Virtual power plant optimization scheduling method and system

Abstract

The invention relates to the technical field of virtual power plant scheduling, and discloses a virtual power plant optimal scheduling method and system, and the method comprises the steps: selecting a to-be-optimized and evaluated target resource system, obtaining the real-time operation data of the target resource system, and constructing a plurality of operation data sequences; performing state analysis and fluctuation analysis on the operation data sequence, determining a state coefficient and a fluctuation coefficient, and calculating a to-be-optimized coefficient in combination with a corresponding weight coefficient; performing classification according to the to-be-optimized coefficients of all the resource systems to obtain a first-class system set and a second-class system set, and generating a first optimization scheduling strategy based on a preset optimization scheduling model; and performing analogue simulation on the first optimal scheduling strategy, judging whether an optimal scheduling target is met or not, if so, issuing a scheduling instruction, and if not, generating a second optimal scheduling strategy, accurately quantifying the operation state and the optimization demand of the resource system, and improving the rationality of the optimal scheduling strategy and the operation efficiency of the virtual power plant.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of virtual power plant scheduling, in particular to a virtual power plant optimization scheduling method and system. BACKGROUND

[0002] As the core carrier of integrating distributed energy, controllable load and energy storage resources, the virtual power plant (VPP) can effectively improve the energy utilization efficiency and suppress the output fluctuation by the "aggregation-coordination-optimization" mechanism, and become one of the key supporting technologies for the construction of new power systems.

[0003] In the prior art, the virtual power plant optimization scheduling technology relies on historical data or static models for resource evaluation, resulting in poor accuracy of the running state judgment of the resource system, and the actual optimization demand of the resource system cannot be reflected, which reduces the rationality of the optimization scheduling strategy and the efficient and stable operation of the virtual power plant. SUMMARY

[0004] To solve the above technical problems, the present application provides a virtual power plant optimization scheduling method and system, which calculates the state coefficient and the fluctuation coefficient of each resource system to obtain the to-be-optimized coefficient of each resource system, classifies them according to the to-be-optimized coefficient to obtain a first system set and a second system set, and combines a preset optimization scheduling model to obtain a first optimization scheduling strategy and simulate it, judges whether the optimization scheduling target is met according to the simulation result, issues a scheduling instruction if it is met, and generates a second optimization scheduling strategy if it is not met, accurately quantifies the running state and optimization demand of the resource system, provides a scientific basis for the formulation of the optimization scheduling strategy, and improves the rationality of the optimization scheduling strategy and the operation efficiency of the virtual power plant.

[0005] In some embodiments of the present application, a virtual power plant optimization scheduling method is provided, which comprises: selecting a target resource system to be optimized and evaluated, obtaining real-time running data of the target resource system, and constructing a plurality of running data sequences; performing state analysis and fluctuation analysis on each running data sequence to determine the state coefficient and the fluctuation coefficient of each real-time running data, and combining the weight coefficient of the corresponding real-time running data to calculate the to-be-optimized coefficient of the corresponding target resource system; classifying the to-be-optimized coefficients of all resource systems to obtain a first system set and a second system set, and generating a first optimization scheduling strategy of the real-time running demand, the first system set and the second system set based on a preset optimization scheduling model; Simulate the first optimization scheduling strategy, judge whether the optimization scheduling target is met according to the simulation result, if yes, issue the scheduling instruction, if not, generate the second optimization scheduling strategy.

[0006] In some embodiments of the present application, before acquiring the real-time running data of the target resource system, further comprising: acquiring historical running demand of a previous historical monitoring period; performing similarity analysis on the historical running demand and the real-time running demand of the current monitoring period to obtain a similarity degree; if the similarity degree is greater than a preset similarity threshold, extracting a historical fluctuation coefficient of the target resource system in the previous historical monitoring period, and setting a monitoring time interval of the current monitoring period according to the historical fluctuation coefficient; if the similarity degree is not greater than the preset similarity threshold, traversing the historical monitoring log according to the real-time running demand to obtain a similar historical monitoring log; extracting a plurality of historical fluctuation coefficients of the target resource system in the similar historical monitoring log, and performing mean value processing, and setting a monitoring time interval of the target resource system in the current monitoring period according to the historical fluctuation coefficients after mean value processing; generating a plurality of monitoring time nodes of the current monitoring period according to the monitoring time interval, and collecting real-time running data of the target resource system according to the monitoring time nodes.

[0007] In some embodiments of the present application, the state coefficient and the fluctuation coefficient of each real-time running data are determined, comprising: setting a standard running data interval at each monitoring time node according to the real-time running demand; comparing the real-time running data at each monitoring time node in the running data sequence with the corresponding standard running data interval, if it is in the corresponding standard running data interval, then calculating the state sub-coefficient at the corresponding monitoring time node according to the first feature critical value between the real-time running data and the corresponding standard running data interval; if it is not in the corresponding standard running data interval, then calculating the state sub-coefficient at the corresponding monitoring time node according to the second feature critical value between the real-time running data and the corresponding standard running data interval; arranging the state sub-coefficients in the order of the monitoring time nodes to obtain a state sub-coefficient sequence of each real-time running data; calculating the state coefficient of the corresponding real-time running data according to each state sub-coefficient in the state sub-coefficient sequence and the weight coefficient of the corresponding monitoring time node; differencing each state sub-coefficient in the state sub-coefficient sequence with the adjacent state sub-coefficients before and after it, and generating a fluctuation sub-coefficient of each state sub-coefficient according to the difference result; The volatile coefficients are classified, and the volatile coefficients of the corresponding real-time running data are calculated based on the classification results.

[0008] In some embodiments of this application, the fluctuation coefficient of the corresponding real-time running data is calculated based on the classification results, including: Pre-set the threshold for the wavelet coefficient; Volatile coefficients below the volatile coefficient threshold are classified as normal volatile coefficients, and the degree of normality of each normal volatile coefficient is calculated. Volatile coefficients that are not less than the volatile coefficient threshold are set as volatile coefficients of the anomalous type, and the degree of anomalousness of each volatile coefficient of the anomalous type is calculated. Calculate the first duration of the continuous fluctuation coefficients for the normal type and the second duration of the continuous fluctuation coefficients for the abnormal type; The volatility coefficient is calculated based on the number of normal type volatility coefficients, the corresponding degree of normality, the first duration, the number of abnormal type volatility coefficients, the corresponding degree of abnormality, and the second duration.

[0009] In some embodiments of this application, calculating the optimization coefficients of the corresponding target resource system includes: The compensation coefficient is generated based on the fluctuation coefficient to produce the state coefficient corresponding to the real-time operating data. The coefficients to be optimized are generated based on the state coefficients, corresponding compensation coefficients, and weight coefficients of each real-time running data of the target resource system.

[0010] In some embodiments of this application, the systems are classified according to the optimization coefficients of all resource systems to obtain a first set of systems and a second set of systems, including: Generate the optimization coefficients for each resource system sequentially; Pre-set the threshold values ​​of the optimization coefficients for each resource system; The optimization coefficient of each resource system is compared with the corresponding optimization coefficient threshold. Resource systems with optimization coefficients less than the optimization coefficient threshold are set as the first type of system, and a set of the first type of system is constructed. Each system in the first category of the system set is mapped to a first difference in coefficients to be optimized and a number of operational data to be optimized. Resource systems whose optimization coefficient is not less than the optimization coefficient threshold are designated as second-type coefficients, and a set of second-type systems is constructed. Each system in the second category of the system set is mapped to a second difference in coefficients to be optimized and a number of preferred operating data.

[0011] In some embodiments of this application, a first optimized scheduling strategy is generated based on a preset optimized scheduling model to determine real-time operating requirements, a first set of systems, and a second set of systems, including: Obtain several historical optimized scheduling logs, and extract the historical operational requirements, historical first-type system set, historical second-type system set, and corresponding historical optimized scheduling strategy from each historical optimized scheduling log; The historical first-class system set includes several historical first-class systems, and each historical first-class system is mapped with a historical first-to-optimize coefficient difference and several historical to-optimize operating data. The set of historical second-class systems includes several historical second-class systems, and each historical second-class system is mapped with the difference of the historical second-class system to be optimized and several historical preferred operating data. Based on the historical operational requirements, historical first-class system set, and historical second-class system set in each historical optimized scheduling log as training input data, and the corresponding historical optimized scheduling strategy as training output data, a neural network is trained to obtain a preset optimized scheduling model. The real-time operational requirements, the first set of systems, and the second set of systems in the current monitoring period are input into the preset optimization scheduling model to obtain the first optimization scheduling strategy.

[0012] In some embodiments of this application, determining whether the optimized scheduling objective is met based on simulation results includes: The optimized scheduling objectives include user demand objectives, power supply reliability objectives, economic benefit objectives, and carbon emission objectives, and a first value threshold for the user demand objective, a second value threshold for the power supply reliability objective, a third value threshold for the economic benefit objective, and a fourth value threshold for the carbon emission objective are determined. A simulation operation scenario is constructed based on the structural information and real-time operation data of each resource system in the current monitoring period; The first optimized scheduling strategy was simulated according to the simulation operation scenario to obtain the simulation operation data of each resource system. The simulation operation data of all resource systems are correlated with each target. Based on the analysis results, a simulation demand dataset for user demand target, a simulation power supply dataset for power supply reliability target, a simulation revenue dataset for economic benefit target, and a simulation carbon emission dataset for carbon emission target are constructed, and the corresponding simulation demand value, simulation reliability value, simulation revenue value, and simulation carbon emission value are determined. The simulated demand value, simulated reliability value, simulated revenue value, and simulated carbon emission value are compared with the corresponding first value threshold, second value threshold, third value threshold, and fourth value threshold, respectively. If all of them are greater than the corresponding value threshold, the optimization scheduling objective is satisfied; if any of them are not greater than the corresponding value threshold, the optimization scheduling objective is not satisfied.

[0013] In some embodiments of this application, generating a second optimized scheduling strategy includes: When the judgment result is that the optimization scheduling objective is not met, the type of optimization scheduling objective that is not met is obtained. The objective type includes user demand objective, power supply reliability objective, economic benefit objective and carbon emission objective. Based on the unmet optimization scheduling target type, historical optimization scheduling logs with similar unmet target types are selected from the historical optimization scheduling logs and used as reference optimization scheduling logs. The historical optimized scheduling strategies in the reference optimized scheduling log are parsed to extract optimized scheduling strategy fragments that do not meet the target type; Based on the real-time operational requirements of the current monitoring cycle, the first type of system set, the second type of system set, and the specific circumstances that do not meet the target type, the extracted optimized scheduling strategy fragments are adaptively adjusted. The adjusted optimized scheduling strategy fragment is merged with the part of the first optimized scheduling strategy that meets the target type to obtain the second optimized scheduling strategy.

[0014] In some embodiments of this application, a virtual power plant optimized scheduling system is also included: The acquisition module is used to select the target resource system to be optimized and evaluated, acquire the real-time operating data of the target resource system, and construct several operating data sequences; The analysis module is used to perform state analysis and fluctuation analysis on each running data sequence, determine the state coefficient and fluctuation coefficient of each real-time running data, and calculate the optimization coefficient of the corresponding target resource system by combining the weight coefficient of the corresponding real-time running data. The generation module is used to classify all resource systems according to the optimization coefficients to obtain a first set of systems and a second set of systems. Based on the preset optimization scheduling model, it generates real-time operation requirements, a first set of systems, and a first optimization scheduling strategy for the second set of systems. The scheduling module is used to simulate the first optimized scheduling strategy, determine whether the optimization scheduling objective is met based on the simulation results, and if so, issue a scheduling instruction; otherwise, generate a second optimized scheduling strategy.

[0015] The virtual power plant optimization scheduling method and system of this application have the following advantages compared with the prior art: By calculating the state coefficient and fluctuation coefficient of each resource system, the optimization coefficient of each resource system is obtained. The systems are then classified according to the optimization coefficient to obtain a first set of systems and a second set of systems. Combined with a preset optimization scheduling model, a first optimization scheduling strategy is obtained and simulated. The simulation results are used to determine whether the optimization scheduling objective is met. If it is met, a scheduling instruction is issued; otherwise, a second optimization scheduling strategy is generated. This process accurately quantifies the operating status and optimization requirements of the resource systems, providing a scientific basis for the formulation of optimization scheduling strategies and improving the rationality of optimization scheduling strategies and the operating efficiency of the virtual power plant. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a virtual power plant optimization scheduling method in an embodiment of this application; Figure 2 This is a schematic diagram of a virtual power plant optimization scheduling system in an embodiment of this application. Detailed Implementation

[0017] The specific embodiments of this application will be described in further detail below with reference to the accompanying drawings and examples. The following examples are used to illustrate this application, but are not intended to limit the scope of this application.

[0018] In the description of this application, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0019] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0020] In the description of this application, it should be noted that, unless otherwise expressly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0021] like Figure 1 As shown in the figure, a virtual power plant optimization scheduling method according to an embodiment of this application includes: S101: Select the target resource system to be optimized and evaluated, obtain the real-time operating data of the target resource system, and construct several operating data sequences; S102: Perform state analysis and fluctuation analysis on each running data sequence, determine the state coefficient and fluctuation coefficient of each real-time running data, and calculate the optimization coefficient of the corresponding target resource system in combination with the weight coefficient of the corresponding real-time running data. S103: Classify all resource systems according to the optimization coefficients to obtain a first set of systems and a second set of systems. Generate a first optimization scheduling strategy for real-time operation requirements, the first set of systems, and the second set of systems based on a preset optimization scheduling model. S104: Simulate the first optimized scheduling strategy, and determine whether the optimization scheduling objective is met based on the simulation results. If yes, issue a scheduling instruction; otherwise, generate the second optimized scheduling strategy.

[0022] In this embodiment, several resource systems in the virtual power plant are identified, including but not limited to distributed energy, energy storage devices, and controllable loads.

[0023] In some embodiments of this application, before obtaining the real-time operating data of the target resource system, the method further includes: Obtain historical operational requirements from the previous monitoring period; A similarity analysis is performed between historical operational requirements and real-time operational requirements in the current monitoring period to obtain the similarity score. If the similarity is greater than the preset similarity threshold, extract the historical fluctuation coefficient of the current target resource system in the previous historical monitoring period, and set the monitoring time interval of the current monitoring period according to the historical fluctuation coefficient. If the similarity is not greater than the preset similarity threshold, the historical monitoring logs are traversed according to the real-time operation requirements to obtain similar historical monitoring logs; Extract several historical fluctuation coefficients of the target resource system from similar historical monitoring logs, perform mean processing, and set the monitoring time interval of the target resource system in the current monitoring period according to the mean-processed historical fluctuation coefficients; Several monitoring time nodes for the current monitoring period are generated according to the monitoring time interval, and real-time operation data of the target resource system are collected according to the monitoring time nodes.

[0024] In this embodiment, the similarity between real-time and historical operating demands is calculated by comparing data indicators from multiple dimensions, including power output patterns, load fluctuation characteristics, and energy supply stability. The data indicators of real-time and historical operating demands are converted into quantifiable feature vectors. Then, a specific similarity algorithm, such as cosine similarity or Euclidean distance, is used to calculate the similarity between the two feature vectors, resulting in a similarity value between 0 and 1. The closer the value is to 1, the higher the similarity between real-time and historical operating demands; conversely, the lower the similarity, the lower the similarity. This method accurately measures the similarity between real-time and historical operating demands, providing crucial information for subsequent optimized scheduling strategies.

[0025] In this embodiment, the larger the historical fluctuation coefficient, the shorter the monitoring time interval, i.e., the more monitoring time nodes there are; conversely, the smaller the historical fluctuation coefficient, the longer the monitoring time interval, i.e., the fewer monitoring time nodes there are. This enables dynamic adjustment of the monitoring frequency to more accurately capture the operational changes of the target resource system.

[0026] In some embodiments of this application, determining the state coefficient and fluctuation coefficient of each real-time running data includes: Set the standard operating data range for each monitoring time node according to real-time operating requirements; The real-time operating data at each monitoring time node in the operating data sequence is compared with the corresponding standard operating data interval. If it is within the corresponding standard operating data interval, the state sub-coefficient at the corresponding monitoring time node is calculated based on the first characteristic critical value of the real-time operating data and the corresponding standard operating data interval. If it is not within the corresponding standard operating data range, the state sub-coefficient at the corresponding monitoring time node is calculated based on the second characteristic critical value of the real-time operating data and the corresponding standard operating data range. The state sub-coefficients are arranged according to the time sequence of the monitoring time nodes to obtain the state sub-coefficient sequence for each real-time running data. The state coefficient of the corresponding real-time running data is calculated based on each state coefficient in the state coefficient sequence and the weight coefficient of the corresponding monitoring time node. The wave coefficient of each state coefficient in the state coefficient sequence is subtracted from the state coefficients before and after it, and the wave coefficient of each state coefficient is generated based on the subtraction result. The volatile coefficients are classified, and the volatile coefficients of the corresponding real-time running data are calculated based on the classification results.

[0027] In this embodiment, a final standard operating data range is set according to real-time operating requirements. Combined with the number of monitoring time nodes and the requirements for stable data changes, the final standard operating data range is divided into several standard operating data ranges and mapped to the corresponding monitoring time nodes, ensuring that each monitoring time node has a clear and suitable standard operating data range as a comparison benchmark.

[0028] In this embodiment, the first characteristic threshold value refers to the maximum value of the standard operating data range, and the second characteristic threshold value refers to the threshold value that is closest to the real-time operating data in the standard operating data range.

[0029] In this embodiment, when calculating the state sub-coefficient, the closer the real-time running data is to the first characteristic critical value, the larger the state sub-coefficient is; if it is not in the corresponding standard running data range, the further the real-time running data is from the second characteristic critical value, the smaller the state sub-coefficient is.

[0030] In this embodiment, the larger the average difference between each state coefficient and its adjacent state coefficients, the larger the wavelet coefficient is, and vice versa. The wavelet coefficient ranges from 0 to 1.

[0031] In this embodiment, by calculating the state sub-coefficient and the corresponding fluctuation sub-coefficient at each monitoring time node, the operating status of the real-time operating data at each monitoring time node is accurately evaluated to determine whether the operation meets the operating requirements and whether the operation is stable. This provides a reliable basis for the subsequent optimization of the scheduling strategy and improves the operating efficiency and stability of the entire virtual power plant.

[0032] In some embodiments of this application, the fluctuation coefficient of the corresponding real-time running data is calculated based on the classification results, including: Pre-set the threshold for the wavelet coefficient; Volatile coefficients below the volatile coefficient threshold are classified as normal volatile coefficients, and the degree of normality of each normal volatile coefficient is calculated. Volatile coefficients that are not less than the volatile coefficient threshold are set as volatile coefficients of the anomalous type, and the degree of anomalousness of each volatile coefficient of the anomalous type is calculated. Calculate the first duration of the continuous fluctuation coefficients for the normal type and the second duration of the continuous fluctuation coefficients for the abnormal type; The volatility coefficient is calculated based on the number of normal type volatility coefficients, the corresponding degree of normality, the first duration, the number of abnormal type volatility coefficients, the corresponding degree of abnormality, and the second duration.

[0033] In this embodiment, the normality level is set based on the difference between the wavelet coefficient threshold and the wavelet coefficient. The larger the difference, the greater the normality level, and vice versa. The abnormality level is set based on the difference between the wavelet coefficient and the wavelet coefficient threshold. The larger the difference, the greater the abnormality level, and vice versa.

[0034] In this embodiment, the fluctuation coefficient threshold is set based on the maximum fluctuation data corresponding to the fluctuation of the state coefficient when it is in a normal state, in order to effectively distinguish between normal fluctuations and abnormal fluctuations in real-time running data.

[0035] In this embodiment, the more normal type fluctuation coefficients there are, the greater the degree of normality and the longer the first duration, the fewer abnormal type fluctuation coefficients there are, the smaller the corresponding degree of abnormality and the shorter the second duration, the smaller the fluctuation coefficient, and vice versa.

[0036] In this embodiment, by calculating the state coefficient and fluctuation coefficient of each real-time operating data, the monitoring accuracy of each real-time operating data is improved, laying the foundation for determining the optimization coefficient of each resource system. Based on the state coefficient and fluctuation coefficient, combined with the weight coefficient of the corresponding real-time operating data, the optimization coefficient is calculated, which can comprehensively and objectively reflect the operating status and optimization needs of the target resource system, thereby improving the accuracy and reliability of the entire virtual power plant optimization scheduling.

[0037] In some embodiments of this application, calculating the optimization coefficients of the corresponding target resource system includes: The compensation coefficient is generated based on the fluctuation coefficient to produce the state coefficient corresponding to the real-time operating data. The coefficients to be optimized are generated based on the state coefficients, corresponding compensation coefficients, and weight coefficients of each real-time running data of the target resource system.

[0038] In this embodiment, the larger the fluctuation coefficient, the smaller the compensation coefficient, and vice versa. The range of the compensation coefficient is (0.8, 1.2).

[0039] In this embodiment, the accuracy of the state coefficient is improved by calculating the fluctuation coefficient and generating the compensation coefficient. When the state coefficient is larger, the compensation coefficient is larger, and the corresponding coefficient to be optimized is larger. Conversely, the smaller the coefficient to be optimized, the worse the operating state of the corresponding resource system is, and thus optimization scheduling is required. Conversely, the operating state is good.

[0040] In some embodiments of this application, the systems are classified according to the optimization coefficients of all resource systems to obtain a first set of systems and a second set of systems, including: Generate the optimization coefficients for each resource system sequentially; Pre-set the threshold values ​​of the optimization coefficients for each resource system; The optimization coefficient of each resource system is compared with the corresponding optimization coefficient threshold. Resource systems with optimization coefficients less than the optimization coefficient threshold are set as the first type of system, and a set of the first type of system is constructed. Each system in the first category of the system set is mapped to a first difference in coefficients to be optimized and a number of operational data to be optimized. Resource systems whose optimization coefficient is not less than the optimization coefficient threshold are designated as second-type coefficients, and a set of second-type systems is constructed. Each system in the second category of the system set is mapped to a second difference in coefficients to be optimized and a number of preferred operating data.

[0041] In this embodiment, the threshold for the coefficient to be optimized refers to the reference value of the coefficient to be optimized when each resource system is running normally. This reference value can be flexibly set according to the type of resource system, historical operating data and actual operating needs. By setting the threshold for the coefficient to be optimized, resource systems that need to be optimized and scheduled can be accurately distinguished from resource systems with good operating status, providing a clear classification basis for subsequent optimization scheduling strategies.

[0042] In this embodiment, preferred operating data refers to real-time operating data where the compensated state coefficient is greater than the corresponding state coefficient threshold, laying the foundation for the subsequent generation of the first optimized scheduling strategy.

[0043] In this embodiment, the first difference of the coefficient to be optimized refers to the threshold of the coefficient to be optimized minus the coefficient to be optimized, the second difference of the coefficient to be optimized refers to the coefficient to be optimized minus the threshold of the coefficient to be optimized, and the running data to be optimized refers to the real-time running data where the compensated state coefficient is not greater than the corresponding state coefficient threshold, that is, the real-time running data that needs to be optimized and scheduled.

[0044] In this embodiment, by determining the first type of system set and the second type of system set, the resource systems that need to be optimized, the resource systems that are operating well, and the corresponding operating data of the corresponding resources in the current monitoring period are identified. Combined with the preset optimization scheduling model, the first optimization scheduling strategy is obtained, which improves the pertinence and effectiveness of the optimization scheduling strategy, forms an efficient and accurate optimization scheduling system, and provides strong support for the intelligent management of virtual power plants.

[0045] In some embodiments of this application, a first optimized scheduling strategy is generated based on a preset optimized scheduling model to determine real-time operating requirements, a first set of systems, and a second set of systems, including: Obtain several historical optimized scheduling logs, and extract the historical operational requirements, historical first-type system set, historical second-type system set, and corresponding historical optimized scheduling strategy from each historical optimized scheduling log; The historical first-class system set includes several historical first-class systems, and each historical first-class system is mapped with a historical first-to-optimize coefficient difference and several historical to-optimize operating data. The set of historical second-class systems includes several historical second-class systems, and each historical second-class system is mapped with the difference of the historical second-class system to be optimized and several historical preferred operating data. Based on the historical operational requirements, historical first-class system set, and historical second-class system set in each historical optimized scheduling log as training input data, and the corresponding historical optimized scheduling strategy as training output data, a neural network is trained to obtain a preset optimized scheduling model. The real-time operational requirements, the first set of systems, and the second set of systems in the current monitoring period are input into the preset optimization scheduling model to obtain the first optimization scheduling strategy.

[0046] In this embodiment, the historical optimized scheduling logs are all optimized scheduling logs that meet the historical operational requirements and optimized scheduling objectives after the historical optimized scheduling strategy.

[0047] In this embodiment, the real-time operation requirements, the first type of system set, and the second type of system set in the current monitoring cycle are input into a pre-trained optimized scheduling model. This allows for the rapid and accurate generation of a first optimized scheduling strategy that adapts to the current situation, thereby improving the overall stability and economy of the virtual power plant operation.

[0048] In some embodiments of this application, determining whether the optimized scheduling objective is met based on simulation results includes: The optimized scheduling objectives include user demand objectives, power supply reliability objectives, economic benefit objectives, and carbon emission objectives, and a first value threshold for the user demand objective, a second value threshold for the power supply reliability objective, a third value threshold for the economic benefit objective, and a fourth value threshold for the carbon emission objective are determined. A simulation operation scenario is constructed based on the structural information and real-time operation data of each resource system in the current monitoring period; The first optimized scheduling strategy was simulated according to the simulation operation scenario to obtain the simulation operation data of each resource system. The simulation operation data of all resource systems are correlated with each target. Based on the analysis results, a simulation demand dataset for user demand target, a simulation power supply dataset for power supply reliability target, a simulation revenue dataset for economic benefit target, and a simulation carbon emission dataset for carbon emission target are constructed, and the corresponding simulation demand value, simulation reliability value, simulation revenue value, and simulation carbon emission value are determined. The simulated demand value, simulated reliability value, simulated revenue value, and simulated carbon emission value are compared with the corresponding first value threshold, second value threshold, third value threshold, and fourth value threshold, respectively. If all of them are greater than the corresponding value threshold, the optimization scheduling objective is satisfied; if any of them are not greater than the corresponding value threshold, the optimization scheduling objective is not satisfied.

[0049] In this embodiment, the first threshold value is obtained by averaging the historical actual demand data of the user's demand target and the corresponding historical satisfaction coefficient. This satisfaction coefficient can be flexibly adjusted according to factors such as the importance of the user and the historical demand satisfaction status. The second threshold value is determined comprehensively based on the historical power supply reliability records of the resource system, the current equipment health status of the resource system, and the preset power supply reliability standards, aiming to ensure the stability and reliability of the virtual power plant's power supply. The third threshold value is dynamically adjusted based on historical economic benefit data, current market electricity price fluctuations, and the operating costs of the resource system to maximize the economic benefits of the virtual power plant. The fourth threshold value considers carbon emission quotas, the carbon emission intensity of the resource system, and the implementation effect of emission reduction technologies, aiming to promote the development of the virtual power plant towards a low-carbon and environmentally friendly direction. By setting these four threshold values, it is possible to comprehensively and objectively evaluate whether the first optimized scheduling strategy meets the optimized scheduling objectives, providing a scientific basis for the subsequent issuance of scheduling instructions or the generation of the second optimized scheduling strategy.

[0050] In this embodiment, calculating the simulation demand value based on the simulation demand dataset refers to a comprehensive calculation of the degree to which each resource system in the simulation demand dataset meets user needs under the simulated operating scenario. This value can intuitively reflect the effectiveness of the current optimized scheduling strategy in meeting user needs. Similarly, the simulation reliability value is calculated based on the simulation power supply dataset, taking into account the power supply stability and reliability of each resource system during simulation operation, reflecting the achievement of the optimized scheduling strategy's power supply reliability target. The simulation benefit value is a value that reflects the economic effect of the optimized scheduling strategy, calculated by analyzing the simulation benefit dataset and combining it with the economic benefit target. The simulation carbon emission value is a quantitative indicator derived by measuring the carbon emissions generated by each resource system during simulation operation based on the simulation carbon emission dataset and comparing it with the carbon emission target. By comparing these four values ​​with their corresponding threshold values, it is possible to comprehensively and accurately determine whether the first optimized scheduling strategy meets the optimized scheduling target.

[0051] In some embodiments of this application, generating a second optimized scheduling strategy includes: When the judgment result is that the optimization scheduling objective is not met, the type of optimization scheduling objective that is not met is obtained. The objective type includes user demand objective, power supply reliability objective, economic benefit objective and carbon emission objective. Based on the unmet optimization scheduling target type, historical optimization scheduling logs with similar unmet target types are selected from the historical optimization scheduling logs and used as reference optimization scheduling logs. The historical optimized scheduling strategies in the reference optimized scheduling log are parsed to extract optimized scheduling strategy fragments that do not meet the target type; Based on the real-time operational requirements of the current monitoring cycle, the first type of system set, the second type of system set, and the specific circumstances that do not meet the target type, the extracted optimized scheduling strategy fragments are adaptively adjusted. The adjusted optimized scheduling strategy fragment is merged with the part of the first optimized scheduling strategy that meets the target type to obtain the second optimized scheduling strategy.

[0052] In this embodiment, by identifying the types of unmet optimization scheduling targets, the direction requiring optimization can be accurately pinpointed, improving the targeting and effectiveness of the optimization scheduling strategy. Filtering historical optimization scheduling logs with similar unmet target types as a reference allows for learning from past successes and reducing the randomness of optimization scheduling. Adaptively adjusting historical optimization scheduling strategy fragments allows them to better adapt to the actual situation of the current monitoring cycle, improving the practicality and operability of the optimization scheduling strategy. Merging the adjusted strategy fragments with the parts that meet the target types yields a second optimization scheduling strategy, realizing dynamic adjustment and optimization of the optimization scheduling strategy, which helps improve the overall operating efficiency and stability of the virtual power plant.

[0053] In some embodiments of this application, such as Figure 2 As shown, it also includes a virtual power plant optimization scheduling system: The acquisition module is used to select the target resource system to be optimized and evaluated, acquire the real-time operating data of the target resource system, and construct several operating data sequences; The analysis module is used to perform state analysis and fluctuation analysis on each running data sequence, determine the state coefficient and fluctuation coefficient of each real-time running data, and calculate the optimization coefficient of the corresponding target resource system by combining the weight coefficient of the corresponding real-time running data. The generation module is used to classify all resource systems according to the optimization coefficients to obtain a first set of systems and a second set of systems. Based on the preset optimization scheduling model, it generates real-time operation requirements, a first set of systems, and a first optimization scheduling strategy for the second set of systems. The scheduling module is used to simulate the first optimized scheduling strategy, determine whether the optimization scheduling objective is met based on the simulation results, and if so, issue a scheduling instruction; otherwise, generate a second optimized scheduling strategy.

[0054] The above description is only a preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of this application, and these improvements and substitutions should also be considered within the scope of protection of this application.

Claims

1. A virtual power plant optimized scheduling method, characterized in that, include: Select the target resource system to be optimized and evaluated, obtain the real-time operating data of the target resource system, and construct several operating data sequences; Perform state analysis and fluctuation analysis on each running data sequence to determine the state coefficient and fluctuation coefficient of each real-time running data, and calculate the optimization coefficient of the corresponding target resource system by combining the weight coefficient of the corresponding real-time running data. Based on the optimization coefficients of all resource systems, a first set of systems and a second set of systems are obtained. Based on the preset optimization scheduling model, a first optimization scheduling strategy is generated for the real-time operation requirements, the first set of systems, and the second set of systems. The first optimized scheduling strategy is simulated, and the simulation results are used to determine whether the optimized scheduling objective is met. If yes, a scheduling instruction is issued; otherwise, a second optimized scheduling strategy is generated.

2. The virtual power plant optimization scheduling method as described in claim 1, characterized in that, Before obtaining the real-time operating data of the target resource system, the method further includes: Obtain historical operational requirements from the previous monitoring period; A similarity analysis is performed between historical operational requirements and real-time operational requirements in the current monitoring period to obtain the similarity score. If the similarity is greater than the preset similarity threshold, extract the historical fluctuation coefficient of the current target resource system in the previous historical monitoring period, and set the monitoring time interval of the current monitoring period according to the historical fluctuation coefficient. If the similarity is not greater than the preset similarity threshold, the historical monitoring logs are traversed according to the real-time operation requirements to obtain similar historical monitoring logs; Extract several historical fluctuation coefficients of the target resource system from similar historical monitoring logs, perform mean processing, and set the monitoring time interval of the target resource system in the current monitoring period according to the mean-processed historical fluctuation coefficients; Several monitoring time nodes for the current monitoring period are generated according to the monitoring time interval, and real-time operation data of the target resource system are collected according to the monitoring time nodes.

3. The virtual power plant optimization scheduling method as described in claim 2, characterized in that, Determine the state coefficients and fluctuation coefficients for each real-time running data point, including: Set the standard operating data range for each monitoring time node according to real-time operating requirements; The real-time operating data at each monitoring time node in the operating data sequence is compared with the corresponding standard operating data interval. If it is within the corresponding standard operating data interval, the state sub-coefficient at the corresponding monitoring time node is calculated based on the first characteristic critical value of the real-time operating data and the corresponding standard operating data interval. If it is not within the corresponding standard operating data range, the state sub-coefficient at the corresponding monitoring time node is calculated based on the second characteristic critical value of the real-time operating data and the corresponding standard operating data range. The state sub-coefficients are arranged according to the time sequence of the monitoring time nodes to obtain the state sub-coefficient sequence for each real-time running data. The state coefficient of the corresponding real-time running data is calculated based on each state coefficient in the state coefficient sequence and the weight coefficient of the corresponding monitoring time node. The wave coefficient of each state coefficient in the state coefficient sequence is subtracted from the state coefficients before and after it, and the wave coefficient of each state coefficient is generated based on the subtraction result. The volatile coefficients are classified, and the volatile coefficients of the corresponding real-time running data are calculated based on the classification results.

4. The virtual power plant optimization scheduling method as described in claim 3, characterized in that, The fluctuation coefficient of the corresponding real-time running data is calculated based on the classification results, including: Pre-set the threshold for the wavelet coefficient; Volatile coefficients below the volatile coefficient threshold are classified as normal volatile coefficients, and the degree of normality of each normal volatile coefficient is calculated. Volatile coefficients that are not less than the volatile coefficient threshold are set as volatile coefficients of the anomalous type, and the degree of anomalousness of each volatile coefficient of the anomalous type is calculated. Calculate the first duration of the continuous fluctuation coefficients for the normal type and the second duration of the continuous fluctuation coefficients for the abnormal type; The volatility coefficient is calculated based on the number of normal type volatility coefficients, the corresponding degree of normality, the first duration, the number of abnormal type volatility coefficients, the corresponding degree of abnormality, and the second duration.

5. The virtual power plant optimization scheduling method as described in claim 4, characterized in that, Calculate the coefficients to be optimized for the corresponding target resource system, including: The compensation coefficient is generated based on the fluctuation coefficient to produce the state coefficient corresponding to the real-time operating data. The coefficients to be optimized are generated based on the state coefficients, corresponding compensation coefficients, and weight coefficients of each real-time running data of the target resource system.

6. The virtual power plant optimization scheduling method as described in claim 5, characterized in that, Based on the optimization coefficients of all resource systems, we obtain a first-class system set and a second-class system set, including: Generate the optimization coefficients for each resource system sequentially; Pre-set the threshold values ​​of the optimization coefficients for each resource system; The optimization coefficient of each resource system is compared with the corresponding optimization coefficient threshold. Resource systems with optimization coefficients less than the optimization coefficient threshold are set as the first type of system, and a set of the first type of system is constructed. Each system in the first category of the system set is mapped to a first difference in coefficients to be optimized and a number of operational data to be optimized. Resource systems whose optimization coefficient is not less than the optimization coefficient threshold are designated as second-type coefficients, and a set of second-type systems is constructed. Each system in the second category of the system set is mapped to a second difference in coefficients to be optimized and a number of preferred operating data.

7. The virtual power plant optimization scheduling method as described in claim 6, characterized in that, Based on a preset optimized scheduling model, a first optimized scheduling strategy is generated for real-time operational requirements, a first set of systems, and a second set of systems, including: Obtain several historical optimized scheduling logs, and extract the historical operational requirements, historical first-type system set, historical second-type system set, and corresponding historical optimized scheduling strategy from each historical optimized scheduling log; The historical first-class system set includes several historical first-class systems, and each historical first-class system is mapped with a historical first-to-optimize coefficient difference and several historical to-optimize operating data. The set of historical second-class systems includes several historical second-class systems, and each historical second-class system is mapped with the difference of the historical second-class system to be optimized and several historical preferred operating data. Based on the historical operational requirements, historical first-class system set, and historical second-class system set in each historical optimized scheduling log as training input data, and the corresponding historical optimized scheduling strategy as training output data, a neural network is trained to obtain a preset optimized scheduling model. The real-time operational requirements, the first set of systems, and the second set of systems in the current monitoring period are input into the preset optimization scheduling model to obtain the first optimization scheduling strategy.

8. The virtual power plant optimization scheduling method as described in claim 7, characterized in that, Determine whether the optimization scheduling objective is met based on the simulation results, including: The optimized scheduling objectives include user demand objectives, power supply reliability objectives, economic benefit objectives, and carbon emission objectives, and a first value threshold for the user demand objective, a second value threshold for the power supply reliability objective, a third value threshold for the economic benefit objective, and a fourth value threshold for the carbon emission objective are determined. A simulation operation scenario is constructed based on the structural information and real-time operation data of each resource system in the current monitoring period; The first optimized scheduling strategy was simulated according to the simulation operation scenario to obtain the simulation operation data of each resource system. The simulation operation data of all resource systems are correlated with each target. Based on the analysis results, a simulation demand dataset for user demand target, a simulation power supply dataset for power supply reliability target, a simulation revenue dataset for economic benefit target, and a simulation carbon emission dataset for carbon emission target are constructed, and the corresponding simulation demand value, simulation reliability value, simulation revenue value, and simulation carbon emission value are determined. The simulated demand value, simulated reliability value, simulated revenue value, and simulated carbon emission value are compared with the corresponding first value threshold, second value threshold, third value threshold, and fourth value threshold, respectively. If all of them are greater than the corresponding value threshold, the optimization scheduling objective is satisfied; if any of them are not greater than the corresponding value threshold, the optimization scheduling objective is not satisfied.

9. The virtual power plant optimization scheduling method as described in claim 8, characterized in that, Generate a second optimized scheduling policy, including: When the judgment result is that the optimization scheduling objective is not met, the type of optimization scheduling objective that is not met is obtained. The objective type includes user demand objective, power supply reliability objective, economic benefit objective and carbon emission objective. Based on the unmet optimization scheduling target type, historical optimization scheduling logs with similar unmet target types are selected from the historical optimization scheduling logs and used as reference optimization scheduling logs. The historical optimized scheduling strategies in the reference optimized scheduling log are parsed to extract optimized scheduling strategy fragments that do not meet the target type; Based on the real-time operational requirements of the current monitoring cycle, the first type of system set, the second type of system set, and the specific circumstances that do not meet the target type, the extracted optimized scheduling strategy fragments are adaptively adjusted. The adjusted optimized scheduling strategy fragment is merged with the part of the first optimized scheduling strategy that meets the target type to obtain the second optimized scheduling strategy.

10. A virtual power plant optimized dispatching system, characterized in that, include: The acquisition module is used to select the target resource system to be optimized and evaluated, acquire the real-time operating data of the target resource system, and construct several operating data sequences; The analysis module is used to perform state analysis and fluctuation analysis on each running data sequence, determine the state coefficient and fluctuation coefficient of each real-time running data, and calculate the optimization coefficient of the corresponding target resource system by combining the weight coefficient of the corresponding real-time running data. The generation module is used to classify all resource systems according to the optimization coefficients to obtain a first set of systems and a second set of systems. Based on the preset optimization scheduling model, it generates real-time operation requirements, a first set of systems, and a first optimization scheduling strategy for the second set of systems. The scheduling module is used to simulate the first optimized scheduling strategy, determine whether the optimization scheduling objective is met based on the simulation results, and if so, issue a scheduling instruction; otherwise, generate a second optimized scheduling strategy.

Images