricity consumption coefficient, load standard deviation, etc.

[0039] This embodiment constructs a standardized feature system by dividing intuitive descriptive features into intuitive descriptive features and ratio descriptive features according to the time scale, avoiding the confusion of multi-dimensional features and providing a clear framework for subsequent quantitative calculations and resource classification. By using real-time running data as the data source and combining mathematical formulas to quantify features, qualitative heterogeneity and response features are transformed into quantitative values, solving the problem of vague characterization in traditional features and thus improving feature accuracy. By updating the feature library in real time, it dynamically adapts to changes such as distributed resource network entry / exit and output fluctuations, ensuring the timeliness of feature data and providing reliable support for subsequent resource cluster partitioning and uncertainty analysis, thus ensuring the adaptability and robustness of aggregation scheduling.

[0040] S3: Perform uncertainty analysis on various resource clusters to obtain fluctuating scenarios, and aggregate resources in the fluctuating scenarios based on a dynamic selection mechanism to obtain virtual power plant clusters.

[0041] In this embodiment, the resource cluster includes a new energy cluster, a load cluster, and a multi-energy coupling cluster.

[0042] This embodiment, by dividing new energy clusters, can centrally manage distributed power sources such as wind power and photovoltaics, which are intermittent and fluctuating, thereby formulating targeted strategies to cope with uncertainties and improving the new energy absorption rate. By dividing load clusters, it can accurately focus on the differences in the scheduling potential of flexible loads, thereby adapting to differentiated scheduling needs such as peak shaving and valley filling, and maximizing the value of load-side regulation. By dividing multi-energy coupling clusters, it can integrate multiple energy resources such as electricity, heat, and cooling with coupling equipment, giving full play to the advantages of multi-energy complementarity, providing a clear carrier for subsequent construction of coupling models and optimization of comprehensive benefits, and improving the economy and flexibility of cluster operation.

[0043] In this embodiment, uncertainty analysis is performed on various resource clusters to obtain fluctuation scenarios, including the following steps: Latin hypercube sampling method is used to sample the wind and solar power output fluctuations in various resource clusters to generate the original power output scenarios. The Monte Carlo simulation method is used to simulate load fluctuations in various resource clusters to generate original load scenarios; Time-series sampling of historical electricity prices in various resource clusters generates the original electricity price scenario for a time period. The original power output scenario, original load scenario, and original electricity price scenario are reduced and scenario weights are assigned by using the probabilistic distance method to obtain the fluctuation scenario.

[0044] This solution employs differentiated methods—Latin hypercube sampling, Monte Carlo simulation, and time-series sampling—for wind and solar power output, load, and electricity price, respectively. This approach accurately adapts to the fluctuation characteristics of various resources, comprehensively covers sources of uncertainty, and solves the problem that traditional single sampling cannot simultaneously ensure accuracy across multiple scenarios. By reducing scenarios and assigning weights using the probabilistic distance method, the solution retains key fluctuation scenarios (such as extreme power output and electricity price scenarios) while reducing computational complexity, thus providing efficient risk input for subsequent dynamic aggregation.

[0045] In this embodiment, a virtual power plant cluster is obtained by aggregating resources in fluctuating scenarios based on a dynamic selection mechanism, including the following steps: In fluctuating scenarios, the aggregation characteristics of resource clusters are calculated by dynamically matching the corresponding aggregation methods to the resource clusters. Associating functional tags with resource clusters based on their aggregation characteristics; Physical boundaries are defined based on the geographical location of resource clusters, and functional boundaries are defined based on the functional labels of resource clusters. The sub-clusters formed by the intersection of physical boundaries and functional boundaries are used as virtual power plant clusters.

[0046] Specifically, Monte Carlo simulation is used for distributed resource clusters, geometric calculation method is used for load clusters, and dynamic programming or linear programming solution method is used for energy storage clusters.

[0047] This embodiment uses a dynamic matching aggregation method in fluctuating scenarios to adapt to the heterogeneity and fluctuation characteristics of new energy, load and multi-energy coupled clusters, accurately calculates aggregation characteristics, and avoids the limitations of a single aggregation method. By associating functional tags with resource clusters, the scheduling positioning of each cluster such as peak shaving, valley filling and backup is clarified, providing a clear basis for subsequent coordinated control, thereby maximizing the functional value of different clusters.

[0048] In this embodiment, the aggregation method includes direct aggregation, Monte Carlo simulation, dynamic programming, and geometric calculation. The aggregation characteristics include cluster response time, resource adjustment range, and cluster adjustable capacity limit; The functional labels include peak shaving cluster, valley filling cluster, and backup cluster.

[0049] Specifically, the cluster response time is the maximum response time of each resource in the corresponding resource cluster, the cluster adjustment range is the sum of the adjustment ranges of each resource in the corresponding resource cluster, and the cluster adjustable capacity limit is obtained by superimposing the flexibility area using a geometric calculation method. The aggregation characteristics generally also include the expected available capacity and the probability of insufficient capacity obtained by statistically analyzing the cluster output distribution in multiple scenarios using the Monte Carlo simulation method.

[0050] This embodiment provides a multi-element aggregation method to accurately adapt to the heterogeneity and fluctuation characteristics of new energy sources, loads, and multi-energy coupled clusters, avoiding the limitations of a single method and improving the aggregation suitability. By clearly defining labels such as peak shaving, valley filling, and backup functions, the cluster can accurately match differentiated scheduling needs, thereby maximizing the control value of heterogeneous resources and improving the control effect.

[0051] S4: Construct a corresponding multi-resource coupling model for each virtual power plant cluster, build an objective function with the goal of maximizing the overall operating efficiency, and solve the objective function under the constraints of the multi-resource coupling model to obtain the collaborative control strategy of the virtual power plant cluster.

[0052] In this embodiment, a corresponding multi-resource coupling model is constructed for each virtual power plant cluster. An objective function is constructed with the goal of maximizing overall operational efficiency. The objective function is solved under the constraints of the multi-resource coupling model to obtain the collaborative control method for virtual power plant clusters, including the following steps: An objective function is constructed with the sum of the costs of purchasing natural gas, energy storage, unit maintenance, and charging / discharging compensation for the virtual power plant cluster as the objective. A two-stage robust optimization method is used to solve the objective function to obtain the distributed output, load curves, and optimal electricity prices for each time period of each virtual power plant cluster; The distributed output, load curves, and optimal electricity prices for each time period of each virtual power plant cluster are used as the collaborative control strategy for virtual power plant clusters.

[0053] Specifically, the objective function is expressed as follows: in, and These represent the costs of purchasing natural gas, charging and discharging compensation, unit maintenance, electrical energy storage, and thermal energy storage for the virtual power plant during time period t, respectively. This represents the electricity sales revenue of the virtual power plant during time period t.

[0054] The specific formula for calculating the cost of purchasing natural gas is as follows: in, This refers to the unit price of natural gas. Let t be the electrical power output of the gas turbine. For gas turbine power generation efficiency, Let be the electrical power output of the i-th gas turbine at time t.

[0055] The specific formula for calculating the charging and discharging compensation cost is as follows. This embodiment takes the charging and discharging of an electric vehicle as an example: in, Let $\frac{i}{i}$ be the battery purchase cost for the $i$-th electric vehicle. Let be the number of charge-discharge cycles during the lifespan of the battery of the i-th electric vehicle. Let be the battery capacity of the i-th electric vehicle. Let be the depth of battery discharge of the i-th electric vehicle.

[0056] The specific formula for calculating unit maintenance costs is as follows: in, and These represent the total number of waste heat recovery devices and electric boilers, respectively. and These are the maintenance cost coefficients for the i-th gas turbine, waste heat recovery unit, electric boiler, photovoltaic system, and wind turbine, respectively. and Let t be the thermal power output of the i-th waste heat recovery unit and electric boiler. and These are the power outputs of the i-th photovoltaic and wind turbine units at time t, respectively.

[0057] Energy storage costs include electrical energy storage costs and thermal energy storage costs, and the specific calculation formulas are as follows: in The unit price for energy storage maintenance. Let be the charging and discharging power of the i-th energy storage unit during time period t, respectively. The unit price for thermal energy storage maintenance. Let be the heat storage and heat release power of the i-th thermal energy storage unit during time period t, respectively. These represent the total amount of electrical energy storage and thermal energy storage, respectively.

[0058] The specific formula for calculating the revenue from electricity sales of a virtual power plant is as follows: in, This represents the revenue from electricity sales by the virtual power plant during time period t. These represent the power sold to and purchased by the virtual power plant from the grid at time t, respectively. These refer to the electricity sales price from the virtual power plant to the grid and the electricity purchase price.

[0059] The calculation process of solving the objective function using the two-stage robust optimization method is expressed by the following formula: in, Let be the output decision variables for each piece of equipment in the virtual power plant at time t. The actual power output of photovoltaic and wind turbine units, respectively. Fines for renewable energy output falling short of forecast power. It is the first penalty coefficient. , The difference between the predicted power output and the actual output of renewable energy. Fines for renewable energy sources whose actual output exceeds forecast power. It is the second penalty coefficient. , This is the difference between the actual output and the predicted power of renewable energy sources. These are the projected power output values ​​for photovoltaic and wind turbine units for the current period. For photovoltaic prediction bias, a margin of error of ±10% is typically taken. To account for wind power forecasting error, a margin of error of ±15% is typically taken. To avoid overly conservative decision-making, robustness coefficients for both solar and wind turbine units are introduced. and .

[0060] Ultimately, a coordinated control strategy for virtual power plant clusters is derived. For example, the basic output curves of wind power / photovoltaic equipment in each virtual power plant's new energy cluster are formulated in advance; control plans for peak shaving / valley filling clusters of load clusters are formulated, such as reducing load in peak shaving clusters from 17:00 to 21:00 and increasing electricity consumption in valley filling clusters from 00:00 to 06:00; and coordinated output plans and power purchase and sale plans for gas turbines, chillers, electric boilers, etc. in multi-energy coupled clusters are formulated.

[0061] Specifically, taking a certain region as an example, where the natural gas price is 247.43 yuan / MWh, the output of each device in the virtual power plant is obtained after solving the objective function using a two-stage robust optimization method, such as... Figure 2 and Figure 3 As shown, Figure 2 This represents the day-ahead power supply and demand dispatch results of the virtual power plant. PV stands for photovoltaic (PV) generators, WT for wind turbines, GT for gas turbines, Exchange with Grid for power exchange with the grid, EV for electric vehicles, ESS for energy storage devices, EB for electric boilers, and EL for electrical load. Because gas turbines participate in power generation and waste heat recovery, their overall energy utilization rate is high, and their cost is lower than the electricity price during most periods. Gas turbines account for a large proportion of output in all periods, while electric boilers increase their power output during peak heat load periods. Through day-ahead output optimization of the virtual power plant, the amount of electricity supplied by the virtual power plant during off-peak periods decreases, while the amount supplied during peak-price periods increases, achieving the economic benefits of coordinated dispatch.

[0062] Figure 3 This represents the current-day virtual power plant heat supply and demand scheduling results. WHR represents the gas turbine waste heat recovery equipment, HES represents the thermal storage equipment, and HL represents the heat load. During periods of low heat load, the electric boiler operates at minimum power, the gas turbine waste heat recovery equipment meets most of the heat load demand, and the thermal storage equipment stores heat energy for use during peak periods. During periods of peak heat load, the gas turbine waste heat recovery equipment, electric boiler, and thermal storage equipment work together to supply energy to meet the heat load demand.

[0063] This embodiment constructs a full-cost objective function covering gas purchase, energy storage, maintenance, and charge / discharge compensation, accurately anchoring the core objective of maximizing overall operational efficiency, thus ensuring that the control strategy balances economy and practicality. By employing a two-stage robust optimization method, it effectively addresses uncertainties such as wind and solar power output and load fluctuations, thereby enhancing the anti-interference capability and robustness of the dispatch strategy. By outputting cluster distributed power output, load curves, and optimal electricity prices, it forms a feasible collaborative control strategy, supporting the precise execution of dispatch instructions by the virtual power plant cluster and maximizing the value of multi-energy coupling.

[0064] Example 2: This example also provides a virtual power plant cluster collaborative control system based on dynamic aggregation, such as... Figure 4 As shown, it includes a data processing module, a feature library generation module, a virtual power plant generation module, and a collaborative control module; The data processing module performs basic characteristic analysis on the power plant's distributed resources to obtain heterogeneity characteristics, and performs response analysis on the distributed resources to obtain resource response characteristics. The feature library generation module quantifies heterogeneous features and resource response features based on real-time operational data of user-side distributed resources to obtain a dynamic feature library, and classifies the features in the dynamic feature library to obtain resource clusters. The virtual power plant generation module performs uncertainty analysis on various resource clusters to obtain fluctuating scenarios, and aggregates resources from the fluctuating scenarios based on a dynamic selection mechanism to obtain virtual power plant clusters. The collaborative control module constructs a corresponding multi-resource coupling model for each virtual power plant cluster, builds an objective function with the goal of maximizing comprehensive operational efficiency, and solves the objective function under the constraints of the multi-resource coupling model to obtain the collaborative control method for virtual power plant clusters.

[0065] This embodiment implements the virtual power plant cluster collaborative control method in this solution by constructing a corresponding system, thereby realizing human-computer interaction and improving the user experience.

[0066] As can be seen from the above embodiments, it has at least the following substantial effects: (1) This invention accurately captures the heterogeneous differences of wide-area resources such as wind power, photovoltaic, and flexible loads through dual analysis of the basic characteristics and response features of distributed resources, providing data support for subsequent targeted aggregation scheduling and solving the problem of vague characterization of resource characteristics by traditional methods. (2) This invention achieves homogenization of heterogeneous resources by constructing a dynamic feature library based on real-time running data and dividing resource clusters, thereby improving the adaptability of resource aggregation and avoiding the limitations of static aggregation in dealing with dynamic changes in resources. (3) This invention comprehensively covers scenarios such as wind and solar power output fluctuations and random load changes by analyzing the uncertainty of resource clusters and using dynamic aggregation mechanisms, thereby enhancing the robustness of virtual power plant clusters in dealing with uncertainty and reducing scheduling deviations. (4) By constructing a multi-resource coupling model and an optimal objective function for comprehensive benefits, this invention fully explores the potential for multi-energy complementarity of electricity, heat and cold, thereby improving the new energy consumption rate, realizing the synergistic optimization of the economy, safety and low carbon of cluster operation, and reducing the overall operating cost of the system.

[0067] The specific embodiments described above are preferred embodiments of the present invention and are not intended to limit the specific scope of the present invention. The scope of the present invention includes, but is not limited to, these specific embodiments. All equivalent changes made in accordance with the shape and structure of the present invention are within the protection scope of the present invention.

Claims

1. A method for coordinated control of virtual power plant clusters based on dynamic aggregation, characterized in that: Includes the following steps: S1. The heterogeneity characteristics of the distributed resources of the power plant are obtained by performing basic characteristic analysis. Response analysis of distributed resources yields resource response characteristics; S2. Based on the real-time operation data of user-side distributed resources, the heterogeneity characteristics and resource response characteristics are quantified to obtain a dynamic feature library. The features in the dynamic feature library are classified to obtain resource clusters. S3. Perform uncertainty analysis on various resource clusters to obtain fluctuation scenarios, and aggregate resources in the fluctuation scenarios based on a dynamic selection mechanism to obtain virtual power plant clusters. S4. Construct a corresponding multi-resource coupling model for each virtual power plant cluster, build an objective function with the goal of maximizing the overall operating efficiency, and solve the objective function under the constraints of the multi-resource coupling model to obtain the collaborative control strategy of the virtual power plant cluster.

2. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 1, characterized in that: In S1, the basic characteristics of the power plant's distributed resources are analyzed to obtain heterogeneity features, including the following steps: The distributed resources include distributed power sources, energy storage resources, and flexible loads; the distributed power sources include wind power output and photovoltaic power output. Analysis of distributed power generation output yields the periodic characteristics of wind power output, the diurnal characteristics of photovoltaic power output, seasonal fluctuations, and climate-sensitive fluctuations. The load characteristics and response potential of energy storage resources are obtained by analyzing the load. The scheduling potential is obtained by analyzing the load adjustability time, adjustable load ratio, and response speed of flexible loads. The cyclical characteristics of wind power output, the diurnal characteristics of photovoltaic power output, seasonal fluctuations, climate-sensitive fluctuations, load characteristics, response potential, and dispatch potential are considered as heterogeneous characteristics.

3. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 2, characterized in that: In S1, response analysis of distributed resources is performed to obtain resource response characteristics, including the following steps: The influence of wind speed on wind power output is quantified to obtain wind power response characteristics, and the influence of sunlight and temperature on photovoltaic power output is quantified to obtain photovoltaic response characteristics. The energy storage response characteristics are obtained by quantifying the changes in charging and discharging power and capacity of energy storage resources; The load response characteristics are obtained by quantifying the load changes of flexible loads.

4. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 1, characterized in that: In S2, a dynamic feature library is obtained by quantifying heterogeneity characteristics and resource response characteristics based on real-time operational data of user-side distributed resources, including the following steps: Based on the time scale, heterogeneous features and resource response features are initially divided into intuitive descriptive features and ratio descriptive features; Using real-time operational data from user-side distributed resources as the data source, and assigning mathematical calculation formulas to intuitive descriptive features and ratio descriptive features, various numerical features are obtained through calculation. A dynamic feature library is obtained by organizing various numerical features and updating them in real time.

5. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 1, characterized in that: In S3, the resource cluster includes a new energy cluster, a load cluster, and a multi-energy coupling cluster.

6. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 5, characterized in that: In S3, uncertainty analysis is performed on various resource clusters to obtain fluctuation scenarios, including the following steps: Latin hypercube sampling method is used to sample the wind and solar power output fluctuations in various resource clusters to generate the original power output scenarios. The Monte Carlo simulation method is used to simulate load fluctuations in various resource clusters to generate original load scenarios; Time-series sampling of historical electricity prices in various resource clusters generates the original electricity price scenario for a time period. The original power output scenario, original load scenario, and original electricity price scenario are reduced and scenario weights are assigned by using the probabilistic distance method to obtain the fluctuation scenario.

7. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 6, characterized in that: In S3, a virtual power plant cluster is obtained by aggregating resources in fluctuating scenarios based on a dynamic selection mechanism, including the following steps: In fluctuating scenarios, the aggregation characteristics of resource clusters are calculated by dynamically matching the corresponding aggregation methods to the resource clusters. Associating functional tags with resource clusters based on their aggregation characteristics; Physical boundaries are defined based on the geographical location of resource clusters, and functional boundaries are defined based on the functional labels of resource clusters. The sub-clusters formed by the intersection of physical boundaries and functional boundaries are used as virtual power plant clusters.

8. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 7, characterized in that: The aggregation methods include direct aggregation, Monte Carlo simulation, dynamic programming, and geometric computation. The aggregation characteristics include cluster response time, resource adjustment range, and cluster adjustable capacity limit; The functional labels include peak shaving cluster, valley filling cluster, and backup cluster.

9. The method for coordinated control of virtual power plant clusters based on dynamic aggregation according to claim 1, characterized in that: In S4, a corresponding multi-resource coupling model is constructed for each virtual power plant cluster. An objective function is built with the goal of maximizing overall operational efficiency. The objective function is solved under the constraints of the multi-resource coupling model to obtain the collaborative control method for the virtual power plant cluster, including the following steps: An objective function is constructed with the sum of the costs of purchasing natural gas, energy storage, unit maintenance, and charging / discharging compensation for the virtual power plant cluster as the objective. A two-stage robust optimization method is used to solve the objective function to obtain the distributed output, load curves, and optimal electricity prices for each time period of each virtual power plant cluster; The distributed output, load curves, and optimal electricity prices for each time period of each virtual power plant cluster are used as the collaborative control strategy for virtual power plant clusters.

10. A virtual power plant cluster collaborative control system based on dynamic aggregation, applicable to the virtual power plant cluster collaborative control method based on dynamic aggregation as described in any one of claims 1-9, characterized in that: It includes a data processing module, a feature library generation module, a virtual power plant generation module, and a collaborative control module; The data processing module performs basic characteristic analysis on the power plant's distributed resources to obtain heterogeneity characteristics, and performs response analysis on the distributed resources to obtain resource response characteristics. The feature library generation module quantifies heterogeneous features and resource response features based on real-time operational data of user-side distributed resources to obtain a dynamic feature library, and classifies the features in the dynamic feature library to obtain resource clusters. The virtual power plant generation module performs uncertainty analysis on various resource clusters to obtain fluctuating scenarios, and aggregates resources from the fluctuating scenarios based on a dynamic selection mechanism to obtain virtual power plant clusters. The collaborative control module constructs a corresponding multi-resource coupling model for each virtual power plant cluster, builds an objective function with the goal of maximizing comprehensive operating efficiency, and solves the objective function under the constraints of the multi-resource coupling model to obtain the collaborative control method for virtual power plant clusters.

Images