Virtual power plant task scheduling processing method, electronic equipment and storage medium

Abstract

The invention provides a virtual power plant task scheduling processing method, electronic equipment and a storage medium, and belongs to the technical field of power distribution network regulation and control. The method comprises the following steps: receiving a to-be-scheduled task, wherein the to-be-scheduled task comprises target power and a power deviation threshold value; selecting a plurality of aggregation units with different aggregation types and distribution positions from an adjustment unit library according to the to-be-scheduled task as first units, and issuing a first scheduling instruction for executing target power to the first units; selecting a plurality of aggregation units matched with the aggregation type and the distribution position of the first unit from the margin unit library as second units, and issuing a second scheduling instruction for executing a power deviation threshold to the second units; according to response results of the first scheduling instruction and the second scheduling instruction, the first units which do not meet the first condition are adjusted to the margin unit library, and the second units which meet the second condition are adjusted to the adjustment unit library. According to the invention, the test dynamics and accuracy of the response deviation degree of the aggregation unit are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution network regulation, and in particular to a virtual power plant task scheduling processing method, an electronic device and a storage medium. BACKGROUND

[0002] With the intelligentization and marketization of the power system, as an innovative form of distributed energy aggregation, the virtual power plant (VPP) has gradually become an important means of power grid regulation. By integrating energy storage units, adjustable loads and other distributed energy resources, the virtual power plant can provide key services such as peak load shifting and load regulation in grid dispatching, thereby enhancing the flexibility and stability of the power grid. However, the efficient operation of the virtual power plant highly depends on the accurate scheduling of its internal aggregated units and the response deviation degree evaluation.

[0003] Due to the extremely high real-time requirements of grid dispatching, embedding precision evaluation into real-time scheduling will interfere with the normal scheduling process, causing scheduling delays or untimely responses, thereby affecting the stability and reliability of the power grid. Therefore, traditional precision evaluation methods are usually performed in an offline environment or for a short period of time when the grid load is light, and are relatively independent of the actual scheduling process. For example, when the energy storage unit is first connected to the virtual power plant, it is tested whether the energy storage unit can accept the virtual power plant dispatching command to charge and discharge, and whether the air conditioner can reduce the load. This kind of method not only cannot produce economic value in actual operation, but also due to the lack of dynamic tracking mechanism, it is difficult to capture the time-varying characteristics of the performance of the aggregated units.

[0004] Therefore, there is an urgent need for a virtual power plant task scheduling method that can dynamically test the response deviation degree of the aggregated units without affecting the overall response deviation degree of the virtual power plant. SUMMARY

[0005] The purpose of the present application is to provide a virtual power plant task scheduling processing method, an electronic device and a storage medium to solve at least one of the above technical problems.

[0006] To achieve the above purpose, in a first aspect, the present application provides a virtual power plant task scheduling processing method, which comprises: receiving a to-be-scheduled task, the to-be-scheduled task comprising a target power, a target execution time period and a power deviation threshold, the power deviation threshold being less than the target power; According to the to-be-scheduled task and the unit characteristics of each aggregated unit in the adjustment unit library, first unit balance selection is performed to select multiple aggregated units with different aggregation types and distribution positions from the adjustment unit library as first units, and first scheduling instructions for executing the target power are issued to the first units, the adjustment unit library being composed of aggregated units with a first comprehensive response deviation less than a first preset deviation; Multiple aggregated units with aggregation types and distribution positions matching those of the first units are selected from the margin unit library as second units, and second scheduling instructions for executing the power deviation threshold are issued to the second units, the margin unit library being composed of aggregated units with a second comprehensive response deviation greater than or equal to a second preset deviation and aggregated units with an execution frequency less than a preset frequency threshold; According to the response results of the first scheduling instructions and the second scheduling instructions, the first units that do not meet a first condition are adjusted to the margin unit library, and the second units that meet a second condition are adjusted to the adjustment unit library.

[0007] In some embodiments, the unit characteristics include aggregation type, distribution position, real-time occupied power, and unit power constraint parameter, and the first unit balance selection from the adjustment unit library includes: An energy map is constructed based on the distribution positions of each aggregated unit in the adjustment unit library, each node in the energy map corresponding to an aggregated unit; An energy-time space is constructed on the energy map based on the unit characteristics of each node; Based on the target power and the target execution time period, first unit balance selection is performed in the energy-time space to generate multiple groups of candidate power output paths, each candidate power output path including a plurality of aggregated units with different aggregation types and distribution positions; The comprehensive performance indicators of each candidate power output path are calculated respectively, the candidate power output path with the lowest comprehensive performance indicator is determined as a target power output path, and the plurality of aggregated units included in the target power output path are selected as first units. In some embodiments, the energy-time space is constructed on the energy map based on the unit characteristics of each node, including: The target execution time period is divided into multiple sub-time periods, and the available power range of each node in the multiple sub-time periods is predicted based on the unit characteristics of each node; On the energy map, an energy layer is built for each node in each sub-time period, and the energy layer is used to represent the available power range of the corresponding aggregation unit in the corresponding sub-time period; The energy layers of each node in all sub-time periods are superimposed to form an energy-time space, wherein the energy layer change track of each node on the time axis is used to represent the predicted available power of the corresponding aggregation unit in different time periods. In some embodiments, the selecting, from the margin unit library, a plurality of aggregation units matching the aggregation type and distribution position of the first unit as second units comprises: The selecting, from the margin unit library, a plurality of aggregation units matching the aggregation type and distribution position of the first unit as second units is based on the execution times of each aggregation unit in the margin unit library; Before the issuing of the second scheduling instruction for the second units to execute the power deviation threshold, comprising: Obtaining historical execution data of each second unit, and calculating a power fluctuation value of each second unit based on the historical execution data, unit characteristics and the power deviation threshold; Determining a current execution power of each second unit based on the power fluctuation value and the last execution power, so that the sum of the current execution powers of each second unit matches the power deviation threshold; Forming a second scheduling instruction based on the current execution power of each second unit. In some embodiments, the historical execution data includes historical execution power and corresponding historical response deviation degree, and the unit characteristics include real-time occupied power and unit power constraint parameters, and the calculating of the power fluctuation value of each second unit based on the historical execution data, unit characteristics and the power deviation threshold comprises: Calculating a response deviation degree weight factor based on the historical response deviation degree in the historical execution data; Calculating a basic fluctuation amplitude based on the real-time occupied power and the unit power constraint parameters; Calculating an initial fluctuation value of each second unit based on the response deviation degree weight factor, the basic fluctuation amplitude and a preset influence weight; Calculating an initial sum of the current execution powers of all second units based on the initial fluctuation value of each second unit and the last historical execution power; According to the ratio of the initial sum to the power deviation threshold, performing scaling processing on the initial fluctuation value of each second unit to calculate the power fluctuation value of each second unit. In some embodiments, the first condition comprises that a first comprehensive response deviation degree is less than a first preset deviation degree; the second condition comprises that a second comprehensive response deviation degree is less than a second preset deviation degree and the execution times are greater than or equal to a preset number threshold; the first scheduling instruction comprises a response target value of each first unit, the second scheduling instruction comprises a response target value of each second unit, the response result of the first scheduling instruction comprises an execution value of each first unit, and the response result of the second scheduling instruction comprises an execution value of each second unit. The adjusting, according to the response results of the first scheduling instruction and the second scheduling instruction, the first units that do not meet the first condition to the buffer unit bank and the second units that meet the second condition to the adjustment unit bank comprises: According to the first scheduling instruction and the response result of the first scheduling instruction, the single-time response deviation degree of each first unit is calculated. The first comprehensive response deviation degree of the first unit is calculated based on the single-time response deviation degree of the first unit and a historical response deviation degree in historical execution data of the first unit. The first units that do not meet the first comprehensive response deviation degree being less than the first preset deviation degree are adjusted to the buffer unit bank. According to the second scheduling instruction and the response result of the second scheduling instruction, the single-time response deviation degree of each second unit is calculated. The second comprehensive response deviation degree of the second unit is calculated based on the single-time response deviation degree of the second unit and a historical response deviation degree in historical execution data of the second unit. The second units that meet the second comprehensive response deviation degree being less than the second preset deviation degree and the execution times being greater than or equal to the preset number threshold are adjusted to the adjustment unit bank. In some embodiments, after the receiving the task to be scheduled, the method further comprises: The historical deviation record of the buffer unit bank is obtained, and the historical deviation record is input to a Bayesian estimator to obtain a posterior deviation distribution output by the Bayesian estimator. The buffer power is determined according to the posterior deviation distribution and the power deviation threshold. The second scheduling instruction for the second unit to execute the power deviation threshold comprises: The second scheduling instruction for the second unit to execute the buffer power is issued, so that the total power output by the second unit matches the power deviation threshold. In some embodiments, the obtaining the historical deviation record of the buffer unit bank and inputting the historical deviation record to the Bayesian estimator to obtain the posterior deviation distribution output by the Bayesian estimator comprises: obtain historical bias records of the margin unit library and historical covariate information, the historical covariate information including at least one of weather types and holiday types; obtain current covariate information corresponding to the historical covariate information; input the historical bias records, the historical covariate information and the current covariate information into a Bayesian estimator to obtain a posterior bias distribution output by the Bayesian estimator.

[0008] In a second aspect, to achieve the above object, the present application further provides an electronic device, comprising: one or more processors; a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the processing method for virtual power plant task scheduling as described above.

[0009] In a third aspect, to achieve the above object, the present application further provides a computer storage medium, the storage medium storing executable instructions, the instructions being executed by a processor to make the processor execute the processing method for virtual power plant task scheduling as described above.

[0010] Compared with the prior art, the beneficial effects of the present application include: In a first aspect, by separating the to-be-scheduled tasks of the virtual power plant into target power and power bias threshold, and respectively executing them by the regulation unit library and the margin unit library, the target power is completed by the high-precision regulation unit library, thereby ensuring that the overall response bias degree of the virtual power plant does not decrease, and reducing the risk of power grid assessment. At the same time, the power bias threshold is executed by the low-precision / aggregated units newly connected in the margin unit library to test or bottom out, so as not to affect the power grid assessment settlement, and to dynamically test the comprehensive response bias degree, thereby improving the dynamicity and accuracy of the response bias degree evaluation.

[0011] In a second aspect, when selecting the first units, the distribution position and type of the aggregated units are considered for balanced selection, thereby avoiding the concentration of all first units in the same geographical location or the same type, and reducing the risk of failure of all first units due to local faults (such as device failure in a certain area, natural disasters, etc.). This decentralized strategy enhances the robustness of the virtual power plant, and even if part of the first units fail, the first units in other regions or types can still continue to work.

[0012] In a third aspect, by selecting multiple aggregation units that match the aggregation type and distribution location of the first units as the second units, on the one hand, since the second units match the aggregation type and distribution location of the first units, this helps to enhance the coordination within the virtual power plant. When the power grid is disturbed or the demand changes, the second units can understand and adapt to the working state and mode of the first units more quickly, so as to more effectively cooperate with the first units to adjust power and jointly maintain the stable operation of the power grid. At the same time, this matching also reduces the risk of power fluctuations and system instability caused by the lack of coordination between units, and improves the reliability and quality of the entire virtual power plant in providing external power services. On the other hand, this provides an ideal contrast environment for the evaluation of the response deviation degree. Under the same type and location conditions, by comparing the response of the first units and the second units when performing tasks, other interference factors such as inherent differences between different types of units or differences in power grid conditions at different locations can be effectively reduced, so that the response deviation degree of the second units can be more accurately evaluated. BRIEF DESCRIPTION OF DRAWINGS

[0013] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as limiting the scope of the present application.

[0014] Figure 1 A flowchart of a processing method of virtual power plant task scheduling in an embodiment; Figure 2 A detailed flowchart of how to select first units with different aggregation types and distribution locations from the regulation unit library according to the to-be-scheduled task and the unit characteristics of each aggregation unit in the regulation unit library in an embodiment; Figure 3 A flowchart of how to form a second scheduling instruction of the second units in an embodiment; Figure 4 A flowchart of how to determine the margin power in an embodiment; Figure 5 A structural diagram of an electronic device involved in the processing method of virtual power plant task scheduling in an embodiment of the present application. DETAILED DESCRIPTION

[0015] In order to make the purposes, technical solutions and advantages of the present application more clear, the following will further describe the present application in combination with the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0016] All the terms used in the present application (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted to have meanings consistent with the context of the present specification, and should not be interpreted in an idealized or overly formal manner.

[0017] For example, the terms "first", "second", and the like as used in the present application can be used herein to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from another element. For example, without departing from the scope of the present application, a first unit can be referred to as a second unit, and similarly, a second unit can be referred to as a first unit. Both the first unit and the second unit are units, but they are not the same unit.

[0018] For another example, the terms "include", "contain", and the like used in the present application indicate the presence of a feature, step, operation and / or component, but do not exclude the presence or addition of one or more other features, steps, operations or components.

[0019] As mentioned earlier, due to the extremely high real-time requirement of power grid dispatching, embedding accuracy evaluation into real-time dispatching will interfere with the normal dispatching process, causing delay in dispatching or untimely response, thereby affecting the stability and reliability of the power grid. Therefore, the traditional accuracy evaluation method is usually carried out in an offline environment or for a short test when the power grid load is light, and is relatively independent of the actual dispatching process. For example, when the energy storage unit is first connected to the virtual power plant, test whether the energy storage unit can accept the virtual power plant dispatching command to charge and discharge, and air conditioner load reduction, etc. This kind of method not only cannot produce economic value in actual operation, but also due to the lack of dynamic tracking mechanism, it is difficult to capture the time-varying characteristics of the performance of the aggregation unit. Therefore, at present, it is urgent to propose a method that can dynamically test the response deviation degree of the aggregation unit without affecting the overall response deviation degree of the virtual power plant. To this end, the present application provides a virtual power plant task scheduling processing method, electronic equipment and storage medium, which improves the dynamicity and accuracy of testing the response deviation degree of the aggregation unit.

[0020] As shown in Figure 1 The present application embodiment provides a virtual power plant task scheduling processing method, which comprises the following steps: Step S10, receiving a task to be dispatched.

[0021] In this embodiment, the to-be-scheduled task refers to a task issued by a power grid dispatching center to a virtual power plant, including a target power, a target execution time period, and a power deviation threshold. For example, in t1-t2, t2-t3, and t3-t4, the load is increased by P1, P2, and P3, respectively. The target power corresponding to the three target execution time periods is P0+P1, P0+P2, and P0+P3, respectively. P0 is the baseline power before the increase, and P1, P2, and P3 are the increase powers in the three time periods, respectively.

[0022] The target power refers to the power output target that the virtual power plant needs to achieve. The target execution time period refers to the time range in which the target power needs to be executed. The power deviation threshold refers to the maximum deviation between the actual power and the target power, and the power deviation threshold is less than the target power. It should be noted that the power deviation threshold is set according to the evaluation rules of the power grid dispatching center. For example, the evaluation rules of a certain power grid clearly stipulate that "filling valley: the actual adjustment power of the third party independent principal is greater than 120% of the day-ahead winning power, and the part exceeding 120% at this sampling point is not compensated. The actual adjustment power is less than 70% of the day-ahead winning power, and this sampling point is not compensated." The power deviation threshold is 20% of the target power.

[0023] In step S20, a first unit balancing selection is performed according to the to-be-scheduled task and the unit characteristics of each aggregated unit in the adjustment unit library, a plurality of aggregated units with different aggregation types and distribution positions are selected from the adjustment unit library as first units, and a first scheduling instruction for executing the target power is issued to the first units.

[0024] It should be noted that in this embodiment, the virtual power plant divides all schedulable aggregated units into an adjustment unit library and a margin unit library. The adjustment unit library is composed of aggregated units with a first comprehensive response deviation degree less than a first preset deviation degree, and the margin unit library is composed of aggregated units with a second comprehensive response deviation degree greater than or equal to a second preset deviation degree and aggregated units with an execution frequency less than a preset frequency threshold. The first comprehensive response deviation degree and the second comprehensive response deviation degree are the comprehensive response deviation degrees calculated by the aggregated units in the adjustment unit library and the margin unit library in the task scheduling, respectively.

[0025] The first preset deviation degree is an index for measuring the response deviation degree of the aggregated units in the regulation unit library, representing the upper limit of the response deviation degree of these units in historical scheduling tasks. Only the units with a comprehensive response deviation degree lower than the first preset deviation degree are included in the regulation unit library to ensure the high response deviation degree of the regulation unit library as a whole, thereby enabling the virtual power plant to stably and accurately respond to the scheduling instructions of the power grid. The second preset deviation degree is an index for measuring the response deviation degree of the aggregated units in the margin unit library, representing the lower limit of the response deviation degree of these units in historical scheduling tasks. The numerical values of the first preset deviation degree and the second preset deviation degree can be the same or different.

[0026] The preset number threshold is related to the number of times the aggregated unit is executed for scheduling instructions. It can be set to 0 to directly place the newly accessed aggregated unit in the margin unit library for scheduling testing, thereby allowing the newly accessed aggregated unit to test the response deviation degree in actual scheduling. The preset number threshold can also be set to any other suitable numerical value, indicating that only when the aggregated unit has undergone a certain number of scheduling tests and meets the corresponding accuracy requirements, it will be placed in the regulation unit library. Such a setting can ensure that the units entering the regulation unit library not only meet the accuracy requirements, but also have a certain level of actual execution experience, thereby improving the reliability and stability of the regulation unit library.

[0027] The unit characteristics of the aggregated unit include one or more of the aggregation type, distribution location, real-time occupied power, and unit power constraint parameters. Among them, the aggregation type refers to the functional category of the aggregated unit, such as energy storage unit, air conditioning load, photovoltaic power station, and traditional power station. Different types of aggregated units play different roles and functions in the power system. For example, energy storage units can be used for power regulation and emergency backup, while photovoltaic power stations mainly provide renewable power. The distribution location refers to the geographical distribution of the aggregated unit, which can be a specific physical location (such as a certain community, a certain street, etc.), or a regional distribution (such as a certain urban area), without necessarily being accurate to a specific latitude and longitude location point. The real-time occupied power represents the power actually output or consumed by the aggregated unit at a certain time, which is used to reflect the real-time running state of the aggregated unit. The unit power constraint refers to the safety boundary constraint of each aggregated unit. Taking the energy storage aggregated unit as an example, the unit power constraint parameters include power constraints and energy constraints.

[0028] The power constraint refers to the fact that at any time, the sum of the baseline power and the upward adjusted target power cannot exceed the maximum charging power, or the sum of the baseline power and the downward adjusted target power cannot exceed the maximum discharging power.

[0029] The energy constraint means that the integral of the power after upward adjustment over the adjustment period cannot exceed the maximum remaining chargeable amount at the start time of the scheduling, or the integral of the power after downward adjustment over the adjustment period cannot exceed the maximum remaining dischargeable amount at the start time of the scheduling.

[0030] The first unit balanced selection means selecting multiple aggregation units with different aggregation types and different distribution positions from the adjustment unit library as the first units. Through balanced selection, the risk of regional failure can be reduced by avoiding the selection of units with too concentrated or single aggregation type or position distribution.

[0031] In some optional embodiments of the first unit balanced selection, an optimization algorithm (such as linear programming, genetic algorithm, etc.) can be used for unit selection according to the scheduling task and the unit characteristics of each unit in the adjustment unit library. A machine learning model can also be introduced to predict the performance of different combined units according to historical scheduling data to assist the optimization algorithm in selection. Thus, the selected first units can efficiently and reliably complete the target power output, while improving the stability and flexibility of the virtual power plant.

[0032] The first scheduling instruction means the specific operation command issued by the virtual power plant to the first units. The first scheduling instruction contains the response target value of each first unit corresponding to each target execution time period. The response target value of each unit can be calculated according to the target power and the unit characteristics of each first unit, so as to generate the first scheduling instruction.

[0033] In step S30, multiple aggregation units matching the aggregation type and distribution position of the first units are selected from the margin unit library as the second units, and the second scheduling instruction for executing the power deviation threshold is issued to the second units.

[0034] In this embodiment, the second units refer to the aggregation units selected from the margin unit library for executing the power deviation threshold. The aggregation type and distribution position of the second units match those of the first units, enhancing the coordination between the adjustment unit library and the margin unit library. The response target value of each second unit is contained in the second scheduling instruction.

[0035] In some embodiments, the second unit balancing selection can be performed based on the execution times of each aggregated unit in the margin unit library, and a plurality of aggregated units matching the aggregation type and distribution position of the first unit are selected from the margin unit library as the second units. The second unit balancing selection includes: selecting a plurality of aggregated units matching the aggregation type and distribution position of the first unit as margin units, sorting the margin units according to the execution times, and dividing the sorting result into a plurality of time intervals, and balancing selecting a number of margin units in each time interval as the second units to ensure that aggregated units with different execution times can have the opportunity to participate in scheduling for testing.

[0036] In step S40, the first units that do not meet the first condition are adjusted to the margin unit library according to the response results of the first scheduling instruction and the second scheduling instruction, and the second units that meet the second condition are adjusted to the adjustment unit library.

[0037] In this embodiment, the first condition includes that the first comprehensive response deviation degree is less than the first preset deviation degree. The second condition includes that the second comprehensive response deviation degree is less than the second preset deviation degree, and the execution time is greater than or equal to the preset time threshold. The response result of the first scheduling instruction contains the execution value of each first unit. The response result of the second scheduling instruction contains the execution value of each second unit.

[0038] In some embodiments, the single-time response deviation degree of each first unit can be calculated according to the first scheduling instruction and the response result of the first scheduling instruction, the first comprehensive response deviation degree of the first unit can be calculated based on the single-time response deviation degree of the first unit and the historical response deviation degree in the historical execution data of the first unit, the first unit that does not meet the first comprehensive response deviation degree being less than the first preset deviation degree is adjusted to the margin unit library, the single-time response deviation degree of each second unit can be calculated according to the second scheduling instruction and the response result of the second scheduling instruction, the second comprehensive response deviation degree of the second unit can be calculated based on the single-time response deviation degree of the second unit and the historical response deviation degree in the historical execution data of the second unit, and the second unit that meets the second comprehensive response deviation degree being less than the second preset deviation degree and the execution time being greater than or equal to the preset time threshold is adjusted to the adjustment unit library.

[0039] For example, the calculation formula of the single-time response deviation degree of an aggregated unit for a to-be-scheduled task is shown in formula (1): (1) Wherein, is the response deviation degree, T is the execution time of all target execution time periods of the to-be-scheduled task, t is a specific execution time, and n is the number of target execution time periods. is the response target value issued to the aggregated unit, This is the execution value of the aggregation unit.

[0040] The formula for calculating the overall response deviation is shown in formula (2): (2) Where cR is the overall response deviation, and m is the number of times the aggregation unit executes the scheduling instruction. The single response deviation when executing the m-th scheduling instruction. The historical response deviation for executing the i-th scheduling instruction.

[0041] Formula (2) combines the single response deviation ( ) and historical response bias ( The overall response deviation (cR) is calculated using this method, which comprehensively reflects the overall response performance of the aggregation unit across multiple scheduling tasks. This comprehensive evaluation method not only considers the unit's response deviation in the current task but also takes into account its historical response performance, avoiding test bias caused by the randomness of a single task, thus more accurately measuring the response capability of the aggregation unit.

[0042] In the virtual power plant task scheduling method proposed in this application, firstly, by separating the tasks to be scheduled in the virtual power plant into target power and power deviation threshold, and executing them respectively by the regulating unit library and the margin unit library, the target power is achieved by the high-precision regulating unit library, thereby ensuring that the overall response deviation of the virtual power plant does not decrease and reducing the risk of grid assessment. Simultaneously, the power deviation threshold is executed by the low-precision / newly connected aggregation unit in the margin unit library for testing or as a fallback. This does not affect grid assessment and settlement, and allows for dynamic testing of the comprehensive response deviation, improving the dynamism and accuracy of response deviation assessment.

[0043] Secondly, when selecting the first unit, a balanced selection is made by considering the distribution location and type of the aggregated units, avoiding the concentration of all first units in the same geographical location or of the same type. This reduces the risk of all first units failing due to localized failures (such as equipment failure in a certain area or natural disasters). This decentralized strategy enhances the robustness of the virtual power plant, ensuring that even if some first units fail, first units in other regions or of different types can continue to operate.

[0044] In a third aspect, by selecting multiple aggregation units that match the aggregation type and distribution location of the first units as the second units, on the one hand, since the second units match the aggregation type and distribution location of the first units, this helps to enhance the coordination within the virtual power plant. When the power grid is disturbed or the demand changes, the second units can understand and adapt to the working state and mode of the first units more quickly, so as to more effectively cooperate with the first units to adjust power and jointly maintain the stable operation of the power grid. At the same time, this matching also reduces the risk of power fluctuations and system instability caused by the lack of coordination between units, and improves the reliability and quality of the entire virtual power plant in providing external power services. On the other hand, this provides an ideal contrast environment for the evaluation of the response deviation degree. Under the same type and location conditions, by comparing the response of the first units and the second units when performing tasks, other interference factors such as inherent differences between different types of units or differences in power grid conditions at different locations can be effectively reduced, so that the response deviation degree of the second units can be more accurately evaluated.

[0045] In one embodiment, as shown in FIG. 2, the step S20 comprises: Figure 2 Step S21, constructing an energy map based on the distribution location of each aggregation unit in the adjustment unit library.

[0046] In this embodiment, each aggregation unit in the adjustment unit library can be mapped to the energy map in the form of a node. Each node in the energy map corresponds to an aggregation unit, and the node position corresponds to the actual geographical position of the aggregation unit.

[0047] Step S22, constructing an energy-time space on the energy map based on the unit characteristics of each node.

[0048] In some embodiments, the energy-time space can be constructed on the energy map based on the target execution time period and the unit characteristics of each node.

[0049] ​Specifically, the target execution time period can be divided into multiple sub-time periods. Based on the unit characteristics of each node, such as aggregation type, real-time power consumption, unit power constraint parameters, and associated real-time weather forecast data, the available power range of each node within the multiple sub-time periods is predicted using techniques such as time series analysis or deep learning models. On the energy map, an energy layer is constructed for each node in each sub-time period. The energy layer is a visualization layer used to represent the available power range of the corresponding aggregation unit within the corresponding sub-time period. The visualization parameters such as the color, shape, and transparency of the energy layer can be adjusted according to the available power range to enhance the visualization effect. The energy layers of each node are superimposed across all sub-time periods to form an energy-time space, where the energy layer change trajectory of each node on the time axis is used to represent the predicted available power of the corresponding aggregation unit in different time periods. This embodiment makes power prediction more refined by dividing the target execution time period into multiple sub-time periods and predicting the available power range of each aggregation unit within each sub-time period.

[0050] For example, for a virtual power plant containing multiple solar photovoltaic units, the energy-time space can display the power generation variation of each photovoltaic unit at different times of the day (such as morning, noon, and evening).

[0051] Step S23: Based on the target power and the target execution time period, perform a first unit equalization selection in the energy-time space to generate multiple sets of candidate power output paths.

[0052] In this embodiment, based on the target power and the target execution time period, multiple candidate power output paths can be generated in the energy-time space using optimization algorithms (such as dynamic programming and genetic algorithms). Each candidate power output path includes a set of multiple aggregation units with different aggregation types and distribution locations.

[0053] From a visualization perspective, the candidate power output path can be viewed as a curve formed in the energy-time space of the energy map. This curve connects aggregation units of different aggregation types and distribution locations within the target execution time period.

[0054] For example, suppose a virtual power plant needs to provide a target power of 100 MW within a target execution period (e.g., from 2 PM to 5 PM). Optimization algorithms can be used to obtain multiple candidate power output paths that meet the 100 MW target power by combining aggregation units of different aggregation types and distribution locations.

[0055] Candidate power output path example one, energy storage unit A (east), discharge power 40 MW; energy storage unit B (west), discharge power 30 MW; adjustable industrial load (middle), adjustment power 20 MW; distributed power generation unit (south), power generation power 10 MW.

[0056] Total power: 40 MW + 30 MW + 20 MW + 10 MW = 100 MW.

[0057] Candidate power output path example two, distributed photovoltaic power station (south), power generation power 50 MW; distributed wind power station (north), power generation power 30 MW; energy storage unit B (west), discharge power 20 MW. Total power: 50 MW + 30 MW + 20 MW = 100 MW.

[0058] Candidate power output path example three, energy storage unit (east), discharge power 25 MW; adjustable industrial load (middle), adjustment power 35 MW; distributed photovoltaic power station (south), power generation power 25 MW; small gas turbine (west), power generation power 15 MW. Total power: 25 MW + 35 MW + 25 MW + 15 MW = 100 MW.

[0059] In step S23, based on the target power and the target execution time period, an optimization algorithm (such as dynamic programming, genetic algorithm) can be used to select the first unit in the energy-time space, and generate a plurality of groups of candidate power output paths. Each candidate power output path includes a plurality of aggregation units with different aggregation types and distribution positions, and also includes corresponding execution time period and response target value. These units work together on the time axis to meet the target power demand.

[0060] Step S24, respectively calculate the comprehensive performance index of each candidate power output path, determine the candidate power output path with the highest comprehensive performance index as the target power output path, and select the plurality of aggregation units included in the target power output path as the first unit.

[0061] The comprehensive performance indicator in this embodiment is a quantitative indicator for comprehensively evaluating the performance of the candidate power output path, which can include economic cost indicators, response deviation indicators, and environmental impact indicators, etc. The economic cost indicators include fuel costs, operation and maintenance costs, etc., which can be calculated by integrating the power output and cost coefficients. The response deviation indicator is used to measure the total response deviation of each candidate power output path, which can be predicted by weighted summation of the historical response deviation of each aggregation unit in the candidate power output path. The weight of each aggregation unit is determined according to its power output proportion. The environmental impact indicator can be calculated according to the emission factor and power output of each aggregation unit. The emission factor reflects the amount of pollutants emitted per unit of electricity generated, such as carbon dioxide, sulfur dioxide, etc. By multiplying the emission factor of each aggregation unit by its power output and then accumulating, the environmental impact indicator can be obtained.

[0062] It can be understood that when evaluating the performance of the candidate power output path, the economic cost indicator, the response deviation indicator, and the environmental impact indicator are all the smaller, the better the performance. Therefore, a comprehensive performance indicator can be calculated by summation, weighted summation, etc. The lower the comprehensive performance indicator, the better the performance. The candidate power output path with the lowest comprehensive performance indicator is determined as the target power output path, and the multiple aggregation units included in the target power output path are selected as the first unit.

[0063] For example, candidate power output path 1: economic cost indicator (yuan), 1000; response deviation indicator (%), 5; environmental impact indicator (kg CO2), 200. Candidate power output path 2: economic cost indicator (yuan), 800; response deviation indicator (%), 8; environmental impact indicator (kg CO2), 150. Candidate power output path 3: economic cost indicator (yuan), 900; response deviation indicator (%), 6; environmental impact indicator (kg CO2), 180.

[0064] Assuming the weight of the economic cost indicator is 0.4, the weight of the response deviation indicator is 0.3, and the weight of the environmental impact indicator is 0.3, the comprehensive performance indicator of candidate power output path 1 is: 0.4x1000+0.3x5+0.3x200=400+1.5+60=461.5. The comprehensive performance indicator of candidate power output path 2 is: 0.4x800+0.3x8+0.3x150=320+2.4+45=367.4. The comprehensive performance indicator of candidate power output path 3 is: 0.4x900+0.3x6+0.3x180=360+1.8+54=415.8.

[0065] Therefore, candidate power output path 2 has the lowest overall performance index, and is selected as the target power output path.

[0066] In the virtual power plant task scheduling processing method proposed in this application embodiment, by constructing an energy map and energy-time space, the distribution location of each aggregation unit and the available power range in different time periods can be intuitively displayed, which improves the accuracy and efficiency of determining the first unit by filtering out the target power output path.

[0067] In one embodiment, such as Figure 3 As shown, in step S30, after selecting multiple aggregation units from the margin unit library that match the aggregation type and distribution location of the first unit as the second unit, and before issuing a second scheduling instruction to the second unit to execute the power deviation threshold, the following steps are included: Step S31: Obtain historical execution data for each second unit, and calculate the power fluctuation value for each second unit based on the historical execution data, unit characteristics, and the power deviation threshold.

[0068] In this embodiment, historical execution data includes historical execution power and the corresponding historical response deviation. The power fluctuation value refers to the magnitude of change in the target response value when the second unit executes historical scheduling instructions.

[0069] Specifically, let the historical execution data sequence H = { , ,..., }, where the i-th scheduling instruction is executed. Includes historical execution power and the corresponding single response deviation Where n is the total number of executions. The characteristics of each second unit include real-time power consumption. and power constraint parameters (including maximum power limit) and minimum power limit ).

[0070] Calculate the response deviation weighting factor based on historical response deviation data. The calculation formula is shown in formula (3): (3) Here, λ is the historical data decay factor, which is used to give higher weight to recent historical execution data.

[0071] For each second unit, the basic fluctuation amplitude B is calculated based on the real-time occupied power, maximum power limit, and minimum power limit. The calculation process is shown in formula (4): (4) The basic fluctuation amplitude B obtained by formula (4) is used to reflect the relative position of the real-time occupied power of the corresponding second unit in its power range, and serves as the basic amplitude of the fluctuation.

[0072] Based on the response deviation weighting factor, the basic volatility B, and the preset influence weight (Empirical data) to calculate the initial fluctuation value The specific calculation process is shown in formula (5): (5) Based on the initial fluctuation value corresponding to each second unit and the previous execution power, the initial sum of the current execution power of all second units is calculated. Then, according to the ratio of the initial sum to the power deviation threshold, the initial fluctuation value is scaled to calculate the power fluctuation value of each second unit.

[0073] Specifically, when the initial sum is greater than the power deviation threshold, the initial fluctuation values ​​of all second units are proportionally reduced according to the ratio of the initial sum to the power deviation threshold to obtain the corresponding power fluctuation values ​​for all second units. When the initial sum is less than the power deviation threshold, the initial fluctuation values ​​of all second units are proportionally increased according to the ratio of the initial sum to the power deviation threshold to obtain the corresponding power fluctuation values ​​for all second units. When the initial sum equals the power deviation threshold, the initial fluctuation value is determined to be the power fluctuation value.

[0074] It should be noted that the power fluctuation value is proportional to the response deviation weighting factor. That is, the lower the response deviation in the historical execution data, the higher the historical response accuracy, and the wider the power range that can be tested, which means the larger the power fluctuation value.

[0075] Step S32: Determine the current execution power of each second unit based on the power fluctuation value and the previous execution power, so that the sum of the current execution power of each second unit matches the power deviation threshold.

[0076] Step S33: Form a second scheduling instruction based on the current execution power of each second unit.

[0077] In this embodiment, the current execution power can be determined by summing the power fluctuation values ​​of each second unit and the previous execution power. Based on the determined second units and the current execution power of each second unit during the target execution time period, a specific operation command, namely the second scheduling instruction, can be generated for the second unit.

[0078] In the virtual power plant task scheduling processing method proposed in this application, firstly, the current execution power of the second unit is adjusted by calculating the power fluctuation value to ensure that each second unit can operate at a different power level during each scheduling. This differentiated testing method helps to comprehensively evaluate the response capability of each unit under different power outputs and avoids test deviations caused by long-term operation at a fixed power. Secondly, the power fluctuation value dynamically determined based on historical execution data allows units with higher historical response accuracy to have a larger power fluctuation range, enabling testing and adjustment within a wider power range. It also avoids allocating excessive power changes to units with low historical response accuracy, reducing power adjustment failures due to insufficient unit performance, thereby improving the overall system efficiency.

[0079] In one embodiment, such as Figure 4 As shown, after step S10, the following steps are also included: Step S11: Obtain the historical deviation records of the margin unit library and input the historical deviation records into the Bayesian estimator to obtain the posterior deviation distribution output by the Bayesian estimator.

[0080] In this embodiment, the historical deviation record includes the target scheduling power and actual output power of the margin unit library in historical scheduling tasks. A Bayesian estimator is a statistical tool that can calculate the posterior probability distribution based on the deviation record between the target scheduling power and the actual output power in historical scheduling tasks. The posterior deviation distribution describes the possible distribution of future deviations, such as the possible range of future deviation values ​​and their corresponding probabilities under different target scheduling power conditions.

[0081] In some implementations, historical deviation records and historical covariate information from the margin unit library can be obtained. The historical covariate information includes at least one of weather type (sunny, cloudy, rainy, etc.) and holiday type (weekday, weekend, statutory holiday, etc.). Current covariate information corresponding to the historical covariate information is obtained. The information type of the current covariate information is consistent with the type of the historical covariate information. Specifically, the current weather type for executing the scheduled task can be obtained through a weather forecast interface, and whether the date for executing the scheduled task is a holiday can be determined through a calendar service. The historical deviation records, the historical covariate information, and the current covariate information are input into a Bayesian estimator to obtain the posterior deviation distribution output by the Bayesian estimator. By integrating the covariate information, the predictions of the Bayesian estimator are closer to the actual conditions, improving the accuracy of deviation prediction.

[0082] Step S12: Determine the margin power based on the posterior deviation distribution and the power deviation threshold.

[0083] Based on the posterior deviation distribution, the power deviation threshold is taken as the target scheduling power, the value range of the future deviation and the corresponding probability are obtained, and the value corresponding to the maximum probability is taken as the deviation value (which may be positive, negative or zero). The sum of the deviation value and the power deviation threshold is calculated to obtain the margin power, which is the power that can be additionally reserved based on the target power.

[0084] For example, assuming that the target power is 10 MW, the power deviation threshold is 2 MW, and the posterior deviation distribution shows that when the target scheduling power is 2 MW, the deviation value is between +1 MW and -1 MW, and the deviation value corresponding to the maximum probability is +0.5 MW. Then the margin power is 2 MW + (+0.5 MW) = 2.5 MW.

[0085] Further, in step S20, the second unit is instructed to execute the second scheduling instruction of the power deviation threshold, including: the second unit is instructed to execute the second scheduling instruction of the margin power, so that the total power output by the second unit matches the power deviation threshold.

[0086] In the processing method for virtual power plant task scheduling provided by the embodiments of the present application, the most possible deviation value is predicted through the posterior deviation distribution, so that the calculation of the margin power is more accurate, and the accuracy of the virtual power plant in responding to the grid scheduling is improved. Moreover, the deviation value and the power deviation threshold are combined to calculate the margin power, so as to ensure that the total power actually output by the second unit can meet the grid evaluation requirements and avoid excessive or insufficient reservation, so that the virtual power plant can operate in a more economical way under the premise of meeting the grid evaluation requirements, thereby reducing the operating cost.

[0087] In one embodiment, a computer storage medium is provided, which stores executable instructions, and the instructions are executed by a processor to make the processor execute the steps in the above method embodiments.

[0088] In one embodiment, an electronic device is also provided, which includes one or more processors, and a memory storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the steps in the above method embodiments.

[0089] In one embodiment, as Figure 5As shown, it shows a structural schematic diagram of an electronic device for implementing the embodiments of the present application. The electronic device includes a central processing unit (CPU) 101, which can perform various appropriate actions and processes according to programs stored in a read-only memory (ROM) 102 or programs loaded from a storage portion 108 into a random access memory (RAM) 103. In the RAM 103, various programs and data required for the operation of the electronic device are also stored. The CPU 101, the ROM 102, and the RAM 103 are connected to each other through a bus 104. An input / output (I / O) interface 105 is also connected to the bus 104.

[0090] The following components are connected to the I / O interface 105: an input portion 106 including a keyboard, a mouse, and the like; an output portion 107 including a cathode ray tube (CRT), a liquid crystal display (LCD), and the like, and a speaker, and the like; a storage portion 108 including a hard disk, and the like; and a communication portion 109 including a network interface card such as a LAN card, a modem, and the like. The communication portion 109 performs communication processing via a network such as the Internet. A drive 110 is also connected to the I / O interface 105 as necessary. A removable medium 111 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is mounted on the drive 110 as necessary, so that a computer program read therefrom is installed in the storage portion 108 as necessary.

[0091] In particular, according to the embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, the embodiments of the present application include a computer program product including instructions carried on a computer-readable medium, in such embodiments, the instructions can be downloaded and installed from a network by the communication portion 109, and / or installed from the removable medium 111. When the instructions are executed by the central processing unit (CPU) 101, the various method steps described in the present application are performed.

[0092] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, but not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

[0093] Furthermore, to the extent that the terms "comprises", "comprising", "includes", "including" and "has" or any variation thereof are used in the specification and / or claims, these terms are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises, includes or has a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus. In addition, to the extent that the terms "coupled", "coupling", "connected", "connecting" or "connection" are used in the specification and / or claims, these terms are intended to mean an indirect, direct, optical, optical, mechanical, mechanical, electrical, and / or electromagnetic coupling, connection, communication, and / or interaction of the tangible elements under discussion, and / or of the intangible electrical signals conveyed by a tangible electrical connection and / or in a tangible electromagnetic field. In addition, where appropriate, the terms "coupled", "coupling", "connected", "connecting" or "connection" are intended to mean the joining of two members directly or indirectly to one another. Such joining can be stationary in nature or movable in nature. Such joining can be mechanical, electrical, and / or hydraulic in nature. In addition, such joining can occur with the two members coupled directly to each other or with a third member positioned in between the two members so as to mechanically, electrically

Claims

1. A processing method for virtual power plant task scheduling, characterized in that, The method comprises: receiving a task to be scheduled, the task to be scheduled comprising a target power, a target execution time period and a power deviation threshold, the power deviation threshold being less than the target power; performing first cell balancing selection according to the task to be scheduled and cell characteristics of each aggregated cell in a regulating cell library, selecting a plurality of aggregated cells with different aggregation types and distribution positions from the regulating cell library as first cells, and issuing a first scheduling instruction to the first cells to execute the target power, the regulating cell library being composed of aggregated cells with a first comprehensive response deviation degree less than a first preset deviation degree; selecting a plurality of aggregated cells matching the aggregation types and distribution positions of the first cells from a margin cell library as second cells, and issuing a second scheduling instruction to the second cells to execute the power deviation threshold, the margin cell library being composed of aggregated cells with a second comprehensive response deviation degree greater than or equal to a second preset deviation degree and aggregated cells with an execution frequency less than a preset frequency threshold; adjusting the first cells that do not meet a first condition to the margin cell library and adjusting the second cells that meet a second condition to the regulating cell library according to response results of the first scheduling instruction and the second scheduling instruction.

2. The processing method of virtual power plant task scheduling according to claim 1, characterized in that, The cell characteristics comprise aggregation type, distribution position, real-time occupied power and cell power constraint parameter, and the first cell balancing selection from the regulating cell library comprises: constructing an energy map based on the distribution positions of each aggregated cell in the regulating cell library, each node in the energy map corresponding to an aggregated cell; constructing an energy-time space on the energy map based on the cell characteristics of each node; performing first cell balancing selection in the energy-time space based on the target power and the target execution time period, generating a plurality of groups of candidate power output paths, each candidate power output path comprising a plurality of aggregated cells with different aggregation types and distribution positions; calculating comprehensive performance indicators of each candidate power output path respectively, determining a candidate power output path with the lowest comprehensive performance indicator as a target power output path, and selecting the plurality of aggregated cells included in the target power output path as first cells.

3. The method of claim 2, wherein the method further comprises: The construction of the energy-time space on the energy map based on the cell characteristics of each node comprises: dividing the target execution time period into a plurality of sub-time periods, and predicting available power ranges of each node in the plurality of sub-time periods based on the cell characteristics of each node; constructing an energy layer for each node in each sub-time period on the energy map, the energy layer being used to represent the available power range of the corresponding aggregated cell in the corresponding sub-time period; stacking the energy layers of each node in all sub-time periods to form an energy-time space, wherein the energy layer change track of each node on the time axis is used to represent the predicted available power of the corresponding aggregated cell in different time periods.

4. The processing method of virtual power plant task scheduling according to claim 1, characterized in that, The selecting, from the margin unit library, a plurality of aggregation units matching the aggregation type and distribution position of the first unit as second units comprises: The second unit balancing selection is performed based on the execution times of each aggregation unit in the margin unit library, and a plurality of aggregation units matching the aggregation type and distribution position of the first unit are selected from the margin unit library as second units; Before the second scheduling instruction of executing the power deviation threshold is issued to the second units, the method comprises: Obtaining historical execution data of each second unit, and calculating a power fluctuation value of each second unit based on the historical execution data, unit characteristics and the power deviation threshold; Determining a current execution power of each second unit based on the power fluctuation value and the last execution power, so that the sum of the current execution powers of each second unit matches the power deviation threshold; Forming a second scheduling instruction based on the current execution power of each second unit.

5. The processing method of virtual power plant task scheduling according to claim 4, characterized in that, The historical execution data comprises historical execution power and corresponding historical response deviation degrees, and the unit characteristics comprise real-time occupied power and unit power constraint parameters. The calculation of the power fluctuation value of each second unit based on the historical execution data, unit characteristics and the power deviation threshold comprises: Calculating a response deviation degree weight factor based on the historical response deviation degrees in the historical execution data; Calculating a basic fluctuation amplitude based on the real-time occupied power and the unit power constraint parameters; Calculating an initial fluctuation value of each second unit based on the response deviation degree weight factor, the basic fluctuation amplitude and a preset influence weight; Calculating an initial sum of the current execution powers of all second units based on the initial fluctuation value of each second unit and the last historical execution power; Scaling the initial fluctuation value of each second unit according to the ratio of the initial sum to the power deviation threshold, and calculating the power fluctuation value of each second unit.

6. The processing method of virtual power plant task scheduling according to claim 1, wherein, The first condition comprises that a first comprehensive response deviation degree is less than a first preset deviation degree. The second condition comprises that a second comprehensive response deviation degree is less than a second preset deviation degree, and the execution times are greater than or equal to a preset times threshold. The first scheduling instruction comprises a response target value of each first unit, the second scheduling instruction comprises a response target value of each second unit, the response result of the first scheduling instruction comprises an execution value of each first unit, and the response result of the second scheduling instruction comprises an execution value of each second unit. The adjusting, according to the response results of the first scheduling instruction and the second scheduling instruction, the first units that do not meet the first condition to the margin unit library and the second units that meet the second condition to the adjustment unit library comprises: Calculating a single response deviation degree of each first unit according to the first scheduling instruction and the response result of the first scheduling instruction; Calculating a first comprehensive response deviation degree of the first unit based on the single response deviation degree of the first unit and the historical response deviation degree in the historical execution data of the first unit; Adjusting the first units that do not meet the first comprehensive response deviation degree being less than the first preset deviation degree to the margin unit library. According to the second scheduling instruction and a response result of the second scheduling instruction, a single-time response deviation degree of each second unit is calculated; According to the single-time response deviation degree of the second unit and a historical response deviation degree in historical execution data of the second unit, a second comprehensive response deviation degree of the second unit is calculated; The second unit that meets the second comprehensive response deviation degree being less than a second preset deviation degree and the execution times being greater than or equal to a preset number threshold is adjusted to a regulation unit library.

7. The processing method of virtual power plant task scheduling according to claim 1, wherein, After the receiving of the task to be scheduled, the method further includes: obtaining historical deviation records of the margin unit library, and inputting the historical deviation records into a Bayesian estimator to obtain a posterior deviation distribution output by the Bayesian estimator; determining a margin power according to the posterior deviation distribution and the power deviation threshold; the issuing of the second scheduling instruction of the second unit to execute the power deviation threshold includes: issuing a second scheduling instruction of the second unit to execute the margin power, so that the total power output by the second unit matches the power deviation threshold.

8. The processing method of virtual power plant task scheduling according to claim 7, characterized in that, The obtaining of the historical deviation records of the margin unit library and the inputting of the historical deviation records into the Bayesian estimator to obtain the posterior deviation distribution output by the Bayesian estimator includes: obtaining historical deviation records and historical covariate information of the margin unit library, the historical covariate information including at least one of weather types and holiday types; obtaining current covariate information corresponding to the historical covariate information; inputting the historical deviation records, the historical covariate information and the current covariate information into the Bayesian estimator to obtain the posterior deviation distribution output by the Bayesian estimator.

9. An electronic device, comprising: include: one or more processors; a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the processing method of the virtual power plant task scheduling according to any one of claims 1 to 8.

10. A computer storage medium, characterized in that, The storage medium stores executable instructions, and the instructions make the processor execute the processing method of the virtual power plant task scheduling according to any one of claims 1 to 8 when executed by the processor.

Images