Energy scheduling method, device and equipment based on virtual power plant and storage medium

Abstract

The invention discloses an energy scheduling method, device and equipment based on a virtual power plant, and a storage medium. The method comprises the following steps: obtaining corrected predicted power of distributed renewable energy sources of the virtual power plant; a distribution optimization model of the virtual power plant is constructed, the distribution optimization model comprises a target function and constraint conditions, the target function takes response cost, energy storage cost, regulation and control cost and carbon transaction cost of the virtual power plant as targets, and the constraint conditions are charge and discharge characteristics of a distributed power supply; in the actual operation process, the corrected and predicted power is input into the distribution optimization model, solving optimization is carried out through a path optimization algorithm in an uncertain environment, and a scheduling strategy is obtained; and performing energy scheduling on the power distribution network accessed to the virtual power plant according to the scheduling strategy, thereby realizing cooperative and clustered operation of various power resources in the virtual power plant, improving the operation efficiency of the power distribution network, reducing energy waste, and improving the reliability of overall energy supply.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network optimization and dispatching, and in particular to an energy dispatching method, device, equipment and storage medium based on a virtual power plant. Background Art

[0002] At present, society has an increasingly strong demand for environmental protection and energy conservation. The traditional power system, due to its heavy reliance on fossil energy, produces a large amount of carbon emissions, and has limitations in configuration methods and operating mechanisms, is difficult to cope with complex and changing power demand and environmental protection requirements. In order to solve these problems, the power system must be highly intelligent and green. Among them, virtual power plants, as a new type of power system operation mode, integrate a variety of flexible resources such as renewable energy, power storage equipment, and power demand response. It can optimize operation in a long-term, large-scale, and multi-objective manner, improve energy utilization efficiency, reduce carbon emissions, and achieve green and environmental protection effects.

[0003] Virtual power plants involve a variety of power resources, each of which has its own operating strategies and limitations. However, the current operating strategies of virtual power plants ignore the coordinated and clustered operation of multiple energy resources within the virtual power plants, resulting in low overall performance and efficiency of the power system. Summary of the invention

[0004] This application proposes an energy scheduling method, device, equipment and storage medium based on a virtual power plant, which realizes multi-energy coordination and clustered operation within the virtual power plant, thereby helping to improve the overall performance and efficiency of the power system.

[0005] In a first aspect, a method for energy scheduling based on a virtual power plant is provided, comprising:

[0006] Obtaining the corrected forecast power of distributed renewable energy sources of the virtual power plant;

[0007] Constructing a distribution optimization model of the virtual power plant, the distribution optimization model includes an objective function and constraints, the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power source;

[0008] In the actual operation process, the corrected predicted power is input into the distribution optimization model, and the optimization is solved by the path optimization algorithm under the uncertainty environment to obtain the scheduling strategy;

[0009] Energy dispatch is performed on the distribution network connected to the virtual power plant according to the dispatch strategy.

[0010] In a second aspect, an energy dispatching device based on a virtual power plant is provided, comprising:

[0011] An acquisition module, used for acquiring the corrected predicted power of distributed renewable energy of the virtual power plant;

[0012] A construction module is used to construct a distribution optimization model of the virtual power plant, wherein the distribution optimization model includes an objective function and constraints, wherein the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power source;

[0013] An optimization module, used for inputting the corrected predicted power into the distribution optimization model during actual operation, and performing optimization through a path optimization algorithm under an uncertain environment to obtain a scheduling strategy;

[0014] A scheduling module is used to schedule energy for the distribution network connected to the virtual power plant according to the scheduling strategy.

[0015] Optionally, in some embodiments of the present application, the acquisition module includes:

[0016] The first acquisition submodule is used to obtain the initial predicted power and average prediction error of the distributed renewable energy of the virtual power plant;

[0017] The obtaining submodule is used to obtain the corrected predicted power based on the initial predicted power and the average prediction error.

[0018] Optionally, in some embodiments of the present application, the building blocks include:

[0019] The modeling submodule is used to optimize the virtual power plant scheduling by using the distributed robust optimization model to obtain the distributed optimization model, wherein the distributed optimization model includes two stages. In the first stage, the planned power is determined according to the corrected predicted power; in the second stage, the actual generated power and the initial predicted power or the deviation of the corrected predicted power are balanced according to the flexibility resources of the virtual power plant.

[0020] Optionally, in some embodiments of the present application, the distribution optimization model is expressed as follows:

[0021] min x {c T x+supE F [Q(x, ζ%)]}

[0022] stAx≤b

[0023] Among them, x is the decision variable in the first stage, c T is the coefficient vector of the objective function in the first stage, A is the coefficient vector of the constraint conditions in the first stage, b is the constant vector of the constraint conditions in the first stage, supE F[Q(x, ζ%)] is the objective function of the second stage, which represents the expected adjustment cost of F∈F when the random variable ζ% obeys the worst probability distribution; F is the true distribution and F is the distribution fuzzy set.

[0024] Optionally, in some embodiments of the present application, the expected adjustment cost is expressed as:

[0025] Q(x, ζ%) = mind T y(x, ζ%)

[0026] st$Gy(x,ζ%)≤h(ζ%)

[0027] Among them, d T is the coefficient vector of the objective function in the second stage, G is the coefficient vector of the constraint conditions in the second stage, and h(ζ%) is the constant vector of the constraint conditions in the second stage.

[0028] Optionally, in some embodiments of the present application, the optimization module includes:

[0029] Initialization submodule, used to initialize the population based on chaotic mapping;

[0030] The screening submodule is used to screen target individuals from the initialized population;

[0031] The search submodule is used to perform position search in the search space based on the screened target individuals until convergence to obtain the scheduling strategy.

[0032] Optionally, in some embodiments of the present application, the search submodule includes:

[0033] The search unit is used to perform position search in the search space based on the screened target individuals by adopting a diverse global optimal guidance strategy with inertia weights and a double sample learning strategy.

[0034] In a third aspect, a computer device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned energy scheduling method based on a virtual power plant when executing the computer program.

[0035] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned energy scheduling method based on a virtual power plant are implemented.

[0036] The present application provides an energy scheduling method, device, computer equipment and storage medium based on a virtual power plant, which obtains the corrected predicted power of the distributed renewable energy of the virtual power plant; constructs a distribution optimization model of the virtual power plant, and the distribution optimization model includes an objective function and constraints, the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power sources; in the actual operation process, the corrected predicted power is input into the distribution optimization model, and the solution is optimized by a path optimization algorithm under an uncertain environment to obtain a scheduling strategy; according to the scheduling strategy, energy scheduling is performed on the distribution network connected to the virtual power plant. In the energy dispatching scheme based on virtual power plant provided in this application, by obtaining the corrected predicted power, the power generation of distributed renewable energy can be predicted more accurately, the prediction error can be reduced, the target work function can effectively reduce the comprehensive operating cost of virtual power plant, improve economic benefits, and the constraint conditions are based on the charging and discharging characteristics of distributed power sources, ensuring that the actual operating limitations of the equipment are fully considered during the dispatching process, avoiding equipment overload or damage, and enhancing the stability and reliability of the system. In an uncertain environment, the distribution optimization model is solved and optimized through the path optimization algorithm, which can flexibly respond to various emergencies, such as weather changes, load fluctuations, etc., to ensure the real-time and adaptability of the dispatching strategy, improve the utilization rate of distributed renewable energy, reduce the phenomenon of wind and light abandonment, and promote the efficient consumption of renewable energy. According to the dispatching strategy, the distribution network connected to the virtual power plant is dispatched for energy, and the coordination and clustering operation of various power resources within the virtual power plant are realized, which can improve the operating efficiency of the distribution network, reduce energy waste, and improve the reliability of the overall energy supply. It can be seen that the efficient dispatching of distributed renewable energy in the virtual power plant is achieved, and the economic benefits and system stability are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 An application environment diagram of the energy scheduling method based on a virtual power plant provided in an embodiment of the present application;

[0039] Figure 2 A flowchart of an energy scheduling method based on a virtual power plant provided in an embodiment of the present application;

[0040] Figure 3 A schematic diagram of a curve of an inertia weight iteration process provided in an embodiment of the present application;

[0041] Figure 4 A search schematic diagram of a dual-sample learning strategy provided in an embodiment of the present application;

[0042] Figure 5 A structural block diagram of an energy dispatching device based on a virtual power plant provided in an embodiment of the present application;

[0043] Figure 6 A structural block diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0046] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0047] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0048] The energy scheduling method based on virtual power plant provided by the embodiment of the present invention can be applied in Figure 1 In the application environment, the computer device 110 communicates with the server 120 via the network 130. The computer device 110 can obtain the corrected predicted power of the distributed renewable energy of the virtual power plant;

[0049] A distribution optimization model of the virtual power plant is constructed, wherein the distribution optimization model includes an objective function and constraints, wherein the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power sources; during actual operation, the corrected predicted power is input into the distribution optimization model, and the solution is optimized by a path optimization algorithm under an uncertain environment to obtain a scheduling strategy; energy scheduling is performed on the distribution network connected to the virtual power plant according to the scheduling strategy, and the results are displayed through a computer device 110. In the present invention, by obtaining the corrected predicted power, the power generation of distributed renewable energy can be predicted more accurately, and the prediction error can be reduced. The target work function can effectively reduce the comprehensive operating cost of the virtual power plant and improve the economic benefits. The constraint conditions are based on the charging and discharging characteristics of the distributed power source, ensuring that the actual operating limitations of the equipment are fully considered during the scheduling process, avoiding equipment overload or damage, and enhancing the stability and reliability of the system. In an uncertain environment, the distribution optimization model is solved and optimized by the path optimization algorithm, which can flexibly respond to various emergencies, such as weather changes, load fluctuations, etc., to ensure the real-time and adaptability of the scheduling strategy, improve the utilization rate of distributed renewable energy, reduce the phenomenon of wind and light abandonment, and promote the efficient consumption of renewable energy. According to the scheduling strategy, the distribution network connected to the virtual power plant is energy dispatched, and the coordination and clustering operation of various power resources within the virtual power plant are realized, which can improve the operating efficiency of the distribution network, reduce energy waste, and improve the reliability of the overall energy supply. It can be seen that the efficient scheduling of distributed renewable energy in the virtual power plant is achieved, and the economic benefits and system stability are improved. Among them, the computer device 110 can be but not limited to various smart phones 110-1, tablet computers 110-2 and laptop computers 110-3. The present invention is described in detail below through specific embodiments.

[0050] See also Figure 2 As shown, Figure 2 A flowchart of an energy scheduling method based on a virtual power plant provided in an embodiment of the present invention is provided. The method can be applied to both a terminal and a server. This embodiment is illustrated by applying to a server. The energy scheduling method based on a virtual power plant includes the following steps:

[0051] S101: Obtain the corrected predicted power of distributed renewable energy of the virtual power plant.

[0052] A virtual power plant (VPP) is a system that integrates multiple distributed energy resources (i.e. distributed renewable energy such as wind power generation, photovoltaics, energy storage equipment, and controllable loads) through software and communication technology. It can uniformly manage and dispatch these decentralized distributed renewable energy sources to improve the flexibility and efficiency of the power system).

[0053] Corrected predicted power refers to a more accurate predicted power obtained by correcting and adjusting the initial predicted power of distributed renewable energy (such as solar photovoltaic, wind power, etc.) within a preset time period (such as a time period with months, seasons, years, etc. as time units). The initial predicted power is usually generated based on historical power data (including historical power and meteorological data, etc.), but the initial predicted power deviates from the predicted value due to various factors (such as weather changes, equipment failures, etc.). Therefore, through the correction process, the accuracy and reliability of the corrected predicted power can be improved, and the dispatch and operation of the power system can be better guided.

[0054] Optionally, in view of the problems of data missing, data anomalies and noise points caused by network communication anomalies, insufficient statistical accuracy, data monitoring equipment transformation, etc. during the collection, transmission and storage of monitoring data of each distributed power plant, and taking into account the power characteristics of new energy power generation such as wind power, photovoltaic power and hydropower, the historical power data can be preprocessed. The preprocessing stage includes steps such as data cleaning, data filling and data standardization of the historical power data.

[0055] Data cleaning is mainly to remove obvious outliers in historical power data caused by equipment failure, human error, etc. For example, random forest, isolation forest and other methods can be used to detect abnormal scores of historical power data (which may include power generation, light intensity, wind speed, temperature, etc.). If there is an abnormal score that meets the preset abnormal score value, the value is removed.

[0056] Data filling solves the problem of missing power data due to various reasons. Common methods include interpolation, averaging, nearest neighbor method, etc.; for example, the power generation data of distributed photovoltaic power stations (which may include power generation, light intensity, wind speed, temperature, etc. of wind power, photovoltaic, hydropower, etc.), where some data points are missing, linear interpolation can be used to estimate the missing values ​​based on the values ​​of adjacent data points, thereby performing data filling.

[0057] Data standardization is to convert historical power data into dimensionless relative values, which can reduce the dimension impact of historical power data and make data with different characteristics have the same scale. For example, the historical power data of new energy sources such as wind power, photovoltaic power, and hydropower (such as temperature, power generation, light intensity, wind speed, etc.) are normalized and the data is mapped to [0, 1]. For example, the calculation formula for normalizing power generation is as follows:

[0058]

[0059] Among them, w′ is the normalized result value, w is the power generated by the historical power data, and w maxis the maximum value of the generated power in the historical power data, w min It is the minimum value of the generated power in the historical power data.

[0060] Further considering the uncertainty of renewable energy such as wind power and photovoltaic power in the virtual power plant, the randomness of its output power will affect the internal coordinated operation of the virtual power plant and external market transactions. The uncertainty factors can be corrected by prediction errors. That is, in one embodiment, the correction prediction power of distributed renewable energy of the virtual power plant is obtained, which includes:

[0061] Obtain the initial predicted power and average prediction error of distributed renewable energy of the virtual power plant;

[0062] The corrected prediction power is obtained based on the initial prediction power and the average prediction error.

[0063] The historical power generation of distributed renewable energy in virtual power plants, such as wind power, photovoltaic power, energy storage equipment, and controllable loads, can be collected in the past period of time (such as the past year); the historical power generation data (including historical power generation and meteorological data) can be used for power forecasting through a preset forecasting model to obtain the initial forecast power for a period of time in the future (such as the next year). The forecasting model includes linear regression, time series analysis, machine learning algorithms (such as random forests, neural networks), etc.

[0064] The actual power generation and initial forecast power at each time point in the historical time period can be obtained, and the initial forecast power and the actual operating power can be subtracted. Finally, the power error between the actual power generation and the initial forecast power at each time point (such as day, hour, minute, second, etc.) can be calculated, and the average of each power deviation can be calculated to obtain the average forecast error. The initial forecast power and the average forecast error can be calculated to obtain the corrected forecast power.

[0065] For example, assuming that the average prediction error includes the average prediction error of wind power, the corrected predicted power of wind power is:

[0066] P WT，t % = P WT，t +ζ WT，t %

[0067] Among them, P WT，t % is the corrected predicted power of wind power, ζ WT，t % is the average prediction error of wind power, P WT，t is the initial predicted power of wind power.

[0068] For example, assuming that the average prediction error includes the average prediction error of photovoltaics, the corrected predicted power of photovoltaics is:

[0069] PPV，t % = P PV，t +ζ PV，t %

[0070] Among them, P PV，t % is the corrected predicted power of photovoltaic, ζ PV，t % is the average prediction error of photovoltaic, P PV，t is the initial predicted power of PV.

[0071] In addition, the prediction error between wind power and photovoltaic power can be calculated based on the average prediction error of wind power and the average prediction error of photovoltaic power. This prediction error can be used to adjust and optimize the scheduling strategy in the internal coordinated operation of the virtual power plant (for example, if the prediction error of wind power is large, the capacity of the energy storage system can be increased to smooth the output fluctuation of wind power) and to determine the degree of participation of demand-related resources (such as wind power and photovoltaic power) to balance the supply and demand relationship. The prediction error ζ t The calculation formula of % is as follows:

[0072] ζ t % = ζ WT，t %+ζ PV，t %

[0073] S102: Constructing a distribution optimization model of the virtual power plant, the distribution optimization model includes an objective function and constraints, the objective function targets the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant, and the constraints are the charging and discharging characteristics of the distributed power sources.

[0074] Among them, the distribution optimization model includes two stages. In the first stage, the planned power is determined according to the corrected predicted power of wind power and the corrected predicted power of photovoltaic power. The planned power refers to the target power in the power generation and load scheduling plan pre-formulated according to the corrected power generation power. The planned power is the core part of the virtual power plant scheduling strategy, which is used to guide the operating status and output power of various distributed energy resources (such as power generation equipment, energy storage systems, controllable loads, etc.) in different time periods. In the second stage, the deviation between the actual operating power of wind power and photovoltaic power and the corrected predicted power is balanced by the flexibility resources in the virtual power plant, so as to determine the final actual operating power.

[0075] The constraints include first-stage constraints and second-stage constraints. In the first stage, the planned operating power is determined based on the corrected predicted power of wind power and photovoltaic power, and conventional constraints such as equipment operation and power balance are met. In the second stage, under a given state (i.e., the planned power determined by the first-stage optimization decision), the power of flexible resources is adjusted to balance the deviations of wind power and photovoltaic power, including combined heat and power (CHP), energy storage system (ESS), controllable load / demand response (CC), power to gas (P2G), shiftable load (SL), and interruptible load (IL). That is, based on the affine strategy, the planned power of the first stage is used to represent the actual operating power of the second stage. Taking CHP electric power as an example, the actual electric power of the second stage can be expressed as:

[0076]

[0077] Among them, P CHP,t % is the actual electrical power of the second stage of CHP (i.e., actual operating power); is the adjustment coefficient of CHP electric power,

[0078] Exemplarily, the first-stage constraints may include wind power and photovoltaic constraints, cogeneration operation constraints, gas boiler operation constraints, energy storage system operation constraints, gas tank operation constraints, carbon capture and storage operation constraints, power-to-gas operation constraints and demand response constraints.

[0079] Wind power and photovoltaic constraints are expressed as:

[0080]

[0081]

[0082] in, is the initial predicted power of wind power, is the abandoned power of wind power, P WT，t % is the corrected predicted power of wind power (here it is the actual operating power during actual operation), is the initial predicted power of photovoltaic, is the abandoned power of photovoltaic power, P PV，t % is the corrected predicted power of photovoltaic (here it is the actual operating power during actual operation).

[0083] The core components of cogeneration are gas turbines and waste heat boilers. The gas turbine generates electricity by burning natural gas. The heat generated in this process is recovered by the waste heat boiler and can be used for heating. The cogeneration operation constraints are expressed as:

[0084]

[0085] Among them, a CHP and b CHP is the linear coefficient and constant term of natural gas consumption by CHP when providing electricity and heat. k and b are the slope and intercept of the corresponding straight line, respectively. In addition, the virtual power plant participates in the power auxiliary service market. The frequency regulation capacity and reserve capacity of CHP shall not exceed a certain range of electric power and shall meet power constraints and ramping requirements. The constraints are as follows:

[0086]

[0087]

[0088]

[0089] in, and is the minimum and maximum electrical power of CHP, and Provide frequency regulation, upper reserve and lower reserve capacity for CHP, μ CHP,F and μ CHP,R The maximum proportion of frequency regulation capacity and reserve capacity is and is the maximum up and down ramp of CHP electric power, and is the minimum and maximum thermal power of CHP, and It is the maximum up and down ramp of CHP thermal power.

[0090] Gas boiler (GB) uses natural gas for heating and needs to meet power constraints and ramp constraints. The operating constraints of the gas boiler are expressed as:

[0091]

[0092]

[0093] in, is the calorific value of natural gas, η GB is the GB heating efficiency, u GB,t GB start / stop status 0-1 variable, and is the minimum and maximum power of GB, and It is the maximum uphill and downhill climbing power of GB.

[0094] ESS needs to meet power and state constraints and capacity constraints. After the virtual power plant participates in the power auxiliary service market, ESS can provide frequency regulation and upper backup auxiliary services when discharging, and lower backup auxiliary services when charging. The operation constraints of the energy storage system are expressed as:

[0095]

[0096] in, and 0-1 variables for ESS charging and discharging states; and is the minimum and maximum charging and discharging power of ESS; and is the charging power and discharging power of ESS at time t; and Provide frequency regulation, upper reserve and lower reserve capacity for ESS; μ ESS,R and μ ESS,R is the maximum proportion of frequency regulation capacity and reserve capacity; η ESS is the ESS charging and discharging efficiency; C ESS,t is the power storage capacity of ESS at time t; and is the minimum and maximum storage capacity of ESS; C ESS,0 and C ESS,T The power storage capacity of ESS at the initial and end times.

[0097] The operating constraints of gas storage tanks (GST) are the same as those of ESS, but there is no need to consider the constraints related to frequency regulation and reserve. The stored natural gas comes from power-to-gas, and the released natural gas enters CHP or GB.

[0098] The carbon dioxide captured by Carbon Capture and Storage (CCS) is eventually permanently stored or used by P2G. The electricity consumed by capture can be divided into basic energy consumption and operating energy consumption. CCS needs to meet power constraints and ramp constraints. The operating constraints of carbon capture and storage are expressed as:

[0099] Q CC,t =Q CS,t +Q P2G,t

[0100]

[0101] Among them, Q CS,t and Q P2G,t are the carbon dioxide stored in CSS and utilized in P2G; P CC,t is the energy consumption of CCS at time t; uCC,t CCS start / stop status 0-1 variable; P CCb and They are the basic energy consumption and operating energy consumption coefficient of CCS capture respectively.

[0102] The natural gas produced by P2G has a certain relationship with the carbon dioxide used and also has a certain relationship with the electricity consumed. P2G needs to meet power constraints and ramp constraints. The power-to-gas operation constraints are expressed as:

[0103]

[0104] in, is the density of carbon dioxide; V P2G,t The amount of natural gas produced by P2G; η P2G is the P2G operation efficiency; P P2G,t The power consumed by P2G.

[0105] For transferable loads, the total amount constraint before and after the transfer and the response degree constraint are met. For interruptible loads, the demand response constraint is expressed as:

[0106]

[0107]

[0108] in, and P SL,t out are the loads transferred in and out at time t; and 0-1 variables for SL transfer in and out status; is the maximum transferable load; P IL,t is the load interrupted at time t; u IL,t BL interrupt status 0-1 variable; is the maximum interruptible load.

[0109] The spot market transaction constraints are expressed as:

[0110]

[0111] in, and It is a 0-1 variable for the state of electricity purchase and sale; The maximum transaction power.

[0112] The power balance constraint is expressed as:

[0113]

[0114] Among them, P WT,t and P PV,t Power prediction for wind power and photovoltaic power; PBL,t and are the rigid electrical load and thermal load of VPP.

[0115] In one embodiment, constructing the distribution optimization model of the virtual power plant includes:

[0116] The virtual power plant scheduling is optimized and modeled using a distributed robust optimization model to obtain the distributed optimization model, wherein the distributed optimization model includes two stages. In the first stage, the planned power is determined based on the corrected predicted power; in the second stage, the actual generated power and the initial predicted power or the deviation of the corrected predicted power are balanced based on the flexibility resources of the virtual power plant.

[0117] In one embodiment, the distribution optimization model is expressed as follows:

[0118] min x {c T x+supE F [Q(x, ζ%)]}

[0119] stAx≤b

[0120] Among them, x is the decision variable in the first stage, c T is the coefficient vector of the objective function in the first stage, A is the coefficient vector of the constraint conditions in the first stage, b is the constant vector of the constraint conditions in the first stage, is the objective function of the second stage, which represents the expected adjustment cost when the random variable ζ% obeys the worst probability distribution; F is the true distribution, and F is the distribution fuzzy set.

[0121] It should be noted that the decision variable x in the first stage is a variable that can be freely changed in the distribution optimization model and is used to make decisions in the distribution optimization model. In the first stage, the decision variable x is used to determine the planned power. The coefficient vector c T is the coefficient of the decision variable in the objective function in the first stage, defining the influence of each decision variable x on the objective function. TUsed to construct objective functions to minimize costs or maximize operating benefits. The coefficient vector A and the constant vector b define the constraints that the decision variables need to meet. These constraints can be resource constraints, technical constraints, or other business rules. In mathematical expressions, constraints are usually expressed as Ax≤b. The random variable ζ% represents the uncertainty factors in the distribution optimization model, such as the deviation between the actual operating power of wind power and photovoltaic power and the corrected predicted power, the prediction error between wind power and photovoltaic power, etc. The expected adjustment cost refers to the additional cost caused by the uncertainty of the random variable in the worst case. It is part of the objective function of the second stage and is used to evaluate the robustness of the first stage decision under uncertainty. sup means taking the supremum (ie, the worst case) among all possible distributions in the distribution fuzzy set F, E F It represents the expected value under all possible distributions F in the distribution fuzzy set F. The true distribution refers to the actual probability distribution of the random variable. The true distribution can be estimated through historical empirical data. Distribution fuzzy set is a method for dealing with uncertainty, which is used to optimize all possible distribution ranges in the model using fuzzy logic distribution.

[0122] In the second stage, the actual operating power of each device in the virtual power plant can be expressed by an affine adjustment strategy based on the first stage. y(x, ζ%) is the decision variable of the second stage. In one embodiment, the expected adjustment cost is expressed as:

[0123] Q(x, ζ%) = mind T y(x, ζ%)

[0124] st$Gy(x,ζ%)≤h(ζ%)

[0125] Among them, d T is the coefficient vector of the objective function in the second stage, G is the coefficient vector of the constraint conditions in the second stage, and h(ζ%) is the constant vector of the constraint conditions in the second stage.

[0126] It should be noted that the coefficient vector of the second-stage objective function defines the degree of influence of different decision variables (such as adjustment strategy, spare capacity, etc.) on the objective function (such as adjustment cost, penalty for violating constraints, etc.). In the mathematical expression, d Ty(x, ζ%) represents the objective function, where y(x, ζ%) is a function of the decision variables, which depends on the decision variables x and the uncertainty factor ζ% in the first stage. The coefficient vector of the constraints in the second stage defines the relationship between the decision variables and the constraints. These constraints can include resource constraints, technical constraints, safety constraints, etc., which limit the feasible range of the decision variables. The constant vector of the constraints in the second stage represents the right-hand side constant values ​​of the constraints, which can depend on the uncertainty factor ζ%. For example, in the power system, the load demand changes caused by the uncertainty of renewable energy can be represented.

[0127] In the second stage of the optimization problem, the goal can be to find a set of decision variables y(x, ζ%) such that the objective function d is satisfied while satisfying all constraints Gy(x, ζ%) ≤ h(ζ%). T The value of y(x, ζ%) is minimal, so it is usually necessary to solve for all possible values ​​of the uncertainty factor to ensure that the solution found is still valid in the worst case.

[0128] The Earth Mover's Distance (EMD) is used to measure the distance between the empirical distribution and the true distribution. The calculation formula is:

[0129]

[0130] Among them, W(F N , F) is the “distance” between the two distributions, F N is the empirical distribution, historical power data ζ k Follow the empirical distribution F N , the uncertain parameter ζ% obeys the true distribution F, ζ% represents a random variable, that is, it represents a point sampled from the true distribution F, Ξ is the support set, Ξ 2 represents the Cartesian product of two support sets, ||·|| is the 1-norm, and ∏ is and the joint distribution of ζ%, Represents the empirical distribution F N The sampling points in .

[0131] The support set is the set of all possible values ​​in a probability distribution. For a probability distribution, the support set is the area where its probability density function or probability mass function is non-zero).

[0132] The distribution fuzzy set can be represented as a Wasserstein sphere with $F_N$ as the center and ε(N) as the radius:

[0133]

[0134] in, is all possible values ​​of the true distribution F on Ξ, F s is the distribution fuzzy set of the prediction error. In order to control the conservatism, the confidence level β is introduced, and the fuzzy radius can be expressed as:

[0135]

[0136] in, is the error sample mean, which represents the average value of the error between all actual operating power and the corrected predicted power. represents the kth sample point sampled in the empirical distribution, C is a constant, and α is an auxiliary variable used to adjust the weights of different error samples to control the sensitivity of the model to uncertainty, especially to the error term. By introducing the auxiliary variable α, the weights of different error samples can be adjusted more flexibly, especially for samples with large deviations.

[0137] The error uncertainty set of wind power and photovoltaic power has a certain impact on the model results. In order to meet the confidence level and reduce conservatism, it is necessary to shrink it as much as possible while covering the possible probability distribution. Standardization process, forming a mean of 0 and a variance of 1

[0138]

[0139] in, is the mean of the sample set, is the variance of the sample set, represents the kth sample point sampled in the empirical distribution, and Ω is the standardized uncertainty parameter (i.e., uncertainty factor) The support set of can be defined as:

[0140] Based on the above two characteristics (i.e., the difference between the empirical distribution and the true distribution (measured by the Earth Mover's Distance (EMD)); the error fluctuation range and weight adjustment (dynamically optimized by the auxiliary variable α)), an optimization model is constructed to solve σ

[0141]

[0142] Among them, σ max is the maximum value of σ, τ is the confidence level, and F′ is The probability distribution of s for The distribution fuzzy set of σ is the optimization variable, which represents the limit value of an error range or fluctuation range. The optimization goal of the optimization model is to minimize σ, that is, to find an error range as small as possible so that the confidence level constraint is satisfied. R represents a real number set, which is used to explain the random variable. is a real random variable whose value range is all real numbers. Therefore, we can get:

[0143]

[0144] Among them, κ is an auxiliary variable, ε is a constant used to measure the impact of a certain deviation or error in the constraint condition, N is the sample size, which represents the number of samples or observations of the random variable, (·) + Indicates a positive value, i.e. (·) + =max(·,0). (·) + =max(·, 0) ensures that in the worst case, the risk or uncertainty of the optimization model does not exceed 1-τ; κ is used to adjust the metrics in the constraints to ensure the robustness of the optimization model. This optimization problem can restore the support set Ξ of ζ%:

[0145]

[0146] in, is the estimated value of the covariance matrix, which represents the covariance structure of the random variable. Ω is the support set, which represents the credible interval or range of the random variable.

[0147] S103: During actual operation, the corrected predicted power is input into the distribution optimization model, and the optimization is solved by a path optimization algorithm under an uncertain environment to obtain a scheduling strategy.

[0148] Among them, the uncertain environment refers to an environment in which there are multiple possible outcomes in the decision-making process of the distributed optimization model (such as fluctuations in the power generation of renewable energy, changes in load demand, fluctuations in market prices, etc.), but the probability of these outcomes occurring or the specific impact cannot be accurately predicted.

[0149] The path optimization algorithm can be a sparrow search algorithm. In view of the high computational dimension of the robust distributed optimization model of the virtual power plant, which is difficult to solve by traditional methods, we can introduce multiple strategies based on the theory of swarm intelligence optimization algorithm on the basis of the sparrow search algorithm, such as dynamic adjustment of search strategy, memory mechanism, and diversity maintenance strategy. The dynamic adjustment of search strategy is to dynamically adjust the search direction and step size according to the information in the search process to adapt to different search stages; the memory mechanism is to record the excellent solutions in the search process to guide subsequent searches and avoid repeated searches; the diversity maintenance strategy is to maintain the diversity of the population by introducing random perturbations or crossover operations to prevent premature convergence; therefore, it is conducive to improving the accuracy and efficiency of optimization solutions and finally obtaining a scheduling strategy.

[0150] In one embodiment, performing optimization by using the path optimization algorithm in an uncertain environment to obtain a scheduling strategy includes:

[0151] Population initialization based on chaotic mapping;

[0152] Select target individuals from the initialized population;

[0153] A position search is performed in the search space based on the screened target individuals until convergence, thereby obtaining the scheduling strategy.

[0154] Chaotic mapping is a complex nonlinear dynamic system, and its output is highly random and unpredictable. In this application, the population is initialized by chaotic mapping to generate an initialized population; the fitness of each target individual in the initialized population is obtained by calculating the objective function; the target individuals are screened according to the fitness (the scheme with fitness higher than a certain threshold); the sparrow search algorithm is used to search for positions in the search space, and the scheduling scheme is continuously updated until convergence, and the optimal scheduling scheme is selected as the final scheduling strategy.

[0155] The Sine chaotic map is used to initialize the population. The Sine chaotic map is defined as follows:

[0156] Y i+1 =ρsin(πY i )

[0157]

[0158] Where Y i+1 represents the value of the chaotic sequence at the i+1th iteration, and is obtained from Y by the Sine mapping formula i Generate, Y i ∈[-1,1] is the chaotic sequence, ρ is the control parameter, U d ,L d are the upper and lower limits of the sparrow individuals in the dth dimension, Xi,d is the position of the i-th individual in the population in the d-th dimension, Y i,d is the position of the i-th individual in the population on the d-th dimension.

[0159] Sine chaotic mapping has good chaotic characteristics, and its chaotic properties are closely related to the value of parameter ρ. ρ∈[0.87,1] and the closer ρ is to 1, the better the chaotic performance is. i The more evenly distributed in the [-1,1] region. When ρ = 1, the system is in a completely chaotic state, so the subsequent experiments all take ρ = 1. According to the variable values ​​generated by the above formula mapped to the sparrow individuals, the initial solution position of the population can be obtained, that is, the initialized population.

[0160] When searching for the optimal position, the SSA algorithm is prone to deviate from the direction and miss the optimal foraging area, causing the algorithm to fall into a local optimum. In the pigeon flock optimization algorithm, pigeons use magnetic reception to shape a map in their brains to perceive the earth's magnetic field, and use the height of the sun as a compass to adjust the direction, so that the algorithm converges faster and more stably, improves the local development and global exploration capabilities of the algorithm, has the advantages of strong robustness, and makes the generated optimal path smoother and more satisfactory. Therefore, the map compass operator in the pigeon flock optimization algorithm is introduced into the explorer position update process in the sparrow search algorithm. In one embodiment, the position search based on the screening target individual in the search space includes:

[0161] Based on the screened target individuals, a diverse global optimal guidance strategy with inertia weights and a two-sample learning strategy are used to perform position search in the search space.

[0162] The diversity global optimal guidance strategy includes the exploration process and the development process. The exploration process refers to finding unknown areas in the search space. The stronger the exploration ability, the better the global optimization result, but it will slow down the convergence speed. The development process refers to the ability to use the better solution obtained in the early stage to search for a better solution. The stronger the development ability, the better the local optimization effect and the faster the convergence speed.

[0163] In the explorer position update process, a global guidance item is added to improve the global spatial detection capability of the sparrow optimization algorithm. The sparrow individuals in the initial population are mainly guided by the sparrow in the global optimal position (i.e., the target individual), which guides the entire population to find the optimal food source. Therefore, the sparrow individuals in the population can move to the global optimal position without any interference, which speeds up the convergence to the current global optimal position.

[0164] The improved position update formula of the explorer (i.e., the target individual) is:

[0165]

[0166] Among them, R is the map compass operator, between 0 and 1, ω is the inertia weight, t is the current iteration number, Xb d is the position indicated by the map compass in the dth dimension, that is, the current global optimal position. It guides all sparrows toward the optimal solution, enhancing the algorithm's global exploration ability. represents the position of the i-th sparrow individual in the d-th dimension and its value at the t-th iteration, Iter represents the position of the i-th sparrow individual in the d-th dimension and its value at the t+1-th iteration. max is the maximum number of iterations, which is used to normalize the current number of iterations t, thereby controlling the dynamic change of the attenuation factor. is the position of the current global optimal solution in the dth dimension, Q is the random perturbation factor, which is used to introduce randomness to enhance the diversity of the algorithm, usually distributed between 0 and 1. L is the step size factor, which is used to control the step size when the individual position is updated. ST is a threshold used to control the branching logic of the algorithm. R2 is a random number.

[0167] The inertia weight has a huge impact on the performance of the sparrow search algorithm. In the range of [0, 1], the larger the inertia weight, the stronger the global exploration ability and the richer the diversity of the population; the smaller the inertia weight, the stronger the local mining ability of the algorithm and the faster the convergence speed. Therefore, a nonlinear inertia weight formula is proposed as follows:

[0168]

[0169] When α=0.7, β=0.3, the value of ω shows a nonlinear decreasing trend between [0,1] as the number of iterations increases. (α-β) 2 It is a nonlinear exponential term, which is used to control the nonlinear change degree of inertia weight with the number of iterations. It represents the first part of the inertia weight decrease. This item gradually decreases with the increase of the number of iterations t. α is used to control the decreasing rate of the inertia weight. Represents the second part of the decreasing inertia weight. This term also decreases gradually with the increase of the number of iterations t, and β is used to control the decreasing rate of the inertia weight.

[0170] like Figure 3As shown in the figure, in the early stage of iteration, the decay rate of ω changes from fast to slow as the number of iterations increases, which is conducive to global search and approaches the global optimal or better position at a faster speed; in the middle and late stages of iteration, the decay rate of the inertia weight slowly decreases, which is conducive to more refined local mining, improving the overall optimization ability of the algorithm, and to a certain extent accelerating the convergence speed and improving the quality of the optimal solution. As the inertia weight continues to change, the sparrow individuals in the global exploration stage can better find the approximate location of the global optimal solution or the better solution, so the algorithm can find the optimal value faster and easier in the local fine search process. Therefore, the addition of nonlinear inertia weight effectively enhances the population diversity, improves the solution accuracy of the algorithm, and balances the ability of the algorithm in the global exploration and local development periods.

[0171] In the basic sparrow search algorithm, the sparrow's learning is blind. Each time, the follower only selects an explorer sample with a better foraging position than its own to learn, but it cannot judge whether this learning is conducive to finding the best foraging area. This blind learning method is beneficial to the algorithm's exploration of the optimization space, but it will also cause part of the population's important position information to be lost, resulting in weak algorithm development capabilities, slow convergence speed, and easy to fall into local optimality.

[0172] In order to solve these problems, a two-sample learning strategy is introduced to improve the position update formula of the follower in SSA. Figure 4 As shown in the figure, there are two local optimal positions and one global optimal position in the search space. Sparrows m and n have fallen into the area where the local optimal solution is located. Therefore, sparrow i will fall into the local optimal solution whether it learns from m or n.

[0173] In order to allow i to explore the global optimal area k, i is asked to learn from m and n at the same time. The follower is guided by the mixed information of the two explorers. In this way, there is a greater probability of exploring the food source area that has not been searched in the current optimization space, so that the algorithm jumps out of the local optimality and further improves the population's exploration ability of the optimization space. The improved follower position update formula is as follows:

[0174]

[0175] in, represents the position of the i-th sparrow individual in the d-th dimension, and its value at the t+1-th iteration, represents the position of the i-th sparrow individual in the d-th dimension and its value at the t-th iteration. Q is the random perturbation factor, which is used to introduce randomness to enhance the diversity of the algorithm. N is the number of samples, which represents the number of samples or observations of the random variable. is the position of the sparrow individual in the d dimension at the tth iteration, is the position of the current global optimal solution in the dth dimension at the t+1th iteration, usually distributed between 0 and 1. D represents the total number of dimensions of the problem, that is, the number of decision variables (or features), and each dimension corresponds to a decision variable. At the t+1th iteration, the position of the sparrow individual m in the d dimension is usually distributed between 0 and 1. At the t+1th iteration, the position of the sparrow individual n in the d dimension is usually distributed between 0 and 1.

[0176] S104: Perform energy dispatch on the distribution network connected to the virtual power plant according to the dispatch strategy.

[0177] The dispatch strategy is used to dispatch the system equipment in the virtual power plant. For example, the wind farm's power generation plan is adjusted according to the dispatch strategy to maximize the use of wind energy; the photovoltaic system's power generation time is optimized to improve the utilization of solar energy; the battery's charge and discharge plan is determined to balance supply and demand and optimize costs; and the controllable load's power consumption time is adjusted to respond to grid demand and electricity price signals.

[0178] Exemplarily, the scheduling strategy may include wind farms increasing power generation during periods of higher wind speeds (such as early morning and evening); photovoltaic systems maximizing power generation during periods of strongest sunlight (such as noon); battery energy storage systems charging during periods of lower electricity prices and discharging during periods of higher electricity prices; and load control increasing electricity consumption during periods of lower electricity prices and reducing electricity consumption during periods of higher electricity prices.

[0179] The above is the energy scheduling process based on virtual power plant in this application.

[0180] As mentioned above, the present application provides an energy scheduling method, device, computer equipment and storage medium based on a virtual power plant, by obtaining the corrected predicted power of the distributed renewable energy of the virtual power plant; constructing a distribution optimization model of the virtual power plant, the distribution optimization model includes an objective function and constraints, the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power sources; in the actual operation process, the corrected predicted power is input into the distribution optimization model, and the solution is optimized by a path optimization algorithm under an uncertain environment to obtain a scheduling strategy; according to the scheduling strategy, energy scheduling is performed on the distribution network connected to the virtual power plant. In the energy dispatching scheme based on virtual power plant provided in this application, by obtaining the corrected predicted power, the power generation of distributed renewable energy can be predicted more accurately, the prediction error can be reduced, the target work function can effectively reduce the comprehensive operating cost of virtual power plant, improve economic benefits, and the constraint conditions are based on the charging and discharging characteristics of distributed power sources, ensuring that the actual operating limitations of the equipment are fully considered during the dispatching process, avoiding equipment overload or damage, and enhancing the stability and reliability of the system. In an uncertain environment, the distribution optimization model is solved and optimized through the path optimization algorithm, which can flexibly respond to various emergencies, such as weather changes, load fluctuations, etc., to ensure the real-time and adaptability of the dispatching strategy, improve the utilization rate of distributed renewable energy, reduce the phenomenon of wind and light abandonment, and promote the efficient consumption of renewable energy. According to the dispatching strategy, the distribution network connected to the virtual power plant is dispatched for energy, and the coordination and clustering operation of various power resources within the virtual power plant are realized, which can improve the operating efficiency of the distribution network, reduce energy waste, and improve the reliability of the overall energy supply. It can be seen that the efficient dispatching of distributed renewable energy in the virtual power plant is achieved, and the economic benefits and system stability are improved.

[0181] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0182] In one embodiment, an energy dispatching device based on a virtual power plant is provided, and the energy dispatching device based on a virtual power plant corresponds one-to-one to the energy dispatching method based on a virtual power plant in the above embodiment. Figure 5 As shown, the energy dispatching device based on the virtual power plant includes:

[0183] An acquisition module 201 is used to acquire the corrected predicted power of distributed renewable energy of a virtual power plant;

[0184] A construction module 202 is used to construct a distribution optimization model of the virtual power plant, wherein the distribution optimization model includes an objective function and constraints, wherein the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power source;

[0185] The optimization module 203 is used to input the corrected predicted power into the distribution optimization model during the actual operation, and solve and optimize the model through the path optimization algorithm under the uncertainty environment to obtain the scheduling strategy;

[0186] The scheduling module 204 is used to perform energy scheduling on the distribution network connected to the virtual power plant according to the scheduling strategy.

[0187] In the energy dispatching scheme based on virtual power plant provided in the present application, in this embodiment, by obtaining the corrected predicted power, the power generation of distributed renewable energy can be predicted more accurately, the prediction error can be reduced, the target work function can effectively reduce the comprehensive operating cost of the virtual power plant, and improve the economic benefits. The constraint conditions are based on the charging and discharging characteristics of the distributed power source, ensuring that the actual operating limitations of the equipment are fully considered during the dispatching process, avoiding equipment overload or damage, and enhancing the stability and reliability of the system. In an uncertain environment, the distribution optimization model is solved and optimized by the path optimization algorithm, which can flexibly respond to various emergencies, such as weather changes, load fluctuations, etc., to ensure the real-time and adaptability of the dispatching strategy, improve the utilization rate of distributed renewable energy, reduce the phenomenon of wind and light abandonment, and promote the efficient consumption of renewable energy. According to the dispatching strategy, the distribution network connected to the virtual power plant is dispatched for energy, and the coordination and clustering operation of various power resources within the virtual power plant are realized, which can improve the operating efficiency of the distribution network, reduce energy waste, and improve the reliability of the overall energy supply. It can be seen that the efficient dispatching of distributed renewable energy in the virtual power plant is achieved, and the economic benefits and system stability are improved.

[0188] Optionally, in some embodiments of the present application, the acquisition module includes:

[0189] The first acquisition submodule is used to obtain the initial predicted power and average prediction error of the distributed renewable energy of the virtual power plant;

[0190] The obtaining submodule is used to obtain the corrected predicted power based on the initial predicted power and the average prediction error.

[0191] Optionally, in some embodiments of the present application, the building blocks include:

[0192] The modeling submodule is used to optimize the virtual power plant scheduling by using the distributed robust optimization model to obtain the distributed optimization model, wherein the distributed optimization model includes two stages. In the first stage, the planned power is determined according to the corrected predicted power; in the second stage, the actual generated power and the initial predicted power or the deviation of the corrected predicted power are balanced according to the flexibility resources of the virtual power plant.

[0193] Optionally, in some embodiments of the present application, the distribution optimization model is expressed as follows:

[0194] min x {c T x+supE F [Q(x, ζ%)]}

[0195] stAx≤b

[0196] Among them, x is the decision variable in the first stage, c T is the coefficient vector of the objective function in the first stage, A is the coefficient vector of the constraint conditions in the first stage, b is the constant vector of the constraint conditions in the first stage, is the objective function of the second stage, which represents the expected adjustment cost when the random variable ζ% obeys the worst probability distribution; F is the true distribution, and F is the distribution fuzzy set.

[0197] Optionally, in some embodiments of the present application, the expected adjustment cost is expressed as:

[0198] Q(x, ζ%) = mind T y(x, ζ%)

[0199] st$Gy(x,ζ%)≤h(ζ%)

[0200] Among them, d T is the coefficient vector of the objective function in the second stage, G is the coefficient vector of the constraint conditions in the second stage, and h(ζ%) is the constant vector of the constraint conditions in the second stage.

[0201] Optionally, in some embodiments of the present application, the optimization module includes:

[0202] Initialization submodule, used to initialize the population based on chaotic mapping;

[0203] The screening submodule is used to screen target individuals from the initialized population;

[0204] The search submodule is used to perform position search in the search space based on the screened target individuals until convergence to obtain the scheduling strategy.

[0205] Optionally, in some embodiments of the present application, the search submodule includes:

[0206] The search unit is used to perform position search in the search space based on the screened target individuals by adopting a diverse global optimal guidance strategy with inertia weights and a double sample learning strategy.

[0207] In one embodiment, a computer device is provided, the internal structure diagram of which can be as follows: Figure 6 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server through a network connection. When the computer program is executed by the processor, it realizes the functions or steps of an energy scheduling method based on a virtual power plant.

[0208] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:

[0209] Obtain the corrected predicted power of distributed renewable energy of the virtual power plant; construct a distribution optimization model of the virtual power plant, which includes an objective function and constraints. The objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power sources; in the actual operation process, input the corrected predicted power into the distribution optimization model, and solve and optimize it through the path optimization algorithm under an uncertain environment to obtain a scheduling strategy; perform energy scheduling on the distribution network connected to the virtual power plant according to the scheduling strategy.

[0210] In this embodiment, by obtaining the corrected predicted power, the power generation of distributed renewable energy can be predicted more accurately, the prediction error can be reduced, the target work function can effectively reduce the comprehensive operating cost of the virtual power plant, and improve the economic benefits. The constraint conditions are based on the charging and discharging characteristics of the distributed power source, ensuring that the actual operating limitations of the equipment are fully considered during the scheduling process, avoiding equipment overload or damage, and enhancing the stability and reliability of the system. In an uncertain environment, the distribution optimization model is solved and optimized through the path optimization algorithm, which can flexibly respond to various emergencies, such as weather changes, load fluctuations, etc., to ensure the real-time and adaptability of the scheduling strategy, improve the utilization rate of distributed renewable energy, reduce the phenomenon of wind and light abandonment, and promote the efficient consumption of renewable energy. According to the scheduling strategy, the distribution network connected to the virtual power plant is dispatched for energy, and the coordination and clustered operation of various power resources within the virtual power plant are realized, which can improve the operating efficiency of the distribution network, reduce energy waste, and improve the reliability of the overall energy supply. It can be seen that the efficient scheduling of distributed renewable energy in the virtual power plant is achieved, and the economic benefits and system stability are improved.

[0211] In one embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0212] Obtain the corrected predicted power of distributed renewable energy of the virtual power plant; construct a distribution optimization model of the virtual power plant, which includes an objective function and constraints. The objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power sources; in the actual operation process, input the corrected predicted power into the distribution optimization model, and solve and optimize it through the path optimization algorithm under an uncertain environment to obtain a scheduling strategy; perform energy scheduling on the distribution network connected to the virtual power plant according to the scheduling strategy.

[0213] In this embodiment, by obtaining the corrected predicted power, the power generation of distributed renewable energy can be predicted more accurately, the prediction error can be reduced, the target work function can effectively reduce the comprehensive operating cost of the virtual power plant, and improve the economic benefits. The constraint conditions are based on the charging and discharging characteristics of the distributed power source, ensuring that the actual operating limitations of the equipment are fully considered during the scheduling process, avoiding equipment overload or damage, and enhancing the stability and reliability of the system. In an uncertain environment, the distribution optimization model is solved and optimized through the path optimization algorithm, which can flexibly respond to various emergencies, such as weather changes, load fluctuations, etc., to ensure the real-time and adaptability of the scheduling strategy, improve the utilization rate of distributed renewable energy, reduce the phenomenon of wind and light abandonment, and promote the efficient consumption of renewable energy. According to the scheduling strategy, the distribution network connected to the virtual power plant is dispatched for energy, and the coordination and clustered operation of various power resources within the virtual power plant are realized, which can improve the operating efficiency of the distribution network, reduce energy waste, and improve the reliability of the overall energy supply. It can be seen that the efficient scheduling of distributed renewable energy in the virtual power plant is achieved, and the economic benefits and system stability are improved.

[0214] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can refer to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0215] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, library or other media used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0216] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0217] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. An energy scheduling method based on a virtual power plant, characterized in that: The method comprises: Obtaining the corrected forecast power of distributed renewable energy sources of the virtual power plant; Constructing a distribution optimization model of the virtual power plant, the distribution optimization model includes an objective function and constraints, the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power source; In the actual operation process, the corrected predicted power is input into the distribution optimization model, and the optimization is solved by the path optimization algorithm under the uncertainty environment to obtain the scheduling strategy; Energy dispatch is performed on the distribution network connected to the virtual power plant according to the dispatch strategy.

2. The energy dispatching method based on virtual power plant according to claim 1 is characterized in that: The obtaining of the corrected predicted power of the distributed renewable energy of the virtual power plant comprises: Obtain the initial predicted power and average prediction error of distributed renewable energy of the virtual power plant; The corrected prediction power is obtained based on the initial prediction power and the average prediction error.

3. The energy dispatching method based on virtual power plant according to claim 1 is characterized in that: The construction of the distribution optimization model of the virtual power plant includes: The virtual power plant scheduling is optimized and modeled using a distributed robust optimization model to obtain the distributed optimization model, wherein the distributed optimization model includes two stages. In the first stage, the planned power is determined based on the corrected predicted power; in the second stage, the actual generated power and the initial predicted power or the deviation of the corrected predicted power are balanced based on the flexibility resources of the virtual power plant.

4. The energy dispatching method based on virtual power plant according to claim 3 is characterized in that: The distribution optimization model is expressed as follows: minutes x {c} T x+supE F [Q(x,ζ%)]} stAx≤b Among them, x is the decision variable in the first stage, c T is the coefficient vector of the objective function in the first stage, A is the coefficient vector of the constraint conditions in the first stage, b is the constant vector of the constraint conditions in the first stage, is the objective function of the second stage, which represents the expected adjustment cost when the random variable ζ% obeys the worst probability distribution; F is the true distribution, and F is the distribution fuzzy set.

5. The energy dispatching method based on virtual power plant according to claim 4 is characterized in that: The expected adjustment cost is expressed as: Q(x,ζ%)=mind T y(x,ζ%) st$Gy(x,ζ%)≤h(ζ%) Among them, d T is the coefficient vector of the objective function in the second stage, G is the coefficient vector of the constraint conditions in the second stage, and h(ζ%) is the constant vector of the constraint conditions in the second stage.

6. The energy dispatching method based on virtual power plant according to claim 1 is characterized in that: The optimization is solved by the path optimization algorithm in an uncertain environment to obtain a scheduling strategy including: Population initialization based on chaotic mapping; Select target individuals from the initialized population; A position search is performed in the search space based on the screened target individuals until convergence, thereby obtaining the scheduling strategy.

7. The energy dispatching method based on virtual power plant according to claim 6 is characterized in that: The position search of the target individual based on the screening in the search space includes: Based on the screened target individuals, a diverse global optimal guidance strategy with inertia weights and a two-sample learning strategy are used to perform position search in the search space.

8. An energy dispatching device based on a virtual power plant, characterized in that: include: An acquisition module, used for acquiring the corrected predicted power of distributed renewable energy of the virtual power plant; A construction module is used to construct a distribution optimization model of the virtual power plant, wherein the distribution optimization model includes an objective function and constraints, wherein the objective function takes the response cost, energy storage cost, regulation cost and carbon trading cost of the virtual power plant as targets, and the constraints are the charging and discharging characteristics of the distributed power source; An optimization module, used for inputting the corrected predicted power into the distribution optimization model during actual operation, and performing optimization through a path optimization algorithm under an uncertain environment to obtain a scheduling strategy; A scheduling module is used to schedule energy for the distribution network connected to the virtual power plant according to the scheduling strategy.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the steps of the energy scheduling method based on the virtual power plant as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the energy scheduling method based on a virtual power plant as described in any one of claims 1 to 7 are implemented.

Images