Intelligent scheduling method and system for cross-regional load migration of virtual power plant

Abstract

The invention discloses an intelligent scheduling method and system for cross-regional load migration of a virtual power plant, and relates to the technical field of load scheduling. The method comprises the following steps: collecting power generation cost parameter data, line capacity data, network loss coefficient data, PTDF matrix data, photovoltaic influence factor data, wind power influence factor data, photovoltaic power generation power data, wind power generation power data, conventional power generation power data and node load data of a power generation node; according to the virtual power plant cross-regional load migration intelligent scheduling method and system, by predicting the generation power and the load demand of each generation node in the future set duration, the load can be scheduled according to the predicted generation power and the load demand, the power demand can be met, and the variable quantity of the predicted load demand of each node is calculated; whether the load needs to be rescheduled or not is determined, and the power generation cost under the current load demand condition is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of load scheduling, and in particular to an intelligent scheduling method and system for cross-regional load migration of a virtual power plant. Background Art

[0002] Generally, a virtual power plant aggregates fragmented resources such as distributed photovoltaics, wind power, energy storage batteries, electric vehicles, and controllable loads through technologies such as the Internet of Things, big data, and artificial intelligence for unified coordination, participates in power market transactions and grid dispatching, so as to achieve flexible regulation and optimal allocation of energy. The prior art with the publication number CN119005647A discloses a scheduling method and system for grid loads. The specific solution is as follows: Using the first influence curve of the response delay time of the power consumption demand by the passenger flow of each electricity customer and the second influence curve of the migration time of the passenger flow to the electricity peak value, predict the power consumption demand of each electricity customer in the first future time period, so as to determine the estimated curve of the grid load of the first power grid in the first future time period; Then, based on the adjustment requirements of the adjustable load of the first power grid, optimize the estimated curve of the grid load of the first power grid in the first future time period to obtain the power load scheduling plan of the first power grid in the first future time period; In the first future time period, based on the power load scheduling plan of the first power grid in the first future time period, perform load scheduling on the first power grid.

[0003] Moreover, the generating sets and loads in a virtual power plant may be distributed at different nodes, resulting in different costs for the generating sets at different nodes to supply power to the loads at different nodes. The prior art generally only considers the balance between the total power supply and the load power to schedule the load, and does not consider the situation where the total load does not change significantly, but the loads between different nodes change significantly. This may lead to the problem of increased power supply costs when power supply is carried out according to the original scheduling plan. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent scheduling method and system for cross-regional load migration of a virtual power plant to solve the above deficiencies in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An intelligent scheduling method for cross-regional load migration of a virtual power plant, including the following steps: S1. Collect the power generation cost parameter data, line capacity data, line loss coefficient data, PTDF matrix data, photovoltaic influence factor data, wind power influence factor data, photovoltaic power generation data, wind power generation data, conventional power generation data, and node load data of the power generation nodes; S2. Update the pre-trained photovoltaic power generation prediction model and wind power generation prediction model based on the photovoltaic impact factor data, wind power impact factor data, photovoltaic power generation data, and wind power generation data; S3. Input the photovoltaic impact factor data into the pre-trained photovoltaic power generation prediction model to generate photovoltaic power generation prediction data, and input the wind power impact factor data into the pre-trained wind power generation prediction model to generate wind power generation prediction data; S4. Input the node load data into the pre-trained load prediction model to generate node load prediction data; S5. Calculate and process the change amount of the node load prediction data to generate node load change amount prediction data; S6. Conduct minimum power supply cost power generation node power analysis based on the power generation cost parameter data of the power generation node, line capacity data, network loss coefficient data, PTDF matrix data, photovoltaic power generation prediction data, wind power generation prediction data, and node load change amount prediction data to generate minimum power supply cost power generation node power analysis data; S7. Perform load dispatching operations of the virtual power plant based on the minimum power supply cost power generation node power analysis data.

[0006] Further, the S1 includes the following steps: S11. Collect the power generation cost parameter data of the power generation node to generate a set of power generation cost parameter data of the power generation node , , represents the power generation cost parameter data of the w-th power generation node, represents the maximum number of power generation nodes; among them, the cost parameters of the power generation node are obtained from government and authoritative agency reports, academic research and industry reports, and enterprise data. The cost parameters can be composed of capital cost (CAPEX): power plant construction, equipment procurement, land cost, etc.; operation and maintenance cost (OPEX): daily operation and maintenance, labor, insurance, etc.; fuel cost: coal, natural gas, uranium and other fuel prices; financing cost: loan interest rate, depreciation period, etc.; environmental cost: carbon emission tax, pollution treatment cost, etc.

[0007] S12. Collect the line capacity data to generate a set of line capacity data , , represents the line capacity data of the q-th line, represents the maximum number of lines; the line capacity data can be obtained through industry standards. The line capacity data is used to perform line power flow constraints during the minimum power supply cost analysis to ensure that the line power flow does not exceed the capacity; S13. Collect the network loss coefficient data and generate a set of network loss coefficient data , where represents the network loss coefficient between nodes w and j, , reflecting the impact of line parameters on network loss, represents the single - unit network loss coefficient of node w, characterizing the loss caused by the output of this unit alone, is a constant loss term independent of unit output (such as fixed network loss); S14. Collect the Power Transfer Distribution Factors (PTDF), and generate the PTDF matrix data C. Among them, the PTDF matrix data can be calculated through network topology: the connection relationship between nodes and lines; line parameters: resistance RR, reactance XX, susceptance BB (used to construct the admittance matrix); reference node (Slack Bus): usually select the balanced node as the power reference point, etc. S15. Collect the photovoltaic impact factor data , wind power impact factor data , photovoltaic power generation data , wind power generation data , conventional power generation data ; S16. Collect the node load data and generate a set of node load data , , represents the load data of the u - th node, represents the maximum number of nodes.

[0008] Furthermore, the S2 includes the following steps: S21. Judge whether the collected photovoltaic impact factor data and photovoltaic power generation data reach the set quantity. If so, update the pre - trained photovoltaic power generation prediction model based on the photovoltaic impact factor data and photovoltaic power generation data , and recount the collected photovoltaic impact factor data and photovoltaic power generation data ; S22. Judge whether the collected wind power impact factor data , and wind power generation data reach the set quantity. If so, update the pre - trained wind power generation prediction model based on the wind power impact factor data , and wind power generation data , and recount the collected wind power impact factor data , and wind power generation power data Re - count. For example, collect the above - mentioned photovoltaic impact factor data, photovoltaic power generation power data as the first updated sample data, collect wind power impact factor data, wind power generation power data as the second updated sample data. Every time the first updated sample data / the second updated sample data reaches the corresponding set sample data quantity, use the first updated sample data / the second updated sample data to update the pre - trained photovoltaic power generation power prediction model / wind power generation power prediction model. When training and updating, the photovoltaic impact factor data and wind power impact factor data are the inputs of their respective models, and the photovoltaic power generation power data and wind power generation power data are the outputs of their respective models. Among them, the pre - trained photovoltaic power generation power prediction model and wind power generation power prediction model can use deep learning models, such as the LSTM (Long Short - Term Memory Network) model. The pre - trained photovoltaic power generation power prediction model and wind power generation power prediction model are used to output the average photovoltaic power generation power and average wind power generation power within the set prediction duration in the future according to the input photovoltaic impact factor and wind power impact factor.

[0009] Furthermore, the step S3 includes the following steps: S31: Input the photovoltaic impact factor data into the pre - trained photovoltaic power generation power prediction model to obtain the predicted photovoltaic power generation power prediction data ; S32: Input the wind power impact factor data into the pre - trained wind power generation power prediction model to generate wind power generation power prediction data .

[0010] Furthermore, the step S4 includes the following steps: S41: Input the node load data into the pre - trained load prediction model to obtain the node load prediction data , and generate a set of node load prediction data . Among them, the collected node load data can be counted. When the count reaches the set quantity, use these node load data to train and update the pre - trained load prediction model. The LSTM model can be used during training and updating. The pre - trained load prediction model predicts the average node load data within the set prediction duration in the future.

[0011] Furthermore, the step S5 includes the following steps: S51: For the set of node load prediction data Node load prediction data for all nodes in Perform change amount calculation processing to obtain node load change amount prediction data , and generate a set of node load change amount prediction data .

[0012] Furthermore, S6 includes the following steps: S61. Construct a calculation formula for the power generation of each power generation node at the minimum cost under the condition of meeting power demand: (1); Among them, , , Are respectively the power generation cost parameter data of the power generation node w of the power generation node, that is =( , , ), When the corresponding node is a node for photovoltaic power generation Adopt the corresponding , When the corresponding node is a node for wind power generation Adopt the corresponding , When the corresponding node is a node for conventional power generation Adopt the corresponding (It is defaulted that the power generation of conventional power generation will not have large fluctuations); S62. Construct constraint conditions, and the formula is as follows: ; ; ; Among them, Is the predicted data of network loss power, , Are the minimum and maximum power generations of the wth power generation node. If the power generation node is a photovoltaic power generation node or a wind power generation node, then Is equal to , Is the maximum capacity of the qth line; Based on the predicted data of photovoltaic power generation , the predicted data of wind power generation , the data of conventional power generation , the set of node load prediction data And the set B of network loss coefficient data, perform network loss power calculation to obtain the predicted data of network loss power , and the calculation formula is as follows: ; S63. Solve formula (1) to obtain the power generation of each power generation node w, and generate the analysis data of the power generation of the power generation node with the minimum power supply cost , including the following steps: S631. Initialize the population and randomly generate M cuckoos for searching the power generation of nodes. Each cuckoo represents a potential analysis data of the power generation of the power generation node with the minimum power supply cost ; S632. Based on formula (1) and the constraint conditions, construct a fitness function and calculate the fitness of each cuckoo for searching the power generation of nodes: ; Among them, , is the penalty coefficient; S633. Each cuckoo reaches a new position through Levy flight. The cuckoo position update formula is as follows: ; Among them, represents the i-th cuckoo for searching the power generation of nodes, t represents the current iteration number, represents the step size, is the Levy random number ; S634. Discard the cuckoos for searching the power generation of nodes with non-optimal fitness with a discovery probability , and generate new cuckoos for searching the power generation of nodes to replace them. The formula is as follows: ; Among them, represents the cuckoo for searching the power generation of nodes with non-optimal fitness, represents the cuckoo for searching the power generation of nodes with optimal fitness; S635. Reset the cuckoos for searching the power generation of nodes that exceed the boundary to the nearest boundary; S636. Judge whether the maximum iteration number is reached. If so, output the analysis data of the power generation of the power generation node with the minimum power supply cost corresponding to the cuckoo for searching the power generation of nodes with the best fitness , if not, return to S632.

[0013] Furthermore, the above-mentioned S7 includes the following steps: S71. Judge whether there is node load change prediction data in the node load change prediction data set whose absolute value is greater than the set node load change threshold; S72. If not, then according to the analysis data of the power generation of the power generation node with the minimum power supply cost Perform scheduling operations on each power generation node; S73. If so, return to S4.

[0014] The virtual power plant cross-regional load transfer intelligent scheduling system includes a data acquisition module, an interface, a memory, a processor, and a control module.

[0015] The data acquisition module is used to collect the data required for the virtual power plant cross-regional load transfer intelligent scheduling method provided by the present invention, including power generation cost parameter data of power generation nodes, line capacity data, network loss coefficient data, PTDF matrix data, photovoltaic impact factor data, wind power impact factor data, photovoltaic power generation data, wind power generation data, conventional power generation data, node load data, the set number of sample data, the set prediction duration, the set node load change threshold, etc.; The data acquisition module is connected to the memory through the interface and is used to import the collected data into the system; The memory is used to store the collected data and computer programs; The processor is used to execute the computer program, implement the virtual power plant cross-regional load transfer intelligent scheduling method provided by the present invention, and generate power generation power analysis data of the power generation nodes with the minimum power supply cost; The control module is used to receive the power generation power analysis data of the power generation nodes with the minimum power supply cost, and schedule each power generation node according to the power generation power analysis data of the power generation nodes with the minimum power supply cost.

[0016] 1. Compared with the prior art, the virtual power plant cross-regional load transfer intelligent scheduling method and system provided by the present invention can predict the power generation power and load demand of each power generation node within the set future duration, so that the load can be scheduled according to the predicted power generation power and load demand to meet the power consumption demand.

[0017] 2. Compared with the prior art, the virtual power plant cross-regional load transfer intelligent scheduling method and system provided by the present invention can calculate the change amount of the predicted load demand of each node to determine whether it is necessary to reschedule the load, reduce the power generation cost under the current load demand, and avoid the problem that the total load does not change significantly while the load between different nodes changes significantly, resulting in an increase in the power supply cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings.

[0019] Figure 1It is the method step diagram provided by the embodiment of the present invention; Figure 2 It is the system structure block diagram provided by the embodiment of the present invention. Specific implementation manners

[0020] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0021] In the following, example embodiments will be described more fully with reference to the accompanying drawings. However, the example embodiments may be embodied in different forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0022] In the case of no conflict, the various embodiments of the present disclosure and the various features in the embodiments may be combined with each other.

[0023] As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items.

[0024] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. As used herein, the singular forms "a" and "the" are also intended to include the plural forms unless the context clearly indicates otherwise. It will also be understood that when the terms "comprises" and / or "is made of" are used in this specification, the specified features, wholes, steps, operations, elements, and / or components are present, but do not preclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or groups thereof.

[0025] The embodiments described herein may be described with reference to the plan views and / or cross-sectional views by means of the ideal schematic diagrams of the present disclosure. Therefore, the example illustrations may be modified according to the manufacturing technology and / or tolerances. Therefore, the embodiments are not limited to the embodiments shown in the drawings, but include modifications of the configurations formed based on the manufacturing process. Therefore, the regions illustrated in the drawings have schematic properties, and the shapes of the regions shown in the drawings illustrate the specific shapes of the regions of the elements, but are not intended to be restrictive.

[0026] Please refer to Figure 1 , the intelligent scheduling method for cross-regional load migration of the virtual power plant, including the following steps: S1. Collect the power generation cost parameter data, line capacity data, line loss coefficient data, PTDF matrix data, photovoltaic impact factor data, wind power impact factor data, photovoltaic power generation data, wind power generation data, conventional power generation data, and node load data of the power generation nodes, including the following steps: S11. Collect the power generation cost parameter data of the power generation nodes, and generate a set of power generation cost parameter data of the power generation nodes , , represents the power generation cost parameter data of the w-th power generation node, represents the maximum number of power generation nodes; among them, the cost parameters of the power generation nodes are obtained from government and authoritative agency reports, academic research and industry reports, and enterprise data. The cost parameters can be composed of capital cost (CAPEX): power plant construction, equipment procurement, land cost, etc.; operation and maintenance cost (OPEX): daily operation and maintenance, labor, insurance, etc.; fuel cost: fuel prices such as coal, natural gas, uranium, etc.; financing cost: loan interest rate, depreciation period, etc.; environmental cost: carbon emission tax, pollution treatment cost, etc.

[0027] S12. Collect the line capacity data, and generate a set of line capacity data , , represents the line capacity data of the q-th line, represents the maximum number of lines; the line capacity data can be obtained through industry standards. The line capacity data is used to perform line power flow constraints during the minimum power supply cost analysis to ensure that the line power flow does not exceed the capacity; S13. Collect the line loss coefficient data, and generate a set of line loss coefficient data , where represents the line loss coefficient between nodes w and j, , reflecting the impact of line parameters on line losses, represents the single-machine line loss coefficient of node w, characterizing the loss caused by the output of this unit alone, is a constant loss term independent of the unit output (such as fixed line losses); S14. Collect the Power Transfer Distribution Factors (PTDF), and generate the PTDF matrix data C. Among them, the PTDF matrix data can be calculated through network topology: the connection relationship between nodes and lines; line parameters: resistance RR, reactance XX, susceptance BB (used to construct the admittance matrix); reference node (Slack Bus): usually select the balanced node as the power reference point, etc.; S15. Collect the photovoltaic impact factor data , wind power impact factor data , photovoltaic power generation data , wind power generation data , conventional power generation data ; S16. Collect the node load data, and generate a set of node load data , , represents the load data of the u-th node, represents the maximum number of nodes.

[0028] S2. Update the pre-trained photovoltaic power prediction model and wind power prediction model based on the photovoltaic impact factor data, wind power impact factor data, photovoltaic power generation data, and wind power generation data, including the following steps: S21. Determine whether the collected photovoltaic impact factor data and the photovoltaic power generation data reach the set quantity. If so, update the pre-trained photovoltaic power prediction model based on the photovoltaic impact factor data and the photovoltaic power generation data , and re-count the collected photovoltaic impact factor data and the photovoltaic power generation data . S22. Determine whether the collected wind power impact factor data , and the wind power generation data reach the set quantity. If so, update the pre-trained wind power prediction model based on the wind power impact factor data , and the wind power generation data , and re-count the collected wind power impact factor data , and the wind power generation data . For example, collect the photovoltaic impact factor data, photovoltaic power generation data as update sample data one, collect the wind power impact factor data, wind power generation data as update sample data two. When the update sample data one / update sample data two reaches the corresponding set sample data quantity, use the update sample data one / update sample data two to update the pre-trained photovoltaic power prediction model / wind power prediction model. When training and updating, the photovoltaic impact factor data and the wind power impact factor data are the inputs of their respective models, and the photovoltaic power generation data and the wind power generation data are the outputs of their respective models. Among them, the pre-trained photovoltaic power prediction model and wind power prediction model can use deep learning models, such as the LSTM (Long Short-Term Memory Network) model. The pre-trained photovoltaic power prediction model and wind power prediction model are used to output the average photovoltaic power and average wind power within the set prediction duration in the future according to the input photovoltaic impact factor and wind power impact factor.

[0029] S3. Input the photovoltaic impact factor data into the pre-trained photovoltaic power generation prediction model to generate photovoltaic power generation prediction data, and input the wind power impact factor data into the pre-trained wind power generation prediction model to generate wind power generation prediction data, including the following steps: S31. Input the photovoltaic impact factor data into the pre-trained photovoltaic power generation prediction model to obtain the predicted photovoltaic power generation prediction data ; S32. Input the wind power impact factor data into the pre-trained wind power generation prediction model to generate wind power generation prediction data .

[0030] S4. Input the node load data into the pre-trained load prediction model to generate node load prediction data, including the following steps: S41. Input the node load data into the pre-trained load prediction model to obtain the node load prediction data , generating a set of node load prediction data , where the collected node load data can be counted. When the count reaches the set number, use these node load data to train and update the pre-trained load prediction model. The LSTM model can be used during training and updating. The pre-trained load prediction model predicts the average node load data within the set prediction duration in the future.

[0031] S5. Calculate and process the change amount of the node load prediction data to generate node load change amount prediction data, including the following steps: S51. Perform change amount calculation and processing on the node load prediction data of all nodes in the set of node load prediction data to obtain the node load change amount prediction data , generating a set of node load change amount prediction data . .

[0032] S6. Based on the power generation cost parameter data of the power generation nodes, line capacity data, network loss coefficient data, PTDF matrix data, photovoltaic power generation prediction data, wind power generation prediction data, and node load change amount prediction data, perform minimum power supply cost power generation node power generation power analysis to generate minimum power supply cost power generation node power generation power analysis data, including the following steps: S61. Construct the power generation power calculation formula for each power generation node at the minimum cost under the condition of meeting the power demand: (1); Among them, , , are respectively the power generation cost parameter data of the power generation node w of the power generation node, that is, = ([[]] , , ), When the corresponding node is a node of photovoltaic power generation Adopt the corresponding , When the corresponding node is a node of wind power generation Adopt the corresponding , When the corresponding node is a node of conventional power generation Adopt the corresponding (by default, the power generation power of conventional power generation will not fluctuate greatly); S62. Construct constraint conditions, and the formula is as follows: ; ; ; Among them, is the predicted data of network loss power, , are the minimum and maximum power generation powers of the wth power generation node. If the power generation node is a photovoltaic power generation node or a wind power generation node, then is equal to , is the maximum capacity of the qth line; Based on the predicted data of photovoltaic power generation power , the predicted data of wind power generation power , the data of conventional power generation power , the set of predicted data of node load and the set of network loss coefficient data B, calculate the network loss power to obtain the predicted data of network loss power , and the calculation formula is as follows: ; S63. Solve formula (1) to obtain the power generation power of each power generation node w, and generate the analysis data of the power generation power of the power generation node with the minimum power supply cost , including the following steps: S631. Initialize the population and randomly generate M node power generation power search cuckoos. Each cuckoo represents a potential analysis data of the power generation power of the power generation node with the minimum power supply cost ; S632. Based on formula (1) and constraint conditions, construct a fitness function and calculate the fitness of each node power generation power search cuckoo: ; Among them, and are penalty coefficients; S633. Each cuckoo reaches a new position through Levy flight. The cuckoo position update formula is as follows: ; Among them, represents the cuckoo searching for the power generation power of the i-th node, t represents the current iteration number, represents the step size, is a Levy random number ; S634. Discard the cuckoo searching for the power generation power of the node with non-optimal fitness with a discovery probability of , and generate a new cuckoo searching for the power generation power of the node to replace it. The formula is as follows: ; Among them, represents the cuckoo searching for the power generation power of the node with non-optimal fitness, represents the cuckoo searching for the power generation power of the node with optimal fitness; S635. Reset the cuckoo searching for the power generation power of the node that exceeds the boundary to the nearest boundary; S636. Determine whether the maximum iteration number is reached. If so, output the analysis data of the power generation power of the power generation node with the best fitness corresponding to the minimum power supply cost , if not, return to S632.

[0033] S7. Perform the load dispatching operation of the virtual power plant according to the analysis data of the power generation power of the power generation node with the minimum power supply cost, including the following steps: S71. Determine whether there is a predicted data of the node load change amount in the set of predicted data of the node load change amount whose absolute value is greater than the set node load change amount threshold; S72. If not, perform the dispatching operation on each power generation node according to the analysis data of the power generation power of the power generation node with the minimum power supply cost ; S73. If so, return to S4.

[0034] Please refer to Appendix Figure 2 . The virtual power plant cross-regional load migration intelligent dispatching system is used to execute the virtual power plant cross-regional load migration intelligent dispatching method provided by the present invention, including a data acquisition module, an interface, a storage, a processor, and a control module.

[0035] The data acquisition module is used to collect the data required for the intelligent scheduling method of cross-regional load migration of the virtual power plant provided by the present invention, including the power generation cost parameter data of the power generation nodes, line capacity data, network loss coefficient data, PTDF matrix data, photovoltaic influence factor data, wind power influence factor data, photovoltaic power generation data, wind power generation data, conventional power generation data, node load data, the set number of sample data, the set prediction duration, the set node load change threshold, etc.; The data acquisition module is connected to the storage through an interface and is used to import the collected data into the system; The storage is used to store the collected data and computer programs; The processor is used to execute the computer program, implement the intelligent scheduling method of cross-regional load migration of the virtual power plant provided by the present invention, and generate the power generation power analysis data of the power generation nodes with the minimum power supply cost; The control module is used to receive the power generation power analysis data of the power generation nodes with the minimum power supply cost, and schedule each power generation node according to the power generation power analysis data of the power generation nodes with the minimum power supply cost.

[0036] Only some exemplary embodiments of the present invention have been described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An intelligent scheduling method for cross-regional load migration of a virtual power plant, characterized by: The following steps are involved: S1. Collect power generation cost parameter data, line capacity data, network loss coefficient data, PTDF matrix data, photovoltaic influence factor data, wind power influence factor data, photovoltaic power generation data, wind power generation data, conventional power generation data and node load data of power generation nodes; S2. updating the pre-trained photovoltaic power prediction model and wind power prediction model based on the photovoltaic impact factor data, wind power impact factor data, photovoltaic power generation data, and wind power generation data; S3, inputting the photovoltaic influence factor data into a pre-trained photovoltaic power prediction model to generate photovoltaic power prediction data, and inputting the wind power influence factor data into a pre-trained wind power prediction model to generate wind power prediction data; S4, inputting the node load data into a pre-trained load prediction model to generate node load prediction data; S5, calculating and processing the change amount of the node load prediction data to generate node load change prediction data; S6. Based on the power generation cost parameter data of the power generation node, the line capacity data, the network loss coefficient data, the PTDF matrix data, the photovoltaic power generation power prediction data, the wind power generation power prediction data and the node load change prediction data, the power generation power analysis of the power generation node with the minimum power supply cost is performed to generate the power generation power analysis data of the power generation node with the minimum power supply cost; S7. Perform load dispatching of the virtual power plant based on the power generation analysis data of the power generation node with the minimum power supply cost.

2. The method for intelligent scheduling of cross-regional load migration of virtual power plants according to claim 1 is characterized in that: The S1 comprises the following steps: S11. Collect power generation cost parameter data of power generation nodes and generate a power generation cost parameter data set of power generation nodes , , represents the power generation cost parameter data of the wth power generation node, Indicates the maximum number of power generation nodes; S12. Collect line capacity data and generate line capacity data set , , represents the capacity data of the qth line, Indicates the maximum number of lines; S13. Collect network loss coefficient data and generate network loss coefficient data set ,in represents the network loss coefficient between nodes w and j, , represents the single-machine network loss coefficient of node w, is a constant loss term that is independent of the unit output; S14, collecting power transmission distribution factors and generating PTDF matrix data C; S15. Collect photovoltaic impact factor data , Wind power impact factor data , Photovoltaic power data , Wind power generation data , conventional power generation data ; S16. Collect node load data and generate node load data set , , represents the load data of the u-th node, Indicates the maximum number of nodes.

3. The method for intelligent scheduling of cross-regional load migration of virtual power plants according to claim 2 is characterized in that: The S2 comprises the following steps: S21, judging the collected photovoltaic impact factor data and photovoltaic power data Whether the set quantity is reached, if so, based on the photovoltaic impact factor data and photovoltaic power data Update the pre-trained photovoltaic power generation prediction model and collect the photovoltaic influencing factor data and photovoltaic power data Conduct a recount; S22. Determine the collected wind power impact factor data , and wind power generation data Whether the set number is reached, if so, based on the wind power impact factor data , and wind power generation data Update the pre-trained wind power generation prediction model and collect the wind power influencing factor data , and wind power generation data Recounting is performed.

4. The method for intelligent scheduling of cross-regional load migration of virtual power plants according to claim 3 is characterized in that: The S3 comprises the following steps: S31, the photovoltaic impact factor data Input the pre-trained photovoltaic power generation prediction model to obtain the predicted photovoltaic power generation prediction data ; S32, the wind power impact factor data Input the pre-trained wind power generation prediction model to generate wind power generation prediction data .

5. The method for intelligent scheduling of cross-regional load migration of virtual power plants according to claim 4 is characterized in that: The S4 comprises the following steps: S41, the node load data Input the pre-trained load forecasting model to obtain node load forecasting data , generate node load prediction data set .

6. The method for intelligent scheduling of cross-regional load migration of virtual power plants according to claim 5 is characterized in that: The S5 comprises the following steps: S51, the node load prediction data set Node load forecast data for all nodes in Carry out change calculation and processing to obtain node load change prediction data , generate node load change prediction data set .

7. The method for intelligent scheduling of cross-regional load migration of virtual power plants according to claim 6 is characterized by: The S6 comprises the following steps: S61. Construct the power generation calculation formula of each power generation node at the minimum cost under the condition of meeting the power demand: （1）； in, , , are the power generation cost parameter data of the power generation node w, When the corresponding node is a photovoltaic power generation node Adopt corresponding , The corresponding node is the node of wind power generation Adopt corresponding , When the corresponding node is a conventional power generation node Use the corresponding ; S62. Construct constraint conditions. The formula is as follows: ； ； ； in, is the network loss power prediction data, , is the minimum and maximum power generation of the wth power generation node, is the maximum capacity of the qth line; Based on photovoltaic power generation power prediction data , Wind power generation power forecast data , conventional power generation data , node load prediction data set And the network loss coefficient data set B, the network loss power is calculated to obtain the network loss power prediction data ; S63. Solve formula (1) to obtain the power generation of each power generation node w, and generate the power generation analysis data of the power generation node with the minimum power supply cost. , including the following steps: S631, initialize the population, randomly generate M node power generation search cuckoos, each cuckoo represents a potential minimum power supply cost power generation node power generation analysis data ; S632, based on formula (1) and constraints, construct a fitness function to calculate the fitness of each node power generation search cuckoo: ； in, , is the penalty coefficient; S633, each cuckoo reaches a new location via Lévy flight; S634, based on the probability of discovery Discard the node power generation search cuckoo with non-optimal fitness and generate a new node power generation search cuckoo instead; S635, resetting the node power generation search cuckoo that exceeds the boundary to the nearest boundary; S636: Determine whether the maximum number of iterations has been reached. If so, output the node power generation power with the best fitness and search for the minimum power supply cost power generation node power generation power analysis data corresponding to the cuckoo. , if not, return to S632.

8. The method for intelligent scheduling of cross-regional load migration of virtual power plants according to claim 7 is characterized in that: The S7 comprises the following steps: S71: Determine the node load change prediction data set Is there any node load change prediction data in the The absolute value of is greater than the set node load change threshold; S72, if not, then analyze the power generation data based on the minimum power supply cost power generation node Perform scheduling operations on each power generation node; S73. If yes, return to S4.

9. A virtual power plant cross-regional load migration intelligent scheduling system, used to execute the virtual power plant cross-regional load migration intelligent scheduling method according to any one of claims 1 to 8, characterized in that: It includes a data acquisition module, an interface, a storage device, a processor, and a control module.

Images