A virtual power plant collaborative control and scheduling method and system

Abstract

The present invention is applicable to the field of power grid control technology, and provides a method and system for collaborative control and scheduling of a virtual power plant, comprising the following steps: constructing a multi-dimensional resource library, wherein each resource node in the multi-dimensional resource library corresponds to real-time adjustable power, power regulation cost, response delay time, carbon emission coefficient, and resource availability coefficient; establishing a global optimization model in the cloud, wherein the global optimization model integrates economic efficiency, environmental benefits, and system stability; using an improved non-dominated sorting genetic algorithm to solve the global optimization model, obtain an optimal scheduling strategy, and send the optimal scheduling strategy to a control center and edge terminals; determining the real-time status of the resource node based on the edge terminal, and adjusting the optimal scheduling strategy according to the real-time status. The present invention significantly improves the response accuracy and system stability of the virtual power plant by constructing a global optimization model and dynamic collaborative control, and using an improved non-dominated sorting genetic algorithm for solution.

Description

Technical Field

[0001] The present invention relates to the field of power grid control technology, and in particular to a collaborative control and scheduling method and system for a virtual power plant. Background Art

[0002] With the large-scale integration of renewable energy sources (such as wind power and photovoltaics) into the power grid, the volatility and uncertainty of the power system have increased significantly. Traditional centralized power generation scheduling models are unable to adapt to the flexible characteristics of distributed energy. As a new energy aggregation management model, virtual power plants (VPs) integrate distributed power sources, energy storage systems, controllable loads, and other resources to achieve coordinated and optimized resource scheduling, becoming a key technology for improving grid flexibility and reliability. However, existing VP scheduling technologies face the following challenges: There is an inherent contradiction between economic efficiency, environmental benefits, and system stability, which traditional single-objective optimization methods find difficult to address. Furthermore, multi-objective optimization problems involve high-dimensional decision variables and complex constraints, making traditional algorithms prone to local optimality and computationally inefficient. Therefore, there is a need to provide a VP scheduling method and system for collaborative control to address these issues. Summary of the Invention

[0003] In view of the deficiencies in the prior art, the purpose of the present invention is to provide a virtual power plant collaborative control and scheduling method and system to solve the problems existing in the above-mentioned background technology.

[0004] The present invention is implemented as follows: a virtual power plant collaborative control and scheduling method, the method comprising the following steps:

[0005] Build a multi-dimensional resource library, where each resource node corresponds to real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient, and resource availability coefficient;

[0006] Aggregate all resource nodes in the control center to form a resource pool for the virtual power plant;

[0007] Establishing a global optimization model in the cloud, wherein the global optimization model integrates economic efficiency, environmental benefits and system stability;

[0008] An improved non-dominated sorting genetic algorithm is used to solve the global optimization model and obtain the optimal scheduling strategy, which is then distributed to the control center and edge terminals.

[0009] The real-time status of resource nodes is determined based on the edge terminal, and the optimal scheduling strategy is adjusted according to the real-time status.

[0010] As a further solution of the present invention: the step of constructing a multi-dimensional resource library specifically includes:

[0011] Collect power generation impact data of each resource node and determine real-time adjustable power based on the power generation impact data;

[0012] Determine the power regulation cost, response delay time, and carbon emission coefficient of each resource node based on the resource node type and historical data;

[0013] Determine the predicted power, rated power, and power generation equipment status of the resource node, and calculate the resource availability coefficient;

[0014] The real-time adjustable power, power regulation cost, response delay time, carbon emission coefficient and resource availability coefficient of all resource nodes are integrated to build a multi-dimensional resource library.

[0015] As a further solution of the present invention: the step of establishing a global optimization model in the cloud specifically includes:

[0016] Determine the total regulating power of the virtual power plant based on the load demand of the power grid and the grid interaction power;

[0017] Determine economic goals based on power regulation costs, environmental goals based on carbon emission coefficients, and stability goals based on response delay time and resource availability coefficients;

[0018] A global optimization model is constructed based on the total regulating power, economic goals, environmental protection goals and stability goals.

[0019] As a further solution of the present invention: the step of using the improved non-dominated sorting genetic algorithm to solve the global optimization model to obtain the optimal scheduling strategy specifically includes:

[0020] Perform chromosome encoding, where each gene position corresponds to the regulation power ΔPi of a resource node, and the chromosome is directly mapped to the three objective function values ​​of economy, environmental protection, and stability;

[0021] Initialize the population according to the predetermined number, using stratified sampling: A% individuals are generated by Latin hypercube sampling, B% individuals are generated based on the neighborhood perturbation of the historical optimal solution, and C% individuals are generated using greedy heuristics, A+B+C=100;

[0022] Determine the dominance relationship of each individual and divide the population into multiple non-dominated levels based on the dominance relationship. The first level is the solution that is not dominated by any individual at all, the second level is the solution that is only dominated by the first level, and so on.

[0023] For individuals in each level, the normalized distances of the three target values ​​are calculated respectively; in each target direction, individuals are sorted by the normalized values, the distances between adjacent individuals in the target space are calculated, and the crowding degree is determined;

[0024] The genetic iterative operation of the population is performed according to the non-dominated hierarchy and congestion degree, and the optimal scheduling strategy is output.

[0025] As a further solution of the present invention: the step of adjusting the optimal scheduling strategy according to the real-time status specifically includes:

[0026] When the resource availability coefficient in the real-time state is lower than the availability threshold, the regulation power of the resource node is proportionally reduced;

[0027] When the response delay time is higher than the delay threshold, the instruction is issued in advance, and the advance time is determined according to the response delay time and the safety margin.

[0028] Another object of the present invention is to provide a virtual power plant collaborative control and scheduling system, the system comprising:

[0029] The resource library construction module is used to build a multi-dimensional resource library. Each resource node in the multi-dimensional resource library corresponds to real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient, and resource availability coefficient;

[0030] The resource node aggregation module is used to aggregate all resource nodes in the control center to form a resource pool for the virtual power plant;

[0031] An optimization model building module, used to build a global optimization model in the cloud, wherein the global optimization model integrates economic efficiency, environmental benefits and system stability;

[0032] The optimal scheduling strategy module is used to solve the global optimization model using an improved non-dominated sorting genetic algorithm to obtain the optimal scheduling strategy, and then send the optimal scheduling strategy to the control center and edge terminals;

[0033] The edge terminal adjustment module is used to determine the real-time status of resource nodes based on the edge terminal and adjust the optimal scheduling strategy according to the real-time status.

[0034] As a further solution of the present invention: the resource library construction module includes:

[0035] The adjustable power unit is used to collect power generation impact data of each resource node and determine the real-time adjustable power based on the power generation impact data;

[0036] A historical data analysis unit, configured to determine the power regulation cost, response delay time, and carbon emission coefficient of each resource node based on the resource node type and historical data;

[0037] Resource availability unit, used to determine the predicted power, rated power and power generation equipment status of the resource node, and calculate the resource availability coefficient;

[0038] The resource library construction unit is used to integrate the real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient and resource availability coefficient of all resource nodes to build a multi-dimensional resource library.

[0039] As a further solution of the present invention: the optimization model establishment module includes:

[0040] A total regulating power unit is used to determine the total regulating power of the virtual power plant based on the load demand of the power grid and the grid interaction power;

[0041] A multi-dimensional target determination unit, which is used to determine an economic target based on the power regulation cost, an environmental target based on the carbon emission coefficient, and a stability target based on the response delay time and the resource availability coefficient;

[0042] The global optimization model unit is used to construct a global optimization model based on the total regulation power, economic objectives, environmental protection objectives and stability objectives.

[0043] As a further solution of the present invention: the optimal scheduling strategy module includes:

[0044] The chromosome encoding unit is used to encode chromosomes. Each gene position corresponds to the adjustment power ΔPi of a resource node, and the chromosome is directly mapped to the three objective function values ​​of economy, environmental protection and stability.

[0045] The population initialization unit is used to initialize the population according to the predetermined number, using stratified sampling: A% individuals are generated by Latin hypercube sampling, B% individuals are generated based on the neighborhood perturbation of the historical optimal solution, and C% individuals are generated using greedy heuristics, A+B+C=100;

[0046] The non-dominated hierarchy unit is used to determine the dominance relationship of each individual and divide the population into multiple non-dominated hierarchies according to the dominance relationship. The first level is the solution that is not dominated by any individual, the second level is the solution that is only dominated by the first level, and so on.

[0047] The crowding determination unit is used to calculate the normalized distances of the three target values ​​for individuals in each level; in each target direction, individuals are sorted by the normalized values, the distances between adjacent individuals in the target space are calculated, and the crowding degree is determined;

[0048] The iterative output unit is used to perform genetic iterative operations on the population according to the non-dominated hierarchy and congestion degree, and output the optimal scheduling strategy.

[0049] As a further solution of the present invention: the edge terminal adjustment module includes:

[0050] A power adjustment unit, configured to proportionally reduce the adjustment power of the resource node when the resource availability coefficient in the real-time state is lower than the availability threshold;

[0051] The instruction sending advance unit is used to send instructions in advance when the response delay time is higher than the delay threshold. The advance time is determined according to the response delay time and the safety margin.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] This invention builds a global optimization model and dynamic collaborative control that integrates economic efficiency, environmental benefits, and system stability. It then uses an improved non-dominated sorting genetic algorithm to determine the optimal scheduling strategy. This highly efficient solution significantly improves the response accuracy and system stability of virtual power plants, providing decision makers with a multi-objective optimization strategy. Furthermore, the edge terminal adaptively adjusts the optimal scheduling strategy based on real-time status, ensuring that scheduling instructions match actual operating conditions for even better results. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 The figure is a flow chart of a collaborative control and scheduling method for a virtual power plant.

[0055] Figure 2 A flowchart for building a multi-dimensional resource library in a collaborative control and scheduling method for a virtual power plant.

[0056] Figure 3 Flowchart for establishing a global optimization model in the cloud for a collaborative control and scheduling method for a virtual power plant.

[0057] Figure 4 Flowchart for obtaining the optimal scheduling strategy in a collaborative control and scheduling method for a virtual power plant.

[0058] Figure 5 This is a flow chart for adjusting the optimal scheduling strategy in a collaborative control and scheduling method for a virtual power plant.

[0059] Figure 6 This is a structural diagram of a virtual power plant collaborative control and scheduling system. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present invention clearer, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] The specific implementation of the present invention is described in detail below with reference to specific embodiments.

[0062] like Figure 1As shown, an embodiment of the present invention provides a virtual power plant collaborative control scheduling method, the method comprising the following steps:

[0063] S100: Build a multi-dimensional resource library. Each resource node in the multi-dimensional resource library corresponds to real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient, and resource availability coefficient.

[0064] S200: All resource nodes are aggregated in the control center to form a resource pool of the virtual power plant;

[0065] S300, establishing a global optimization model in the cloud, wherein the global optimization model integrates economic efficiency, environmental benefits, and system stability;

[0066] S400: Use an improved non-dominated sorting genetic algorithm to solve the global optimization model, obtain the optimal scheduling strategy, and send the optimal scheduling strategy to the control center and edge terminals;

[0067] S500 , determining the real-time status of the resource node based on the edge terminal, and adjusting the optimal scheduling strategy according to the real-time status.

[0068] It's important to note that traditional scheduling methods typically convert multiple objectives into a single objective (e.g., weighted summation), requiring manual weighting. This is highly subjective and difficult to balance. Current static scheduling strategies cannot respond in real time to changes in resource status (such as device failures and communication delays), resulting in significant discrepancies between scheduling instructions and actual execution. Furthermore, traditional multi-objective optimization algorithms (such as standard NSGA-II) are prone to local optima when dealing with high-dimensional problems and are computationally expensive. The embodiments of the present invention aim to address these issues.

[0069] In this embodiment of the present invention, a multidimensional resource library must first be constructed. This library contains a large number of resource nodes, each of which corresponds to real-time adjustable power, power adjustment cost, response delay, carbon emission coefficient, and resource availability coefficient. Real-time adjustable power is determined by the device itself and real-time environmental factors. For example, the real-time adjustable power of photovoltaic power generation is determined by light intensity, photovoltaic panel area, photovoltaic conversion efficiency, and inverter capacity. The adjustable power of energy storage devices is limited by their current state of charge (SOC) and charge and discharge rates. Power adjustment cost comprehensively considers electricity price costs, equipment aging costs, and operation and maintenance costs. For example, the adjustment cost of energy storage devices includes electricity price costs and battery life loss due to charging and discharging; the cost of interruptible loads includes user compensation costs and equipment startup and shutdown losses. The resource availability coefficient is primarily determined by the device status. All resource nodes are then aggregated at the control center to form the virtual power plant resource pool. This aggregation can be categorized: resources are divided into fast-response pools and slow-response pools based on their response delay characteristics, facilitating the prioritization of fast-response resources during scheduling. Geographic aggregation can also be performed: resource nodes within the same region are aggregated into sub-resource pools to reduce communication latency and scheduling complexity. In this embodiment of the present invention, each resource node is deployed with an edge terminal, all of which can be controlled by a control center. The edge terminal itself has data processing and analysis capabilities, and both the edge terminal and the control center are connected to the cloud. In this embodiment of the present invention, a global optimization model is established in the cloud. This global optimization model integrates economic efficiency, environmental benefits, and system stability to minimize scheduling costs and carbon emissions, prioritize low-carbon emission resources, and ensure stable grid operation. Preferably, an improved non-dominated sorting genetic algorithm (NSGA-II) is used to solve the global optimization model and obtain the optimal scheduling strategy. This improves the algorithm's global search capability and convergence speed, avoids being trapped in local optima, and then distributes the optimal scheduling strategy to the control center and edge terminals. In addition, the edge terminal determines the real-time status of the resource node, including availability coefficient and response delay, and adjusts the optimal scheduling strategy based on this real-time status to ensure that scheduling instructions match the actual operating status. The embodiment of the present invention significantly improves the response accuracy and system stability of the virtual power plant by constructing a global optimization model and dynamic collaborative control that integrates economic efficiency, environmental benefits and system stability, and adopts an improved non-dominated sorting genetic algorithm for solution, providing decision makers with a multi-objective trade-off optimization strategy, which is particularly suitable for modern power grid scenarios with a high proportion of renewable energy.

[0070] like Figure 2 As shown, as a preferred embodiment of the present invention, the step of constructing a multi-dimensional resource library specifically includes:

[0071] S101, collecting power generation impact data of each resource node and determining real-time adjustable power based on the power generation impact data;

[0072] S102, determining the power regulation cost, response delay time, and carbon emission coefficient of each resource node based on the resource node type and historical data;

[0073] S103, determining the predicted power, rated power, and power generation equipment status of the resource node, and calculating the resource availability coefficient;

[0074] S104: Integrate the real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient, and resource availability coefficient of all resource nodes to build a multi-dimensional resource library.

[0075] In this embodiment of the present invention, power generation impact data for each resource node is collected. It's easy to understand that the power generation impact data for each resource type is different. Taking photovoltaic power generation as an example, the power generation impact data includes the photoelectric conversion efficiency η, the real-time light intensity I, the photovoltaic panel area S, the inverter rated power P1, and the photovoltaic panel rated power P2. The adjustable power of photovoltaic power generation = η × I × S × min (1, P1 / P2). Then, based on the resource node type and historical data, the power regulation cost, response delay time, and carbon emission coefficient of each resource node are determined. For example, the carbon emission coefficient for photovoltaic power generation is 50 kg CO2 / MWh, and the carbon emission coefficient for diesel power generation is 820 kg CO2 / MWh. Next, the predicted power P3, rated power P2, and power generation equipment status SC of the resource node are determined, and the resource availability coefficient is calculated: resource availability coefficient = P3 / P2 × β × SC, where β is the state constraint coefficient. Finally, integration is performed, and the multi-dimensional resource library is constructed.

[0076] like Figure 3 As shown, as a preferred embodiment of the present invention, the step of establishing a global optimization model in the cloud specifically includes:

[0077] S301, determining the total regulation power of the virtual power plant according to the load demand of the power grid and the power interaction power of the power grid;

[0078] S302, determining an economic target based on the power regulation cost, determining an environmental target based on the carbon emission coefficient, and determining a stability target based on the response delay time and resource availability coefficient;

[0079] S303: Construct a global optimization model based on the total regulated power, economic target, environmental target, and stability target.

[0080] In this embodiment of the present invention, the total regulation power PV of the virtual power plant is determined based on the grid's load demand PL and the current grid interaction power PG, where PV = PL - PG. Then, the economic objective G1 is determined based on the power regulation cost: G1 = Σ(Ci × ΔPi), where Ci is the power regulation cost of the i-th resource node and ΔPi is the regulation power of the i-th resource node. The environmental objective G2 is determined based on the carbon emission coefficient: G2 = Σ(Ei × ΔPi), where Ei is the carbon emission coefficient. Finally, the stability objective G3 is determined based on the response delay time and the resource availability coefficient: G3 = Σ(Ti × |ΔPi|) + Σ(Ai × (1-Ui)), where Ti is the response delay time, Ai is the availability penalty coefficient (a constant), and Ui is the resource availability coefficient (0-1). This allows the construction of a global optimization model.

[0081] like Figure 4 As shown in FIG. 1 , as a preferred embodiment of the present invention, the step of using the improved non-dominated sorting genetic algorithm to solve the global optimization model and obtain the optimal scheduling strategy specifically includes:

[0082] S401, chromosome encoding is performed, where each gene position corresponds to the regulation power ΔPi of a resource node, and the chromosome is directly mapped to the three objective function values ​​of economy, environmental protection, and stability;

[0083] S402, initialize the population according to the predetermined number, using stratified sampling: A% of individuals are generated by Latin hypercube sampling, B% of individuals are generated based on the neighborhood perturbation of the historical optimal solution, and C% of individuals are generated using greedy heuristics, A+B+C=100;

[0084] S403, determining the dominance relationship of each individual, and dividing the population into multiple non-dominated levels based on the dominance relationship. The first level is a solution that is not dominated by any individual, the second level is a solution that is only dominated by the first level, and so on.

[0085] S404: For each individual in each level, the normalized distances of the three target values ​​are calculated. In each target direction, the individuals are sorted by the normalized values, and the distances between adjacent individuals in the target space are calculated to determine the crowding degree.

[0086] S405: Perform genetic iteration operations on the population according to the non-dominated hierarchy and the congestion degree, and output the optimal scheduling strategy.

[0087] In this embodiment of the present invention, chromosome encoding is first performed using real vector encoding, with each gene bit corresponding to the regulation power ΔPi of a resource node. The chromosomes are directly mapped to the values ​​of the three objective functions: economy, environmental protection, and stability. A population is then initialized to a predetermined number (e.g., 200) using stratified sampling: 50% of the individuals are generated using Latin hypercube sampling to ensure uniform coverage of the target space, 30% are generated based on neighborhood perturbations of historical optimal solutions, and 20% are generated using a greedy heuristic, prioritizing nodes with low cost and low carbon emissions. Next, the dominance relationship of each individual is determined, and for each individual in the population, the economic, environmental, and stability objective values ​​are compared. If individual a is not inferior to individual b in all objectives and is superior in at least one objective, then a dominates b. Based on the dominance relationships, the population is divided into multiple non-dominated levels: the first level contains solutions that are completely undominated by any individual, the second level contains solutions dominated only by the first level, and so on. Next, for each individual in each hierarchy, the normalized distances between the three target values ​​are calculated. The target values ​​need to be mapped to the interval [0, 1] to avoid the influence of dimensional differences. Within each target direction, individuals are sorted by their normalized values, and the distances between adjacent individuals in the target space are calculated to determine the degree of congestion. It should be noted that a scheduling strategy is considered non-dominated if it is not inferior to other strategies in terms of economic efficiency, environmental benefits, and system stability, and is superior in at least one of the three objectives.

[0088] Next, genetic iterations of the population are performed based on the non-dominated hierarchy and crowding degree to output the optimal scheduling strategy. Specifically, several individuals are randomly selected from the current population, their non-dominated hierarchy and crowding degree are compared, and the best individuals are selected for the next generation. A crossover operation is then performed on the selected parent individuals to generate offspring individuals. This simulates the crossover method of binary coding, achieving local search capabilities in real-number coding. Mutation is performed on the offspring individuals, randomly adjusting gene values ​​(the regulation power ΔPi of the resource node) with a certain probability to enhance population diversity. The parent and offspring populations are merged to form a temporary population. The non-dominated sorting and crowding degree calculations are performed again on the temporary population, and the optimal N individuals are selected to form the next generation population (N is the preset population size). The process stops when the maximum number of iterations is reached. Individuals with a non-dominated hierarchy of 1 are then extracted from the final population to form the Pareto front set of solutions (the set of non-dominated solutions). Based on actual needs (e.g., prioritizing economic efficiency or environmental performance), the most preferred solution from the Pareto front is selected as the optimal scheduling strategy.

[0089] like Figure 5 As shown, as a preferred embodiment of the present invention, the step of adjusting the optimal scheduling strategy according to the real-time status specifically includes:

[0090] S501, when the resource availability coefficient in the real-time state is lower than the availability threshold, proportionally reduce the regulation power of the resource node;

[0091] S502: When the response delay time is higher than the delay threshold, the instruction is issued in advance, and the advance time is determined according to the response delay time and the safety margin.

[0092] In this embodiment of the present invention, if the resource availability coefficient falls below a threshold (e.g., 0.6), the regulated power is proportionally reduced. For example, if the originally scheduled PV capacity is reduced by 1 MW, and the availability drops to 0.5, the actual reduction is 0.5 MW. Furthermore, if the response delay exceeds the delay threshold, indicating a significant delay, the command needs to be issued in advance. The lead time equals the response delay time + a safety margin, where the safety margin is a fixed value, for example, 2 seconds.

[0093] like Figure 6 As shown, an embodiment of the present invention further provides a virtual power plant collaborative control and scheduling system, the system comprising:

[0094] The resource library construction module 100 is used to construct a multi-dimensional resource library. Each resource node in the multi-dimensional resource library corresponds to real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient, and resource availability coefficient;

[0095] The resource node aggregation module 200 is used to aggregate all resource nodes in the control center to form a resource pool of the virtual power plant;

[0096] An optimization model building module 300 is used to build a global optimization model in the cloud, wherein the global optimization model integrates economic efficiency, environmental benefits, and system stability;

[0097] The optimal scheduling strategy module 400 is used to solve the global optimization model using an improved non-dominated sorting genetic algorithm to obtain the optimal scheduling strategy, and then send the optimal scheduling strategy to the control center and edge terminals;

[0098] The edge terminal adjustment module 500 is configured to determine the real-time status of the resource node based on the edge terminal and adjust the optimal scheduling strategy according to the real-time status.

[0099] As a preferred embodiment of the present invention, the resource library construction module 100 includes:

[0100] The adjustable power unit is used to collect power generation impact data of each resource node and determine the real-time adjustable power based on the power generation impact data;

[0101] A historical data analysis unit, configured to determine the power regulation cost, response delay time, and carbon emission coefficient of each resource node based on the resource node type and historical data;

[0102] Resource availability unit, used to determine the predicted power, rated power and power generation equipment status of the resource node, and calculate the resource availability coefficient;

[0103] The resource library construction unit is used to integrate the real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient and resource availability coefficient of all resource nodes to build a multi-dimensional resource library.

[0104] As a preferred embodiment of the present invention, the optimization model building module 300 includes:

[0105] A total regulating power unit is used to determine the total regulating power of the virtual power plant based on the load demand of the power grid and the grid interaction power;

[0106] A multi-dimensional target determination unit, which is used to determine an economic target based on the power regulation cost, an environmental target based on the carbon emission coefficient, and a stability target based on the response delay time and the resource availability coefficient;

[0107] The global optimization model unit is used to construct a global optimization model based on the total regulation power, economic objectives, environmental protection objectives and stability objectives.

[0108] As a preferred embodiment of the present invention, the optimal scheduling strategy module 400 includes:

[0109] The chromosome encoding unit is used to encode chromosomes. Each gene position corresponds to the adjustment power ΔPi of a resource node, and the chromosome is directly mapped to the three objective function values ​​of economy, environmental protection and stability.

[0110] The population initialization unit is used to initialize the population according to the predetermined number, using stratified sampling: A% individuals are generated by Latin hypercube sampling, B% individuals are generated based on the neighborhood perturbation of the historical optimal solution, and C% individuals are generated using greedy heuristics, A+B+C=100;

[0111] The non-dominated hierarchy unit is used to determine the dominance relationship of each individual and divide the population into multiple non-dominated hierarchies according to the dominance relationship. The first level is the solution that is not dominated by any individual, the second level is the solution that is only dominated by the first level, and so on.

[0112] The crowding determination unit is used to calculate the normalized distances of the three target values ​​for individuals in each level; in each target direction, individuals are sorted by the normalized values, the distances between adjacent individuals in the target space are calculated, and the crowding degree is determined;

[0113] The iterative output unit is used to perform genetic iterative operations on the population according to the non-dominated hierarchy and congestion degree, and output the optimal scheduling strategy.

[0114] As a preferred embodiment of the present invention, the edge terminal adjustment module 500 includes:

[0115] A power adjustment unit, configured to proportionally reduce the adjustment power of the resource node when the resource availability coefficient in the real-time state is lower than the availability threshold;

[0116] The instruction sending advance unit is used to send instructions in advance when the response delay time is higher than the delay threshold. The advance time is determined according to the response delay time and the safety margin.

[0117] The above is only a detailed description of the preferred embodiments of the present invention, which is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0118] It should be understood that, although the various steps in the flow chart of each embodiment of the present invention are shown in sequence according to the indication of the arrows, these steps are not necessarily performed in sequence according to the order indicated by the arrows. Unless otherwise specified herein, the execution of these steps is not strictly limited in order, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0119] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When executed, the program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0120] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered merely as exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A virtual power plant collaborative control and scheduling method, characterized in that: The method comprises the following steps: Build a multi-dimensional resource library, where each resource node corresponds to real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient, and resource availability coefficient; Aggregate all resource nodes in the control center to form a resource pool for the virtual power plant; Establishing a global optimization model in the cloud, wherein the global optimization model integrates economic efficiency, environmental benefits and system stability; An improved non-dominated sorting genetic algorithm is used to solve the global optimization model and obtain the optimal scheduling strategy, which is then distributed to the control center and edge terminals. Determine the real-time status of resource nodes based on edge terminals and adjust the optimal scheduling strategy based on the real-time status; The steps of establishing a global optimization model in the cloud specifically include: determining the total regulation power of the virtual power plant based on the load demand of the power grid and the grid interaction power; determining the economic target based on the power regulation cost, determining the environmental target based on the carbon emission coefficient, and determining the stability target based on the response delay time and the resource availability coefficient; and constructing a global optimization model based on the total regulation power, economic target, environmental target, and stability target. Among them, the steps of using the improved non-dominated sorting genetic algorithm to solve the global optimization model and obtain the optimal scheduling strategy specifically include: performing chromosome encoding, where each gene position corresponds to the adjustment power ΔPi of a resource node, and directly mapping the chromosome to the three objective function values ​​of economy, environmental protection, and stability; initializing the population according to a predetermined number, and using stratified sampling: A% individuals are generated through Latin hypercube sampling, B% individuals are generated based on the neighborhood perturbation of the historical optimal solution, and C% individuals are generated using a greedy heuristic, A+B+C=100; determining the dominance relationship of each individual, and dividing the population into multiple non-dominated levels according to the dominance relationship, the first level is a solution that is completely not dominated by any individual, the second level is a solution that is only dominated by the first level, and so on; for individuals in each level, calculating the normalized distances of the three target values ​​respectively; in each target direction, sorting the individuals according to the normalized values, calculating the distances between adjacent individuals in the target space, and determining the congestion degree; performing genetic iterative operations on the population according to the non-dominated level and the congestion degree, and outputting the optimal scheduling strategy.

2. The virtual power plant collaborative control and scheduling method according to claim 1, characterized in that: The steps of constructing a multi-dimensional resource library specifically include: Collect power generation impact data of each resource node and determine real-time adjustable power based on the power generation impact data; Determine the power regulation cost, response delay time, and carbon emission coefficient of each resource node based on the resource node type and historical data; Determine the predicted power, rated power, and power generation equipment status of the resource node, and calculate the resource availability coefficient; The real-time adjustable power, power regulation cost, response delay time, carbon emission coefficient and resource availability coefficient of all resource nodes are integrated to build a multi-dimensional resource library.

3. The virtual power plant collaborative control and scheduling method according to claim 1, characterized in that: The step of adjusting the optimal scheduling strategy according to the real-time status specifically includes: When the resource availability coefficient in the real-time state is lower than the availability threshold, the regulation power of the resource node is proportionally reduced; When the response delay time is higher than the delay threshold, the instruction is issued in advance, and the advance time is determined according to the response delay time and the safety margin.

4. A virtual power plant collaborative control and dispatching system, characterized in that: The system comprises: The resource library construction module is used to build a multi-dimensional resource library. Each resource node in the multi-dimensional resource library corresponds to real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient, and resource availability coefficient; The resource node aggregation module is used to aggregate all resource nodes in the control center to form a resource pool for the virtual power plant; An optimization model building module, used to build a global optimization model in the cloud, wherein the global optimization model integrates economic efficiency, environmental benefits and system stability; The optimal scheduling strategy module is used to solve the global optimization model using an improved non-dominated sorting genetic algorithm to obtain the optimal scheduling strategy, and then send the optimal scheduling strategy to the control center and edge terminals; The edge terminal adjustment module is used to determine the real-time status of resource nodes based on the edge terminal and adjust the optimal scheduling strategy according to the real-time status; The optimization model establishment module includes: a total regulation power unit, which is used to determine the total regulation power of the virtual power plant according to the load demand of the power grid and the grid interaction power; a multi-dimensional target determination unit, which is used to determine the economic target according to the power regulation cost, the environmental target according to the carbon emission coefficient, and the stability target according to the response delay time and the resource availability coefficient; a global optimization model unit, which is used to construct a global optimization model based on the total regulation power, economic target, environmental target and stability target; The optimal scheduling strategy module includes: a chromosome encoding unit for chromosome encoding, where each gene bit corresponds to the regulation power ΔPi of a resource node, and the chromosome is directly mapped to the three objective function values ​​of economy, environmental protection, and stability; a population initialization unit for initializing the population according to a predetermined number, using stratified sampling: A% individuals are generated through Latin hypercube sampling, B% individuals are generated based on the neighborhood perturbation of the historical optimal solution, and C% individuals are generated using a greedy heuristic, where A+B+C=100; a non-dominated hierarchy unit for determining the dominance relationship of each individual and dividing the population into multiple non-dominated hierarchies based on the dominance relationship, where the first hierarchy is a solution that is completely not dominated by any individual, the second hierarchy is a solution that is only dominated by the first hierarchy, and so on; a crowding degree determination unit for calculating the normalized distances of the three target values ​​for individuals in each hierarchy; in each target direction, individuals are sorted according to the normalized values, the distances between adjacent individuals in the target space are calculated, and the crowding degree is determined; and an iterative output unit for performing genetic iterative operations on the population based on the non-dominated hierarchy and crowding degree, and outputting the optimal scheduling strategy.

5. The virtual power plant collaborative control and dispatching system according to claim 4, characterized in that: The resource library construction module includes: The adjustable power unit is used to collect power generation impact data of each resource node and determine the real-time adjustable power based on the power generation impact data; A historical data analysis unit, configured to determine the power regulation cost, response delay time, and carbon emission coefficient of each resource node based on the resource node type and historical data; Resource availability unit, used to determine the predicted power, rated power and power generation equipment status of the resource node, and calculate the resource availability coefficient; The resource library construction unit is used to integrate the real-time adjustable power, power adjustment cost, response delay time, carbon emission coefficient and resource availability coefficient of all resource nodes to build a multi-dimensional resource library.

6. The virtual power plant collaborative control and dispatching system according to claim 4, characterized in that: The edge terminal adjustment module includes: A power adjustment unit, configured to proportionally reduce the adjustment power of the resource node when the resource availability coefficient in the real-time state is lower than the availability threshold; The instruction sending advance unit is used to send instructions in advance when the response delay time is higher than the delay threshold. The advance time is determined according to the response delay time and the safety margin.

Images