Energy management control method for source-grid-load-storage integrated virtual power plant

Abstract

The application provides an energy management control method for a source, network, load and storage integrated virtual power plant, and belongs to the technical field of virtual power plant control. The method comprises the following steps: acquiring real-time and historical operation data of each region of a power grid and future meteorological data; calculating regional power consumption prediction data through a power consumption prediction model based on historical data and meteorological data; calculating a load pressure index according to real-time operation data and power consumption prediction data, identifying a virtual power plant that needs to be reconstructed and a reconstruction type; calculating energy management and control values for each virtual power plant to be reconstructed through an optimization model; matching virtual power plants of different reconstruction types in multiple rounds based on the energy management and control values, determining the reconstruction mode of unpaired virtual power plants, and generating a reconstruction instruction to control the virtual power plant to complete architecture reorganization and energy regulation. The application realizes dynamic balance and adaptive elastic regulation of cross-regional resources of the virtual power plant, and significantly improves new energy consumption capacity and power grid operation economy.

Description

Technical Field

[0001] This invention relates to the field of virtual power plant control technology, and more specifically to an energy management and control method for an integrated virtual power plant with source, grid, load and storage. Background Technology

[0002] The penetration rate of distributed renewable energy sources, represented by wind power and photovoltaics, in the power system is continuously increasing. The randomness and volatility of their output pose significant challenges to the safe and stable operation of the power grid. Virtual power plants, as an advanced technology and management model, aggregate widely distributed distributed power sources, energy storage systems, and controllable loads through advanced information and communication technologies and software systems. They participate in grid operation and electricity market transactions as a special type of power plant, and are one of the key means to improve the system's capacity to absorb distributed energy and enhance grid resilience.

[0003] However, existing virtual power plant energy management methods primarily focus on the optimal scheduling of aggregated resources within the virtual power plant to achieve goals such as minimizing operating costs or maximizing profits. In terms of organizational structure, virtual power plants are typically pre-constructed statically based on geographical or administrative boundaries, with relatively fixed aggregated resource scope and control capabilities. This static architecture struggles to adapt to dynamic changes in grid load: during peak load periods, virtual power plants in some areas may be unable to effectively alleviate grid congestion due to insufficient control capabilities; conversely, during off-peak periods with abundant renewable energy generation, virtual power plants in other areas may waste resources due to excess control. Current technologies lack a mechanism to dynamically adjust the organizational scale and structure of virtual power plants based on real-time and future grid conditions, resulting in insufficient cross-regional resource optimization and allocation capabilities, and an need to improve overall operational efficiency. Summary of the Invention

[0004] This invention provides an energy management and control method for an integrated virtual power plant with source, grid, load and storage, which solves the problems in the existing technology such as the static and fixed organizational structure of virtual power plants, which makes it difficult to adapt to the dynamic fluctuations of the power grid, and the lack of coordination in cross-regional resource regulation, which leads to the inability to effectively alleviate local load pressure.

[0005] To achieve the above objectives, this invention provides an energy management and control method for an integrated virtual power plant (EPP), comprising: Step S1: acquiring real-time and historical operating data of each region in the power grid, as well as meteorological data for a preset future time period; Step S2: calculating electricity consumption forecast data for each region based on the historical operating data and meteorological data using a preset electricity consumption forecast model; Step S3: calculating the load pressure index of each region based on the real-time operating data and electricity consumption forecast data, and identifying virtual power plants requiring reconfiguration and their corresponding reconfiguration types in conjunction with safe load intervals; Step S4: after identifying virtual power plants requiring reconfiguration, calculating energy control values ​​for each virtual power plant based on the corresponding load pressure index using a preset optimization model; Step S5: pairing virtual power plants of different reconfiguration types through multiple rounds of matching based on the calculated energy control values, determining the reconfiguration method of unpaired virtual power plants based on the pairing results, and generating reconfiguration instructions for each virtual power plant requiring reconfiguration to control the virtual power plant to complete architecture reorganization and energy regulation.

[0006] Optionally, the historical operating data includes historical load power and historical meteorological data. The electricity consumption forecasting model is constructed based on a long short-term memory neural network. The electricity consumption forecasting model outputs the electricity consumption forecasting data by analyzing historical load power and historical meteorological data and combining them with future meteorological data.

[0007] Optionally, the real-time operating data includes load power. Combined with the corresponding regional power grid capacity and electricity consumption forecast data, the load pressure index of each region is calculated. Each virtual power plant corresponds to a power grid region. The reconfiguration type includes split type and merge type. The identification of virtual power plant reconfiguration requirements and reconfiguration types includes: when the load pressure index of a virtual power plant is not within the safe load range, the virtual power plant is identified as a virtual power plant that needs to be reconfigured; for load pressure indices below the safe load range, the corresponding virtual power plant's reconfiguration type is identified as split type; for load pressure indices above the safe load range, the corresponding virtual power plant's reconfiguration type is identified as merge type.

[0008] Optionally, the process of calculating the energy control value includes: for a split virtual power plant, calculating the first difference between the current load pressure index of the virtual power plant and the lower limit of the safe load range; based on the first difference and electricity consumption forecast data, determining the maximum resource threshold that can be split; based on the maximum resource threshold and the basic ratio of each type of resource, outputting the split amount of each type of resource as the energy control value through a preset optimization model.

[0009] Optionally, the process of calculating the energy control value further includes: for a merged virtual power plant, calculating a second difference between the current load pressure index of the virtual power plant and the upper limit of the safe load range; based on the second difference, combined with the load growth rate in the electricity consumption forecast data, determining the total amount of energy gap that needs to be supplemented; based on the total energy gap, outputting the acceptance amount of each type of resource as the energy control value through an optimization model.

[0010] Optionally, the pairing process of the virtual power plants includes: establishing a feature matrix based on the reconstruction type, energy control values, and regional power grid topology characteristics of all virtual power plants to be reconstructed; the feature matrix includes the split vector of split-type virtual power plants and the demand vector of merged virtual power plants; comparing the corresponding resource types in the split vector and the demand vector, calculating the overlap, and taking a pair of split vectors and demand vectors with an overlap greater than a preset threshold as candidate pairing groups; for each candidate pairing group, calculating the matching score between the split vector and the demand vector using the Euclidean distance algorithm, combined with the line transmission distance; locking the virtual power plant corresponding to the candidate pairing group with the highest matching score and marking it as paired; repeating the pairing process for virtual power plants that have not been paired until there are no valid candidate groups.

[0011] Optionally, the reconfiguration methods for unpaired virtual power plants include resource sharing, resource invocation, and new construction. The process of determining the reconfiguration method includes: for unpaired split-type virtual power plants, determining their reconfiguration method as resource sharing; for unpaired merged-type virtual power plants, extracting the acceptance amount of each type of resource from their energy management values; updating the inventory of the preset shared resource pool based on the sum of the energy management values ​​of all unpaired split-type virtual power plants, and comparing the acceptance amount; if the inventory meets the acceptance amount, determining the reconfiguration method of the virtual power plant as resource invocation; if the inventory does not meet the acceptance amount, determining the reconfiguration method of the virtual power plant as new construction, and determining the specifications of the new virtual power plant based on the gap and electricity consumption forecast data.

[0012] Optionally, the pairing result includes paired and unpaired virtual power plants, as well as the resource exchange type and energy control value of each virtual power plant. The process of generating the reconfiguration instruction includes: for paired virtual power plants, parsing the resource exchange type and energy control value in the pairing result to generate basic instructions; combining the real-time operating parameters of the regional power grid to generate constraint parameters for the basic instructions; and integrating the basic instructions and constraint parameters into a standardized reconfiguration instruction according to a preset communication protocol.

[0013] Optionally, the constraint parameters include transmission power constraints, voltage stability constraints, frequency response constraints, and timing constraints. The generation of the constraint parameters includes: determining the range of single resource transmission power based on the line transmission capacity in the real-time operating parameters as a transmission power constraint; determining the voltage fluctuation threshold during resource exchange based on the node voltage monitoring values ​​in the real-time operating parameters as a voltage stability constraint; determining the range of frequency deviation amplitude caused by resource exchange based on the real-time frequency data in the real-time operating parameters as a frequency response constraint; and determining the time window for resource transmission based on the grid load change curve as a timing constraint.

[0014] Optionally, the process of generating the reconfiguration instruction further includes: for virtual power plants that require resource sharing, generating resource sharing instructions based on the amount of each type of resource in their energy management values, and storing the divisible resources into the shared resource pool according to type; for virtual power plants that require resource retrieval, generating corresponding resource retrieval instructions based on their energy management values ​​to retrieve the resources stored in the shared resource pool; and for scenarios where new virtual power plants need to be built, generating new control instructions based on the specifications of the new virtual power plants.

[0015] The energy management and control method for integrated virtual power plants (source-grid-load-storage) provided by this invention collects real-time grid operation data and meteorological information, combines long short-term memory neural networks to predict regional electricity demand, dynamically generates load pressure indicators, and accurately identifies low-load virtual power plants that need to be split and high-load virtual power plants that need to be merged. Based on an optimization model, it calculates the precise splitting or acceptance amount of each type of resource as energy control values, driving a multi-round feature matching mechanism to achieve resource characteristic adaptation and distance optimization pairing between split-type and merged power plants. At the same time, it establishes a shared resource pool to dynamically absorb unmatched resources and triggers the process of building new virtual power plants for remaining demand gaps. Finally, it generates standardized reconfiguration instructions by combining line capacity, voltage fluctuation limits, frequency deviation tolerance, and time window constraints, forming a complete control chain from situational awareness and intelligent decision-making to closed-loop execution. This effectively improves the cross-regional collaborative efficiency of source-grid-load-storage resources, eliminates the risk of local grid congestion, enhances the capacity for renewable energy absorption, and realizes the dynamic and elastic reconfiguration of virtual power plant clusters and the safe and economical operation of the power grid. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. In the drawings:

[0017] Figure 1This is a general flowchart of the energy management and control method provided in the embodiments of the present invention;

[0018] Figure 2 This is a flowchart of the reconstruction identification and quantization calculation provided in an embodiment of the present invention;

[0019] Figure 3 This is a flowchart of the multi-round matching algorithm provided in an embodiment of the present invention;

[0020] Figure 4 This is a shared resource pool scheduling logic diagram provided in an embodiment of the present invention;

[0021] Figure 5 This is an instruction generation constraint embedding diagram provided in the embodiments of the present invention. Detailed Implementation

[0022] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the scope of the present invention.

[0023] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application all comply with the relevant provisions of national laws and regulations. In the embodiments of this application, certain existing industry solutions such as software, components, and models may be mentioned. These should be considered exemplary, intended only to illustrate the feasibility of implementing the technical solution of this application, and do not imply that the applicant has already used or necessarily used such solutions.

[0024] As energy systems evolve towards a deeper integration of power generation, grid, load, and storage, existing virtual power plant energy management methods suffer from several problems: rigid static architectures that struggle to respond to dynamic grid fluctuations; a lack of coordination mechanisms in cross-regional regulation leading to inefficient local load relief; and a lack of quantitative support for energy dispatch commands causing execution deviations. Therefore, developing a virtual power plant energy management solution that deeply integrates multi-period situational awareness, dynamic reconfiguration, and precise quantitative regulation is crucial.

[0025] To address this issue, this invention provides an energy management and control method for an integrated virtual power plant (source-grid-load-storage). It achieves multi-dimensional coupling calculation of real-time and predicted data through load pressure indicators to accurately identify reconfiguration needs. It generates quantified energy control values ​​based on optimization models and feature matching mechanisms, and constructs a cross-regional resource elastic scheduling system by combining a shared resource pool and multi-round pairing algorithms. Through the generation of constraint parameters and standardized instruction encapsulation, it forms a closed-loop control from situational awareness to precise execution, achieving synergistic optimization of the virtual power plant's organizational dynamics, regional load balance, and renewable energy absorption capacity.

[0026] The following is combined Figures 1-5 This invention is described in detail.

[0027] like Figure 1 As shown in the figure, this embodiment of the invention provides an energy management and control method for an integrated virtual power plant (source-grid-load-storage), the energy management and control method comprising:

[0028] Step S1: Obtain real-time and historical operating data of each region in the power grid, as well as meteorological data for a future preset time period;

[0029] Step S2: Based on the historical operating data and meteorological data, calculate the electricity consumption forecast data for each region using a preset electricity consumption forecast model;

[0030] Step S3: Based on real-time operation data and electricity consumption forecast data, calculate the load pressure index of each region, and identify the virtual power plants that need to be reconfigured and the corresponding reconfiguration types in combination with the safe load range;

[0031] Step S4: After identifying the virtual power plants that need to be reconfigured, for each virtual power plant to be reconfigured, calculate the energy control value according to the corresponding load pressure index through the preset optimization model;

[0032] Step S5: Based on the calculated energy control values, pair virtual power plants of different reconfiguration types through multiple rounds of matching. Determine the reconfiguration method of unpaired virtual power plants based on the pairing results, and generate reconfiguration instructions for each virtual power plant that needs to be reconfigured to control the virtual power plants to complete architecture reorganization and energy regulation.

[0033] Real-time operational data includes load power, line transmission capacity, and node voltage monitoring values, while historical operational data includes historical load power, historical meteorological data, and historical grid capacity data. These data can be obtained from the grid monitoring terminal. The load pressure index is a quantitative parameter used to comprehensively assess the current and future load carrying capacity of the regional power grid. Its calculation integrates real-time operational data and electricity consumption forecast data to achieve dynamic and forward-looking assessment of load pressure. The safe load range refers to the load pressure range that ensures the stable operation of the regional power grid, preferably 50%-80%. The reconfiguration type refers to the adjustment mode divided according to the load pressure status of the virtual power plant, including splitting and merging. The energy control value refers to the core quantitative parameter used to guide the reconfiguration of the virtual power plant. For splitting virtual power plants, it is reflected in the amount of resources that can be split, and for merging virtual power plants, it is reflected in the amount of resources that need to be accepted. The reconfiguration command refers to the standardized command used to control the virtual power plant to perform architecture reorganization and energy regulation. It includes parameters such as resource interaction amount, transmission constraints, and execution timing, and can be parsed and executed by the virtual power plant management and control system.

[0034] The energy management and control method for integrated virtual power plants (source-grid-load-storage) provided in this invention integrates multi-dimensional data and predictive models to achieve accurate identification and forward-looking judgment of virtual power plant reconfiguration needs, avoiding the lag of traditional regulation. By leveraging the quantitative calculation of energy control values ​​and a multi-round matching mechanism, it improves the resource matching efficiency of split and merged virtual power plants, reducing resource mismatch and idleness. Simultaneously, by clarifying the reconfiguration methods of unpaired virtual power plants and generating standardized instructions, it forms a complete closed loop from data acquisition to execution regulation, effectively enhancing the synergy and stability of the integrated source-grid-load-storage system, and significantly improving the grid's flexibility in responding to load fluctuations and resource utilization efficiency.

[0035] Preferably, the historical operating data includes historical load power and historical meteorological data. The electricity consumption forecasting model is constructed based on a long short-term memory neural network. The electricity consumption forecasting model outputs the electricity consumption forecasting data by analyzing historical load power and historical meteorological data and combining them with future meteorological data.

[0036] Specifically, when predicting electricity consumption data, historical load power and corresponding historical meteorological data are first used as training samples and input into the LSTM model for parameter optimization, so that the model learns the potential correlation between meteorological conditions and load changes, such as load increasing by 5% for every 1°C increase in temperature. In the prediction stage, meteorological data for a preset time period in the future, such as temperature and sunshine forecasts for the next 7 days, are input into the trained model. Combined with the periodic patterns of historical load, the corresponding electricity consumption prediction data for the time period is output.

[0037] In a preferred embodiment of this invention, a long short-term memory (LSTM) neural network is introduced to construct an electricity consumption forecasting model, which integrates historical load power and meteorological data, significantly improving the accuracy and reliability of electricity consumption forecasting. Compared with traditional time series models, the gating mechanism of LSTM can effectively capture the long-term dependencies of load data, avoiding prediction bias caused by excessively long historical data sequences. Simultaneously, by training the model in association with historical meteorological data and load data, the model can accurately identify the impact of meteorological factors on load. When combining future meteorological data for forecasting, it can predict load fluctuation trends in advance, solving the problem of insufficient meteorological sensitivity caused by traditional models relying solely on load data.

[0038] Preferably, the real-time operating data includes load power. Combined with the corresponding regional power grid capacity and electricity consumption forecast data, the load pressure index of each region is calculated. Each virtual power plant corresponds to a power grid region. The reconfiguration type includes split type and merge type. The identification of virtual power plant reconfiguration needs and reconfiguration types includes: when the load pressure index of a virtual power plant is not within the safe load range, the virtual power plant is identified as a virtual power plant that needs to be reconfigured; for load pressure indices below the safe load range, the corresponding virtual power plant's reconfiguration type is identified as split type; for load pressure indices above the safe load range, the corresponding virtual power plant's reconfiguration type is identified as merge type.

[0039] Among them, load power refers to the total instantaneous active load power of the power grid area covered by the virtual power plant at the current moment, including the real-time superposition value of various types of electricity loads such as industrial production load, commercial operation load, and residential load, which is collected and summarized in real time by distributed smart meters and concentrators in the area; regional power grid capacity refers to the maximum carrying capacity of the power grid area corresponding to the virtual power plant, which is determined by the rated capacity of the substation, the current carrying capacity of the transmission line, and the maximum withstand power of the distribution equipment in the area. It is the core parameter for measuring the power supply capacity of the regional power grid; the load pressure index can be calculated as: load pressure index = max(current load power, future maximum load power) ÷ regional power grid capacity × 100%. For example, if the current load power is 30MW, the predicted maximum load in the next 7 days is 40MW, and the regional power grid capacity is 50MW, then the load pressure index = max(30,40) ÷ 50 × 100% = 80%.

[0040] For example, a virtual power plant in a certain region corresponds to a grid capacity of 100MW, with a preset safe load range of 50%-80%. First, the current load power in the real-time operation data of this region is obtained as 60MW, and the maximum load power in the next 7 days in the electricity consumption forecast data is 90MW. Then, the load pressure index is calculated as 90%. This index is compared with the safe load range. Since 90% > 80%, the virtual power plant is identified as a merged virtual power plant that needs to be restructured. If the current load power in another region is 40MW, and the maximum predicted load in the future is 30MW, the calculated load pressure index is 40%. Since 40% < 50%, it is identified as a split virtual power plant that needs to be restructured.

[0041] The preferred embodiments of this invention provide quantifiable and actionable criteria for judging the reconfiguration needs of virtual power plants, significantly improving the objectivity and accuracy of reconfiguration decisions. Compared to the traditional overload-based control mode that relies on experience-based judgment, this scheme combines real-time load with future predicted load to calculate pressure indicators, avoiding short-sighted control caused by relying solely on real-time data. The quantitative setting of the safe load range makes the boundary between reconfiguration and non-reconfiguration clearer, reducing the subjective error of human judgment.

[0042] like Figure 2 As shown, preferably, the process of calculating the energy control value includes: for a split virtual power plant, calculating the first difference between the current load pressure index of the virtual power plant and the lower limit of the safe load range; based on the first difference and electricity consumption forecast data, determining the maximum resource threshold that can be split; based on the maximum resource threshold and the basic ratio of each type of resource, outputting the split amount of each type of resource as the energy control value through a preset optimization model.

[0043] Further preferably, the process of calculating the energy control value also includes: for a merged virtual power plant, calculating a second difference between the current load pressure index of the virtual power plant and the upper limit of the safe load range; based on the second difference, combined with the load growth rate in the electricity consumption forecast data, determining the total amount of energy gap that needs to be supplemented; based on the total energy gap, outputting the acceptance amount of each type of resource as the energy control value through an optimization model.

[0044] Specifically, the first difference refers to the difference between the load pressure index and the lower limit of the safe load range for split-type virtual power plants, reflecting the resource redundancy of the virtual power plant; the second difference refers to the difference between the load pressure index and the upper limit of the safe load range for merged virtual power plants, reflecting the urgency of the resource gap for the virtual power plant; the total energy gap refers to the total amount of resources that merged virtual power plants need to supplement, calculated by combining the second difference with the load growth rate in the electricity consumption forecast data, and the specific calculation formula is: total energy gap = (second difference × regional grid capacity) × (1 + load growth rate × forecast period); the basic allocation ratio refers to the inherent proportion of various resources in split-type virtual power plants.

[0045] Specifically, a regional power grid has a capacity of 120MW, with a safe load range set at 50%-80%. A split-type virtual power plant has a current load of 45MW, and the predicted maximum future load is 50MW. Using the load pressure index calculation formula, this index is 50 ÷ 120 × 100% = 41.67%. The lower limit of the safe load range is 50%, and the first difference is 50% - 41.67% = 8.33%. The corresponding resource redundancy is 8.33% × 120MW = 10MW. Combining this with the predicted minimum future load of 50MW, the maximum resource threshold is determined to be 10MW. The basic resource allocation of this virtual power plant is 20% photovoltaic, 30% energy storage, and 50% adjustable load. Through optimization model calculations, the split amount is 2MW photovoltaic, 3MW energy storage, and 5MW adjustable load, which corresponds to the energy control values.

[0046] Another regional power grid has a capacity of 150MW, with a safe load range of 50%-80%. For the merged virtual power plant, its current load is 110MW, and the maximum future load in the electricity consumption forecast data is 125MW, with a load pressure index of 125÷150×100%=83.33%. The upper limit of the safe load range is 80%, and the second difference is 83.33%-80%=3.33%. The electricity consumption forecast data shows that the load growth rate for the next two days is 5% / day, which, calculated using the total energy gap formula, is 3.33%×150MW×(1+5%×2)=5.5MW. Through optimization modeling, allocating the load in a ratio of 40% for photovoltaic, 40% for energy storage, and 20% for adjustable load, the acceptable capacity is 2.2MW for photovoltaic, 2.2MW for energy storage, and 1.1MW for adjustable load, which is the energy control value.

[0047] In a preferred embodiment of the present invention, the calculation method for energy control values ​​quantifies resource redundancy or shortage by combining the difference between load pressure indicators and safe load ranges. It also incorporates electricity consumption forecast data and outputs specific splitting or acceptance quantities through an optimization model based on resource allocation. This ensures that split-type virtual power plants can still meet their future load demands after releasing redundant resources, avoiding supply shortages caused by excessive splitting. Furthermore, it guarantees that the resources acquired by merged virtual power plants can not only fill current gaps but also cope with future load growth. Simultaneously, the unified resource type allocation provides accurate numerical basis for subsequent virtual power plant pairing, reducing the risk of resource mismatch. Overall, it improves the accuracy, foresight, and collaborative efficiency of resource allocation in the integrated source-grid-load-storage system, enhancing the reliability of virtual power plant architecture reorganization and energy regulation.

[0048] like Figure 3 As shown, preferably, the pairing process of the virtual power plants includes: establishing a feature matrix based on the reconstruction type, energy control values, and regional power grid topology characteristics of all virtual power plants to be reconstructed; the feature matrix includes the split vector of split-type virtual power plants and the demand vector of merged virtual power plants; comparing the corresponding resource types in the split vector and the demand vector, calculating the overlap, and taking a pair of split vectors and demand vectors with an overlap greater than a preset threshold as candidate pairing groups; for each candidate pairing group, using the Euclidean distance algorithm to calculate the matching score between the split vector and the demand vector, combined with the line transmission distance; locking the virtual power plant corresponding to the candidate pairing group with the highest matching score and marking it as paired; repeating the pairing process for virtual power plants that have not been paired until there are no valid candidate groups.

[0049] The feature matrix refers to a dataset that integrates key information of all virtual power plants to be reconstructed in tabular form. Rows represent virtual power plant numbers, and columns include reconstructing type, split vector, demand vector, and regional power grid topology features. Resource type overlap is an indicator measuring the degree of resource matching between split and merged virtual power plants, calculated as: number of overlapping resource types ÷ total number of resource types × 100%. Matching degree is a quantitative indicator combining resource type overlap and energy value matching. It is calculated using the Euclidean distance algorithm to determine the spatial distance between the split vector and the demand vector; the smaller the distance, the closer the resource quantity matching. For example, the Euclidean distance between a split vector (4, 2, 5) and a demand vector (3, 2, 4) is... The Euclidean distance is converted into a score of 0-100. The formula is: Match score = 100 / (1 + Euclidean distance).

[0050] Specifically, the pairing process for virtual power plants is as follows: First, the energy control values ​​of split-type virtual power plants are sorted by resource type, such as photovoltaic (PV), energy storage, and adjustable load, to form a split vector. For example, split-type A has a split volume of 3MW PV, 1MW energy storage, and 2MW adjustable load, corresponding to the split vector (3, 1, 2). The energy control values ​​of merged-type virtual power plants are sorted by the same resource type to form a demand vector. For example, merged-type X needs to accept 2MW PV, 1MW energy storage, and 2MW adjustable load, corresponding to the demand vector (2, 1, 2). Then, the spatial distance between the two types of vectors is calculated using the Euclidean distance algorithm. The distance between A and X is... The score is then converted to a matching score of 100 / (1+1)=50. Multiple rounds of matching are then performed. In the first round, candidate groups with completely overlapping resource types are selected and matched in descending order of score. After removing matched virtual power plants, the remaining virtual power plants are recalculated and matched repeatedly until no valid candidate groups remain.

[0051] In a preferred embodiment of this invention, the matching degree between the split vector and the demand vector is quantified using the Euclidean distance algorithm. Compared with subjective judgment, this more accurately identifies the degree of resource supply and demand matching, avoiding mismatch problems where the types overlap but the numerical deviations are large. The multi-round matching mechanism based on scores ensures that virtual power plants with high matching degrees complete resource interaction first, maximizing the utilization of redundant resources and quickly filling gaps. Unmatched virtual power plants are handled through a closed-loop process of shared resource pools and new additions, which not only avoids the idleness of redundant resources but also ensures the final satisfaction of gap demands. Overall, this improves the efficiency and flexibility of virtual power plant resource allocation and enhances the coordination and stability of the integrated source-grid-load-storage system in dealing with complex load scenarios.

[0052] like Figure 4As shown, preferably, the reconfiguration methods for unpaired virtual power plants include resource sharing, resource invocation, and new construction. The process of determining the reconfiguration method includes: for unpaired split-type virtual power plants, determining their reconfiguration method as resource sharing; for unpaired merged-type virtual power plants, extracting the acceptance amount of each type of resource from their energy management values; updating the inventory of the preset shared resource pool based on the sum of the energy management values ​​of all unpaired split-type virtual power plants, and comparing the acceptance amount; if the inventory meets the acceptance amount, determining the reconfiguration method of the virtual power plant as resource invocation; if the inventory does not meet the acceptance amount, determining the reconfiguration method of the virtual power plant as new construction, and determining the specifications of the new virtual power plant based on the gap and electricity consumption forecast data.

[0053] Among them, resource sharing refers to the regulation method in which unpaired split-type virtual power plants include their own splittable resources into a shared resource pool according to type, which can be used by other merged virtual power plants; resource retrieval refers to the operation of unpaired merged virtual power plants extracting the required resources from the shared resource pool to fill the gap, and the amount of retrieval shall not exceed the inventory limit of the corresponding type of resource pool; the shared resource pool is a centralized reserve system for storing long-term idle resources of virtual power plants in various regions, including information such as resource type, capacity, and available time, as a supplementary regulation resource for unpaired virtual power plants.

[0054] Specifically, the calculation process for the specifications of a new virtual power plant includes: subtracting the corresponding type of inventory in the shared resource pool from the required capacity of the merged virtual power plant to obtain the gap for each resource type. For example, if 5MW of photovoltaic power is needed and the resource pool has 2MW, the gap is 3MW. Then, based on the predicted future load growth rate of electricity consumption, such as 10%, it is used as the increase coefficient for the gap. For example, the 3MW photovoltaic gap becomes 3.3MW.

[0055] For example, a split-type virtual power plant C has not been paired up. Its energy management value is 2MW of photovoltaic (PV) and 1MW of energy storage. These resources are stored in the shared resource pool according to the resource sharing method, so that the pool inventory is updated to 2MW of PV and 1MW of energy storage. The unpaired merged virtual power plant D needs to accept 1.5MW of PV and 0.8MW of energy storage. After comparing with the inventory of the shared resource pool, the reconstruction method is determined to be resource retrieval. The corresponding amount of resources is extracted from the pool, and the pool has 0.5MW of PV and 0.2MW of energy storage remaining. Another unpaired merged virtual power plant E needs to accept 3MW of PV. Since the shared resource pool only has 0.5MW of PV which cannot meet the demand, the calculated gap is 2.5MW. Combined with the electricity consumption forecast data that the future load growth of area E is 10%, the reconstruction method is determined to be new construction. The new construction specification is a 2.75MW PV power plant.

[0056] The preferred embodiment of this invention clarifies the reconfiguration method of unpaired virtual power plants, enabling the efficient utilization of split redundant resources through a shared resource pool, prioritizing the use of shared resources in merged projects to reduce new construction costs and timelines, and accurately determining new construction specifications only when resources are insufficient, based on electricity consumption forecast data. This forms a closed-loop control system covering all unpaired scenarios, improving the overall resource utilization rate and ensuring the final satisfaction of demand gaps. It also enhances the flexibility and reliability of the virtual power plant energy management system in dealing with complex pairing results, ensuring the integrity and stability of integrated source-grid-load-storage control.

[0057] like Figure 5 As shown, preferably, the pairing result includes paired virtual power plants and unpaired virtual power plants, as well as the resource exchange type and energy control value of each virtual power plant. The process of generating the reconfiguration instruction includes: for paired virtual power plants, parsing the resource exchange type and energy control value in the pairing result to generate basic instructions; combining the real-time operating parameters of the regional power grid to generate constraint parameters for the basic instructions; and integrating the basic instructions and constraint parameters into a standardized reconfiguration instruction according to a preset communication protocol.

[0058] Further preferably, the constraint parameters include transmission power constraints, voltage stability constraints, frequency response constraints, and timing constraints. The generation of the constraint parameters includes: determining the range of power for a single resource transmission based on the line transmission capacity in the real-time operating parameters as a transmission power constraint; determining the voltage fluctuation threshold during resource exchange based on the node voltage monitoring values ​​in the real-time operating parameters as a voltage stability constraint; determining the range of frequency deviation amplitude caused by resource exchange based on the real-time frequency data in the real-time operating parameters as a frequency response constraint; and determining the time window for resource transmission based on the grid load change curve as a timing constraint.

[0059] Among them, the basic instruction refers to the prototype of the instruction containing the core information of resource interaction, including the sender, receiver, resource type, and interaction quantity; the standardized reconfiguration instruction refers to the structured instruction that integrates the basic instruction and constraint parameters according to the preset communication protocol such as IEC61850, which can be directly parsed and executed by the virtual power plant control system; the calculation of the transmission power constraint is expressed as follows: based on the line transmission capacity in the real-time operating parameters, such as the line L1 capacity of 10MW, 80% of it is taken as the upper limit of the single resource transmission power, such as 10×80%=8MW, to avoid line overload; the calculation of the voltage stability constraint is expressed as follows: based on the node voltage monitoring value, such as 10.5kV, the fluctuation threshold is set to ±5%, that is, the upper limit is 11.0kV and the lower limit is 10.0kV, to ensure that the voltage is stable within the safe range during resource interaction; the frequency response constraint is expressed as follows, based on the real-time frequency of the power grid, stipulating that the frequency deviation caused by resource transmission shall not exceed ±0.2Hz, to prevent frequency fluctuations from affecting the stability of the power grid; the timing constraint is expressed as follows, based on the analysis of the power grid load change curve, the resource transmission time window is set during the load trough period to reduce the impact on the power grid.

[0060] Specifically, when generating reconfiguration instructions, the pairing results are first parsed, including the resource exchange type and energy control values ​​of paired virtual power plants, and the reconfiguration method of unpaired virtual power plants. Basic instructions are then generated. For paired virtual power plants, the sender, receiver, resource type, and interaction quantity are specified, such as split-type A transmitting 4MW of photovoltaic power and 2MW of energy storage power to merge-type X. For unpaired merge-type power plants that require resource allocation, the caller, resource pool, resource type, and allocation quantity are specified. For new power plants, the new location, resource type, and capacity are specified. Next, based on the real-time operating parameters of the regional power grid and electricity consumption forecast data, four types of constraint parameters are generated: transmission power constraints are determined based on line transmission capacity; voltage stability constraints are set based on node voltage monitoring values; frequency response constraints are defined based on real-time frequency data; and timing constraints are determined by referring to load change curves. Finally, the basic instructions and constraint parameters are integrated into standardized reconfiguration instructions according to the preset communication protocol. For example, the instruction is: A transmits 4MW of photovoltaic power and 2MW of energy storage to X; the constraints are: power ≤8MW, voltage 10.0-11.0kV, frequency 49.8-50.2Hz, and execution time 10:00-11:00. This ensures that the instructions can be directly parsed and executed by the virtual power plant control system, which clarifies the core operation and ensures the safety and stability of the power grid.

[0061] The preferred embodiment of this invention generates standardized reconfiguration instructions that include basic instructions and multi-dimensional constraint parameters. This clarifies the core operations of virtual power plant resource interaction and ensures that the operation adapts to the real-time state of the power grid through constraints such as transmission power and voltage stability. At the same time, timing constraints avoid peak loads and reduce impacts, making the reconfiguration instructions both executable and safe. This improves the accuracy and efficiency of virtual power plant architecture reconfiguration, ensures the stable operation of the power grid during resource regulation, and enhances the collaborative reliability of the integrated source-grid-load-storage system.

[0062] like Figure 4 As shown, preferably, the process of generating the reconfiguration instruction further includes: for virtual power plants that require resource sharing, generating resource sharing instructions based on the amount of each type of resource in their energy control values, and storing the divisible resources into the shared resource pool according to type; for virtual power plants that require resource retrieval, generating corresponding resource retrieval instructions based on their energy control values ​​to retrieve the resources stored in the shared resource pool; and for scenarios where new virtual power plants need to be built, generating new control instructions based on the specifications of the new virtual power plants.

[0063] Among them, the resource sharing instructions include various types of divisible resources and their quantities, as well as the storage time; the resource call instructions include the call list and priority; and the new control instructions include construction parameters, time nodes, etc.

[0064] Specifically, for a split-type virtual power plant G requiring resource sharing, its energy control values ​​are 3MW of photovoltaic (PV) and 1MW of energy storage. The generated instruction is to store G's 3MW PV and 1MW of energy storage into the shared resource pool according to their types, marking their availability as 48 hours. After execution, the resource pool inventory is updated to 3MW PV and 1MW energy storage. For a merged-type virtual power plant H requiring resource allocation, its energy control value is to accept 2MW of PV. The generated instruction is to allocate 2MW of PV from the shared resource pool to H during the allocation period from 14:00 to 16:00 on the same day. After execution, the resource pool has 1MW of PV remaining. For scenarios requiring the construction of new virtual power plants, the energy control value of merged-type virtual power plant I is to accept 2MW of energy storage, while the shared resource pool only stores 1MW of energy. Based on electricity consumption forecast data, a new 1.1MW energy storage power station is determined. The generated instruction is to construct a 1.1MW lithium iron phosphate energy storage power station in the area where I is located, connect it to a 10kV line, and complete commissioning and put it into operation within 3 days.

[0065] This invention provides a method to generate targeted reconfiguration instructions for resource sharing, invocation, and new creation scenarios, thereby enabling the control of unpaired virtual power plants to form a closed loop: split redundant resources are stored in a shared pool according to type, merged resources can be quickly invoked from the pool, and new creation instructions are accurately implemented according to specifications. This not only improves resource utilization but also ensures that demand gaps are quickly met. At the same time, it provides an executable operational basis for long-term power grid planning, enhancing the integrity and effectiveness of the virtual power plant energy management system in complex scenarios.

[0066] This invention provides an energy management and control method for integrated virtual power plants (power generation, grid, load, and storage). The method generates accurate load pressure indicators by real-time acquisition of multi-source grid data and fusion with long short-term memory neural network predictions. This dynamically identifies resource-redundant virtual power plants that need to be split and resource-scarce virtual power plants that need to be merged. Based on an optimization model, the method quantifies the splitting or acceptance of various resource types, driving a multi-round feature matching mechanism to achieve optimal cross-regional resource pairing. It dynamically schedules unmatched resources using a shared resource pool and triggers new resource creation processes for gap needs. Finally, it embeds real-time constraints such as line transmission capacity, voltage fluctuation limits, frequency limits, and timing windows to generate standardized reconfiguration instructions, forming a closed-loop control system from situational awareness and intelligent decision-making to safe execution. This significantly improves the dynamic adaptability of virtual power plant clusters to grid fluctuations, achieving the triple goals of cross-regional load pressure relief, increased renewable energy absorption rate, and optimized system operation economy.

[0067] The above description is merely a preferred embodiment of the technical solution of the present invention and is not intended to limit the scope of protection of the present invention. 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. An energy management and control method for an integrated virtual power plant (source-grid-load-storage), characterized in that, The energy management and control method includes: Step S1: Obtain real-time and historical operating data of each area in the power grid, as well as meteorological data for a future preset time period. The real-time operating data includes load power. Step S2: Based on the historical operating data and meteorological data, calculate the electricity consumption forecast data for each region using a preset electricity consumption forecast model; Step S3: Based on the load power in the real-time operating data, the corresponding regional power grid capacity, and the electricity consumption forecast data, calculate the load pressure index for each region, and identify the virtual power plants that need to be reconfigured and the corresponding reconfiguration type in conjunction with the safe load range; wherein, the safe load range is a preset load pressure range to ensure the stable operation of the regional power grid; the reconfiguration type includes split type and merge type. When the load pressure index is less than the safe load range, it is identified as split type; when the load pressure index is greater than the safe load range, it is identified as merge type. Step S4: After identifying the virtual power plants that need to be reconfigured, for each virtual power plant to be reconfigured, calculate the energy control value according to the corresponding load pressure index through a preset optimization model; wherein, for split-type virtual power plants, the energy control value is the amount of each type of resource that can be split, and for merged-type virtual power plants, the energy control value is the amount of each type of resource that needs to be accepted. Step S5: Based on the calculated energy control values, virtual power plants of different reconfiguration types are paired through multiple rounds of matching. The multiple rounds of matching include: establishing a feature matrix based on the reconfiguration type, energy control values, and regional power grid topology characteristics of all virtual power plants to be reconfigured; the feature matrix includes the split vector of split-type virtual power plants and the demand vector of merged virtual power plants; comparing the corresponding resource types in the split vector and demand vector, calculating the overlap, and taking a pair of split vectors and demand vectors with an overlap greater than a preset threshold as candidate pairing groups; for each candidate pairing group, using the Euclidean distance algorithm to calculate the matching degree score between the split vector and demand vector, combined with the line transmission distance; locking the virtual power plant corresponding to the candidate pairing group with the highest matching degree score and marking it as paired; repeating the pairing process for virtual power plants that have not been paired until there are no valid candidate groups; determining the reconfiguration method of unpaired virtual power plants based on the pairing results, and generating a reconfiguration instruction for each virtual power plant that needs to be reconfigured to control the virtual power plant to complete the architecture reorganization and energy regulation.

2. The energy management and control method according to claim 1, characterized in that, The historical operating data includes historical load power and historical meteorological data. The electricity consumption forecasting model is constructed based on a long short-term memory neural network. The electricity consumption forecasting model outputs the electricity consumption forecasting data by analyzing historical load power and historical meteorological data and combining them with future meteorological data.

3. The energy management and control method according to claim 1, characterized in that, In step S4, for a split-type virtual power plant, the calculation of the split amount includes: Calculate the first difference between the current load pressure index of the virtual power plant and the lower limit of the safe load range; Based on the first difference and electricity consumption forecast data, determine the maximum resource threshold that can be split; Based on the maximum resource threshold and the basic allocation ratio of each type of resource, the splitting amount of each type of resource is output through a preset optimization model.

4. The energy management and control method according to claim 1, characterized in that, In step S4, for a merged virtual power plant, the calculation of the acceptance quantity includes: Calculate the second difference between the current load pressure index and the upper limit of the safe load range for the virtual power plant; Based on the second difference, and combined with the load growth rate in the electricity consumption forecast data, the total amount of energy gap that needs to be supplemented is determined; Based on the total energy gap, the model is optimized to output the acceptance amount of each type of resource.

5. The energy management and control method according to claim 1, characterized in that, The reconfiguration methods for the unpaired virtual power plant include resource sharing, resource allocation, and new creation. The process for determining the reconfiguration method includes: For unpaired, split-type virtual power plants, their reconfiguration method is determined to be resource sharing; For unpaired, merged virtual power plants, extract the acceptance of each type of resource from their energy management values; The preset shared resource pool inventory is updated based on the sum of the energy control values ​​of all unpaired split virtual power plants, and compared with the acceptance quantity; If the inventory meets the acceptance requirements, then the reconstruction method for the virtual power plant is determined to be resource allocation; If the inventory does not meet the acceptance volume, the reconstruction method of the virtual power plant is determined to be new construction, and the specifications of the new virtual power plant are determined based on the gap and electricity consumption forecast data.

6. The energy management and control method according to claim 1, characterized in that, The pairing results include paired and unpaired virtual power plants, as well as the resource exchange type and energy control values ​​of each virtual power plant. The process of generating the reconfiguration instruction includes: For paired virtual power plants, the resource exchange type and energy control value in the pairing results are parsed to generate basic instructions; Based on the real-time operating parameters of the regional power grid, constraint parameters are generated for the basic commands; The basic instructions and constraint parameters are integrated into standardized reconfiguration instructions according to the preset communication protocol.

7. The energy management and control method according to claim 6, characterized in that, The constraint parameters include transmission power constraints, voltage stability constraints, frequency response constraints, and timing constraints. The generation of these constraint parameters includes: Based on the line transmission capacity in the real-time operating parameters, the range of single resource transmission power is determined as a transmission power constraint. Based on the node voltage monitoring values ​​in the real-time operating parameters, determine the voltage fluctuation threshold during resource exchange as a voltage stability constraint. Based on the real-time frequency data in the real-time operating parameters, determine the range of frequency deviation caused by resource exchange as a frequency response constraint. Based on the power grid load change curve, the time window for resource transmission is determined as a timing constraint.

8. The energy management and control method according to claim 5, characterized in that, The process of generating the reconstruction instructions also includes: For virtual power plants that require resource sharing, resource sharing instructions are generated based on the amount of each type of resource in their energy management values, and the divisible resources are stored in the shared resource pool according to their type. For virtual power plants that require resource allocation, corresponding resource allocation instructions are generated based on their energy control values ​​to access resources stored in the shared resource pool. For scenarios requiring the construction of new virtual power plants, new control commands are generated based on the specifications of the new virtual power plants.

Images