Virtual power plant non-cooperative game scheduling method and device based on generalized energy storage

Abstract

The invention discloses a virtual power plant non-cooperative game scheduling method and device based on generalized energy storage, relates to the technical field of power scheduling, and mainly aims to solve the problem that the power utilization comfort of a user is neglected in an existing power scheduling scheme making process. Carrying out screening processing on the environmental parameters to obtain environmental influence factors meeting correlation requirements; performing prediction processing on the new energy output based on the environmental influence factors to obtain a prediction result of the new energy output; establishing a generalized energy storage model, and establishing a virtual gamer non-cooperative game model based on the generalized energy storage model and the prediction result of the new energy output; and carrying out optimization solution on the virtual gamer non-cooperative game model to obtain a scheduling result of the virtual power plant.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power dispatching, and in particular to a non-cooperative game dispatching method and device for a virtual power plant based on generalized energy storage. Background Art

[0002] In recent years, various regions have vigorously developed new energy industries such as wind power and photovoltaic power generation. However, due to the small capacity of single power units, scattered geographical distribution, and greater randomness and volatility in the grid connection process, the large-scale access of wind and photovoltaic power generation has brought many challenges to the reliable operation of the power grid. As a distributed energy coordination management technology, virtual power plants use multiple technologies such as communication and control to integrate multi-regional, large-scale, independent distributed energy, energy storage, and loads into system scheduling, which can promote the consumption of new energy in the power grid while obtaining economic benefits.

[0003] However, when virtual power plants coordinate and dispatch distributed energy, they usually only take economy as a single goal, ignoring the user's power comfort after actual dispatch, which directly reduces the enthusiasm of various entities in virtual power plants to participate in demand response. Therefore, there is an urgent need for a distributed energy coordination and dispatching solution for virtual power plants that takes into account both economy and the optimal acceptance range of user comfort, so as to solve the problem of significantly reduced user comfort while improving economy, and realize flexible coordination between energy economy of virtual power plants and user satisfaction. Summary of the invention

[0004] In view of this, the present invention provides a non-cooperative game scheduling method and device for a virtual power plant based on generalized energy storage, the main purpose of which is to solve the problem of ignoring the user's electricity comfort in the process of formulating existing power scheduling plans.

[0005] According to one aspect of the present invention, a non-cooperative game scheduling method for a virtual power plant based on generalized energy storage is provided, comprising: Collect environmental parameters that affect the output of new energy, and screen and process the environmental parameters to obtain environmental influencing factors that meet the relevance requirements; Performing prediction processing on the output of the new energy source based on the environmental influencing factors to obtain a prediction result of the output of the new energy source; Establishing a generalized energy storage model, and establishing a virtual player non-cooperative game model based on the generalized energy storage model and the prediction results of the new energy output; The non-cooperative game model of the virtual players is optimized and solved to obtain the dispatch result of the virtual power plant.

[0006] Furthermore, the environmental parameters are screened to obtain environmental influencing factors that meet the relevance requirements, including: Setting each of the environmental parameters as node construction graph structure data; Using a multivariate correlation analysis method to calculate the correlation of each of the environmental parameters in the graph structure data, and obtaining a correlation value corresponding to each of the environmental parameters; Based on the correlation values, environmental parameters that do not meet the correlation requirements are eliminated to obtain the environmental influencing factors that meet the correlation requirements.

[0007] Furthermore, the prediction result includes a day-ahead prediction result and an intraday prediction result. The prediction processing of the new energy output based on the environmental influencing factors to obtain the prediction result of the new energy output includes: The gated recurrent unit is used to improve the graph neural network to obtain an improved graph neural network model; Based on the improved graph neural network model and the environmental influencing factors, the new energy output is predicted and processed in the day-ahead stage to obtain the day-ahead prediction result; The intraday stage is divided into multiple prediction cycles, and the new energy output of the current prediction cycle is predicted based on the actual output of new energy in the previous prediction cycle, the improved graph neural network model and the environmental influencing factors until all prediction cycles are processed, and the intraday prediction results corresponding to each prediction cycle are obtained.

[0008] Furthermore, the generalized energy storage model includes an electric vehicle generalized energy storage model, an air conditioning cluster generalized energy storage model, and an air conditioning cluster generalized energy storage model. The establishment of the generalized energy storage model includes: Establishing a generalized energy storage model of the electric vehicle according to the state characteristics of the electric vehicle; Establishing a generalized energy storage model of the building according to the indoor temperature of the building and the set upper and lower limits of the allowable heating temperature; The generalized energy storage model of the air-conditioning cluster is established according to the upper and lower limits of the temperature set by the air-conditioning.

[0009] Furthermore, the generalized energy storage model of the electric vehicle is established according to the state characteristics of the electric vehicle, including: Sampling each electric vehicle to obtain state characteristics of each electric vehicle; and classifying each electric vehicle based on the state characteristics to obtain a vehicle classification result; Based on the automobile classification results and the state characteristics of each type of automobile, determine the probability density function of automobile off-grid and on-grid and the probability density function of automobile driving distance; The dispatchable power of the vehicle at each moment in the dispatch cycle is determined according to the probability density function of the vehicle being off-grid and on-grid and the probability density function of the vehicle's travel distance.

[0010] Furthermore, the generalized energy storage model of the building is established according to the indoor temperature of the building and the set upper and lower limits of the allowable heating temperature, including: Determine the heat load of the building based on the indoor temperature of the building, the equivalent heat capacity coefficient, the equivalent heat loss coefficient, and the ambient temperature around the building; The heat storage state of the building at each moment in the scheduling cycle is determined according to the set upper and lower limits of the allowable heating temperature and the heat load.

[0011] Furthermore, the virtual player non-cooperative game model is established based on the generalized energy storage model and the prediction result of the new energy output, including: Construct an economic objective function based on the operation and maintenance costs and energy purchase costs of the virtual power plant; Based on the generalized energy storage model, the start and stop states of the air conditioner and the electric vehicle at each time in the scheduling cycle are determined, and the comfort objective function is constructed in combination with the thermal comfort index and the time comfort; The economic objective function and the comfort objective function are determined as the profit functions of the virtual game players, and the non-cooperative game model of the virtual game players is optimally established by solving the minimum value of the profit function.

[0012] According to another aspect of the present invention, a non-cooperative game scheduling device for a virtual power plant based on generalized energy storage is provided, comprising: A factor screening module is used to collect environmental parameters that affect the output of new energy sources, and screen the environmental parameters to obtain environmental influencing factors that meet the relevance requirements; A prediction module, used for predicting the output of the new energy source based on the environmental influencing factors to obtain a prediction result of the output of the new energy source; A model building module, used to build a generalized energy storage model, and to build a virtual player non-cooperative game model based on the generalized energy storage model and the prediction result of the new energy output; The solution module is used to optimize and solve the non-cooperative game model of the virtual players to obtain the scheduling result of the virtual power plant.

[0013] Furthermore, the factor screening module is also used for: Setting each of the environmental parameters as node construction graph structure data; Using a multivariate correlation analysis method to calculate the correlation of each of the environmental parameters in the graph structure data, and obtaining a correlation value corresponding to each of the environmental parameters; Based on the correlation values, environmental parameters that do not meet the correlation requirements are eliminated to obtain the environmental influencing factors that meet the correlation requirements.

[0014] Furthermore, the prediction results include day-ahead prediction results and intraday prediction results, and the prediction module is further used for: The gated recurrent unit is used to improve the graph neural network to obtain an improved graph neural network model; Based on the improved graph neural network model and the environmental influencing factors, the new energy output is predicted and processed in the day-ahead stage to obtain the day-ahead prediction result; The intraday stage is divided into multiple prediction cycles, and the new energy output of the current prediction cycle is predicted based on the actual output of new energy in the previous prediction cycle, the improved graph neural network model and the environmental influencing factors until all prediction cycles are processed, and the intraday prediction results corresponding to each prediction cycle are obtained.

[0015] Further, the generalized energy storage model includes an electric vehicle generalized energy storage model, an air conditioning cluster generalized energy storage model, and an air conditioning cluster generalized energy storage model, and the model building module includes a first building unit, a second building unit, and a third building unit; The first establishing unit is used to establish the generalized energy storage model of the electric vehicle according to the state characteristics of the electric vehicle; The second establishing unit is used to establish the generalized energy storage model of the building according to the indoor temperature of the building and the set upper and lower limits of the allowable heating temperature; The third establishing unit is used to establish the generalized energy storage model of the air-conditioning cluster according to the upper and lower limits of the temperature set by the air-conditioning.

[0016] Furthermore, the first establishing unit is further used for: Sampling each electric vehicle to obtain state characteristics of each electric vehicle; and classifying each electric vehicle based on the state characteristics to obtain a vehicle classification result; Based on the automobile classification results and the state characteristics of each type of automobile, determine the probability density function of automobile off-grid and on-grid and the probability density function of automobile driving distance; The dispatchable power of the vehicle at each moment in the dispatch cycle is determined according to the probability density function of the vehicle being off-grid and on-grid and the probability density function of the vehicle's travel distance.

[0017] Furthermore, the second establishing unit is further used for: Determine the heat load of the building based on the indoor temperature of the building, the equivalent heat capacity coefficient, the equivalent heat loss coefficient, and the ambient temperature around the building; The heat storage state of the building at each moment in the scheduling cycle is determined according to the set upper and lower limits of the allowable heating temperature and the heat load.

[0018] Furthermore, the model building module further includes a fourth building unit, which is used to: Construct an economic objective function based on the operation and maintenance costs and energy purchase costs of the virtual power plant; Based on the generalized energy storage model, the start and stop states of the air conditioner and the electric vehicle at each time in the scheduling cycle are determined, and the comfort objective function is constructed in combination with the thermal comfort index and the time comfort; The economic objective function and the comfort objective function are determined as the profit functions of the virtual game players, and the non-cooperative game model of the virtual game players is optimally established by solving the minimum value of the profit function.

[0019] According to another aspect of the present invention, a storage medium is provided, wherein at least one executable instruction is stored in the storage medium, and the executable instruction enables a processor to perform operations corresponding to the above-mentioned non-cooperative game scheduling method for a virtual power plant based on generalized energy storage.

[0020] According to another aspect of the present invention, there is provided a device, comprising a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the above-mentioned non-cooperative game scheduling method for virtual power plants based on generalized energy storage.

[0021] By means of the above technical solution, the technical solution provided by the embodiment of the present invention has at least the following advantages: The present invention provides a non-cooperative game scheduling method and device for a virtual power plant based on generalized energy storage. Compared with the prior art, the present invention collects environmental parameters that affect the output of new energy, and screens and processes the environmental parameters to obtain environmental influencing factors that meet the correlation requirements; predicts and processes the output of new energy based on the environmental influencing factors to obtain the predicted results of the output of new energy; establishes a generalized energy storage model, and establishes a virtual player non-cooperative game model based on the generalized energy storage model and the predicted results of the output of new energy; optimizes and solves the virtual player non-cooperative game model to obtain the scheduling results of the virtual power plant, and completes the power scheduling based on generalized energy storage. The present invention comprehensively considers the economy, user thermal comfort, and time comfort, and proposes a virtual player non-cooperative game model with economy and multiple comfort as the goals, so that the virtual power plant can consider the economy and comfort of each subject in the virtual power plant while supplying energy to the regional power load, maintain a stable, continuous and predictable power supply state and power supply efficiency, and maintain power supply safety.

[0022] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings: Figure 1 A schematic flow chart of a non-cooperative game scheduling method for a virtual power plant based on generalized energy storage provided by an embodiment of the present invention is shown; Figure 2 A flow chart of another non-cooperative game scheduling method for a virtual power plant based on generalized energy storage provided by an embodiment of the present invention is shown; Figure 3 A flow chart of another non-cooperative game scheduling method for a virtual power plant based on generalized energy storage provided by an embodiment of the present invention is shown; Figure 4 A schematic flow chart of another non-cooperative game scheduling method for a virtual power plant based on generalized energy storage provided by an embodiment of the present invention is shown; Figure 5 A structural schematic diagram of a non-cooperative game scheduling device for a virtual power plant based on generalized energy storage provided by an embodiment of the present invention is shown; Figure 6 A schematic structural diagram of a device provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0024] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0025] In the embodiment of the present invention, a virtual power plant (VPP) is a management system that integrates distributed energy resources. It integrates multiple distributed energy sources (such as wind energy, solar energy, biomass energy, small hydropower stations, energy storage equipment, etc.) and adjustable load resources through information technology and Internet technology to achieve coordinated management of power production, storage and consumption. Virtual power plants can not only balance supply and demand and improve energy efficiency, but also participate in power market transactions and provide auxiliary services such as peak load regulation and frequency regulation.

[0026] With the development and in-depth research on the load side of virtual power plants, in addition to physical energy storage devices such as batteries and heat storage tanks, generalized energy storage (including pipeline storage or time and space transfer of electric energy) can also be used to balance power fluctuations at different times within virtual power plants. Although no actual energy storage equipment is built, the actual effect is similar to that of physical energy storage equipment.

[0027] The embodiment of the present invention provides a non-cooperative game scheduling method for a virtual power plant based on generalized energy storage, such as Figure 1 As shown, the method includes: 101. Collect environmental parameters that affect the output of new energy sources, and screen and process the environmental parameters to obtain environmental impact factors that meet the relevance requirements; In the embodiment of the present invention, the current execution end collects environmental parameters that affect the output of new energy, including direct radiation, scattered radiation, total horizontal radiation, clear sky direct radiation, clear sky scattered radiation, clear sky total horizontal radiation, dew point, cloud type, air pressure, solar zenith angle, wind speed, wind direction, relative humidity, temperature, precipitation, etc. that affect wind power and photovoltaic output, which are not specifically limited in the embodiment of the present invention. Since the above environmental parameters have different degrees of influence on the output of new energy, the current execution end needs to screen and process the above environmental parameters, that is, to screen out environmental influencing factors that meet the correlation requirements from the above more environmental parameters.

[0028] 102. Predicting the output of the new energy source based on the environmental influencing factors to obtain a prediction result of the output of the new energy source; In the embodiment of the present invention, the current execution end performs prediction processing on the output of new energy based on the environmental influencing factors screened in step 101 to obtain the prediction result of the output of new energy. The prediction processing can be implemented by using a prediction model, which represents a type of model for the quantitative relationship between things described by mathematical language or formula for prediction. It reveals the inherent regularity between things to a certain extent, and it is used as a direct basis for calculating the prediction value during prediction.

[0029] 103. Establish a generalized energy storage model, and establish a virtual player non-cooperative game model based on the generalized energy storage model and the prediction result of the new energy output; In an embodiment of the present invention, the current execution end establishes a generalized energy storage model, wherein the generalized energy storage model includes a generalized energy storage model for electric vehicles, a generalized energy storage model for buildings, a generalized energy storage model for air conditioning clusters, etc., which are not specifically limited in the embodiment of the present invention. The above-mentioned generalized energy storage model simulates the virtual charging and discharging power, virtual capacity, virtual state of charge, etc. of subjects such as electric vehicles, buildings, and air conditioning clusters, and realizes generalized energy storage under the quantification of power storage and heat storage. Constructing a generalized energy storage model can reduce the installation of conventional energy storage equipment and reduce the system operation and maintenance costs. While changing the spatiotemporal distribution of the original electric energy of the power plant, it further reduces the output deviation of the virtual power plant and improves the power supply reliability. In addition, the current execution end also establishes a virtual player non-cooperative game model based on the constructed generalized energy storage model and the prediction results of the new energy output obtained in step 102. Among them, the virtual player non-cooperative game model belongs to a multi-agent non-cooperative game model. The multi-agent non-cooperative game refers to a strategy environment in which there is no cooperation framework between multiple participants, but each other's actions are regarded as individual actions. In this game, each subject mainly emphasizes making independent decisions, regardless of the decisions of other subjects. Non-cooperative games emphasize decision-making behavior in the absence of cooperation or inability to reach a cooperative agreement. In this game, the gain of one party often means the loss of the other party, and the overall benefit remains unchanged or decreases. Typical scenarios of multi-agent non-cooperative games include market competition, in which participants usually pursue the maximization of their own interests without considering the interests of other participants.

[0030] It should be noted that the non-cooperative game model of virtual players in the embodiment of the present invention comprehensively considers the operating economy of the virtual power plant and the user's comfort requirements, which can avoid the subjectivity brought about by empowerment and effectively protect the user's energy rights and interests.

[0031] 104. Optimize and solve the non-cooperative game model of the virtual players to obtain the dispatch result of the virtual power plant.

[0032] In an embodiment of the present invention, the current execution end optimizes and solves the non-cooperative game model of the virtual players to obtain the scheduling result of the virtual power plant. In the optimization and solving process, when the solution reaches the Nash equilibrium, also known as the non-cooperative game equilibrium, the final scheduling result is obtained. Non-cooperative game equilibrium is a core concept in game theory. It refers to a situation in which all participants in a strategy combination are faced with such a situation: when the strategies of other participants remain unchanged, it is the best choice for the participant to adopt the current strategy. In other words, no participant can increase the profit by unilaterally changing the strategy, because the current strategy of each participant is the best response to the strategies of other participants.

[0033] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to reduce the dimension of multi-dimensional environmental parameters and obtain environmental influencing factors with high correlation with the output of new energy, another non-cooperative game scheduling method for virtual power plants based on generalized energy storage is provided, such as Figure 2 As shown, the step of screening the environmental parameters to obtain environmental influencing factors that meet the relevance requirements includes: 201. Set each of the environmental parameters as node construction graph structure data; In the embodiment of the present invention, the current execution end defines the graph structure as , set each environmental parameter as the node construction graph structure data, and use To indicate that Represents N environmental parameters. F represents the edge set of the graph, A Represented as the adjacency matrix of the graph.

[0034] 202. Calculate the correlation of each of the environmental parameters in the graph structure data using a multivariate correlation analysis method to obtain a correlation value corresponding to each of the environmental parameters; In the embodiment of the present invention, the current execution end uses a multivariate correlation analysis method to calculate the correlation of each of the environmental parameters in the graph structure data. For example, the maximum information coefficient MIC is used to calculate the correlation of each environmental parameter in the graph structure network. The specific formula is as follows:

[0035]

[0036] Where a and b are the number of grids divided by the electrical and thermal loads in the x and y directions, respectively; is the probability of the variable sample being distributed in the grid (x, y) coordinates, that is, the joint probability density distribution function of variables x and y; p ( x ), p ( y ) are the marginal probability density distribution functions of x and y respectively; B is the 0.55 or 0.6 power of the data volume, To calculate the mutual information, is the maximum value of the normalized mutual information, that is, the calculated MIC value.

[0037] 203. Eliminate environmental parameters that do not meet the correlation requirement based on the correlation value to obtain the environmental influencing factors that meet the correlation requirement.

[0038] In the embodiment of the present invention, the current execution end establishes a MIC value correlation judgment condition. For example, if the MIC values ​​are all above the first threshold value (for example, 0.9), it can be inferred that the corresponding environmental parameters are the main factors affecting photovoltaic output. If they are between the second threshold value and the first threshold value (for example, 0.5-0.9), it means that the correlation is strong. If they are below the second threshold value (for example, 0.5), it means that the correlation is weak. The environmental parameters with weak correlation are eliminated, and the remaining environmental parameters are the environmental influencing factors that meet the correlation requirements.

[0039] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to improve the accuracy of the new energy output prediction, another non-cooperative game scheduling method for virtual power plants based on generalized energy storage is provided, such as Figure 3 As shown, the prediction results in the method include day-ahead prediction results and intraday prediction results. The step predicts the output of the new energy source based on the environmental influencing factors, and the prediction results of the output of the new energy source include: 301. Using a gated recurrent unit to improve the graph neural network, and obtaining an improved graph neural network model; In the embodiment of the present invention, the current execution end uses a gated recurrent unit (GRU) to improve the graph convolutional neural network (GCN) to obtain an improved graph neural network model; wherein the graph neural network GCN is used to extract the spatial information of each node and edge in the graph, and the time correlation information of each graph spatial information is extracted by using the gated recurrent unit GRU. The gated recurrent unit GRU in the improved graph neural network model has a simpler structure and fast operation speed, which can well solve the gradient explosion and vanishing problems generated by the graph neural network when processing very long time series.

[0040] 302. Perform a day-ahead forecast process on the new energy output based on the improved graph neural network model and the environmental influencing factors to obtain a day-ahead forecast result; In the embodiment of the present invention, before the current execution end uses the improved graph neural network model for prediction processing, it first constructs the wind power, photovoltaic power station and the transmission line connection relationship between them into a graph structure, in which the node represents the power station, the edge represents the connection relationship between them, and a single pixel in the image is regarded as a node, and then the installed capacity, wind speed, light intensity and other environmental influencing factors and other features are used as node attributes. Then, the graph structure data is input into the improved graph neural network model for prediction processing to obtain the day-ahead prediction result.

[0041] 303. Divide the intraday stage into multiple forecast cycles, and forecast the new energy output of the current forecast cycle based on the actual output of new energy in the previous forecast cycle, the improved graph neural network model and the environmental influencing factors until all forecast cycles are processed, and obtain the intraday forecast results corresponding to each forecast cycle.

[0042] In the embodiment of the present invention, considering that wind power and photovoltaic output are affected by environmental factors such as light intensity, temperature, and humidity, there are large prediction errors, and the output fluctuations of various devices in the distribution network are large, the prediction errors of wind power and photovoltaic output can be effectively reduced by shortening the prediction time. The current execution end divides the intraday stage into multiple prediction cycles. For example, if a day is divided into 24 prediction cycles, a 1h rolling prediction within the day is adopted. The initial time of the intraday prediction is set to t=1. After the previous prediction cycle ends, the actual wind power and photovoltaic output at each moment in the previous prediction cycle is obtained at the beginning of the current prediction cycle, and the latest environmental parameters are input. The improved graph neural network model is used to predict the wind power and photovoltaic output of the current prediction cycle. The specific formula is as follows:

[0043]

[0044] in, i is the scheduling time step; The wind turbine is t Moment prediction t+i Always make contributions, The photovoltaic unit is t Moment prediction t+i Always make contributions; Wind power output is t Initial value at time, For photovoltaic power generation t Initial value of the moment; It is a prediction t+i The increase in wind power output at any given moment, For prediction t+i The increment of photovoltaic output at the moment. Until all forecast periods are processed, the intraday forecast results corresponding to each forecast period are obtained.

[0045] Furthermore, as a refinement and extension of the specific implementation of the above embodiment, in order to quantify the generalized energy storage for effective scheduling and control of the virtual power plant, another non-cooperative game scheduling method for the virtual power plant based on generalized energy storage is provided, such as Figure 4 As shown, the steps to establish a generalized energy storage model include: 401. Establish a generalized energy storage model for electric vehicles based on the state characteristics of electric vehicles; In the embodiment of the present invention, the current execution end performs sampling processing on each electric vehicle to obtain the state characteristics of each electric vehicle; for example, Latin hypercube sampling is used to obtain the parameters of each electric vehicle: network access and network disconnection time , , number of trips Etc., which is not specifically limited in the embodiments of the present invention. Then, the current execution end classifies each of the electric vehicles based on the state features to obtain a vehicle classification result; the specific classification process can be implemented using a spectral clustering algorithm. The spectral clustering algorithm can improve the similarity between classified samples. Its essence is to achieve scene clustering through cutting graphs, and transform the clustering problem into a graph optimal partitioning problem based on the matrix eigenvectors and the weights between each vertex, thereby achieving scene reduction.

[0046] In the embodiment of the present invention, the current execution end uses the data set to collect all the electric vehicle parameters To represent, n is the total number of electric vehicles sampled. According to the following formula, the Euclidean distance between data points and the Gaussian distance between data points are calculated respectively to form a similarity matrix K.

[0047]

[0048]

[0049] in, is the Euclidean distance between sample points; , They represent different electric vehicle individuals; It is the scale parameter that determines the decay rate between sample points; is the Gaussian distance between sample points.

[0050] In the embodiment of the present invention, the current execution end uses the similarity matrix Constructing the Laplacian Matrix , further construct the regularized Laplace matrix .calculate The eigenvectors corresponding to the first i largest eigenvalues ​​of , use it as each column to construct a matrix, convert the row vector of the matrix into a unit vector, and get the matrix , each row of the matrix is ​​regarded as a point, and a clustering result is obtained, that is, electric vehicles with similar off-grid time and similar SOC are clustered into one category, and a vehicle classification result is obtained. In the embodiment of the present invention, the behavior habits of various types of electric vehicles and the proportion of various types of electric vehicles in all electric vehicles can be understood through the classification results, and the embodiment of the present invention does not make specific limitations.

[0051] In the embodiment of the present invention, the current execution end determines the probability density function of the vehicle off-grid and on-grid and the probability density function of the vehicle driving distance based on the vehicle classification result and the state characteristics of each type of vehicle; the off-grid probability density function of the electric vehicle f dep And the grid connection probability density function f arr As shown below:

[0052]

[0053] in, , are the expected values ​​of off-grid time and grid-connected time, respectively; , are the standard deviation of off-grid time and number of trips, respectively.

[0054] Establish the relationship model between electric vehicle driving distance and starting SOC:

[0055]

[0056] in, is the probability density function of daily travel, is the driving distance; is the expected value of the number of trips, which is 3.98; is the standard deviation of the number of trips, is the power consumption per kilometer of electric vehicles, The capacity of a single electric vehicle.

[0057] In the embodiment of the present invention, the current execution end determines the dispatchable amount of the vehicle at each moment in the dispatch cycle according to the probability density function of the vehicle being off-grid and on-grid and the probability density function of the vehicle's driving distance. For example, when the electric vehicle is charging, the state of charge of the electric vehicle at moment t is related to the state of charge of the electric vehicle at the previous moment and the charging power of the electric vehicle at this moment, and the formula is as follows:

[0058] in, is the state of charge of the battery of the electric vehicle at time t; is the state of charge of the battery of the electric vehicle at time t-1; is the charging power of the battery on the electric vehicle at time t; is the charging efficiency coefficient of the battery.

[0059] For example, when an electric vehicle chooses to discharge, its state of charge at time t can be expressed as:

[0060] in, is the discharge power of the battery on the electric vehicle at time t; is the discharge efficiency coefficient of the battery.

[0061] 402. Establish a generalized energy storage model for a building according to the indoor temperature of the building and the set upper and lower limits of the allowable heating temperature; In the embodiment of the present invention, the current execution end determines the heat load of the building based on the indoor temperature of the building, the equivalent heat capacity coefficient, the equivalent heat loss coefficient, and the ambient temperature around the building. The specific formula is as follows:

[0062] in, is the heat load of building i at time t; , are the equivalent heat capacity coefficient and equivalent heat loss coefficient of building i respectively; is the indoor temperature of building i at time t; is the ambient temperature around the building at time t.

[0063] The above non-homogeneous differential equation of building heat load is transformed into an equivalent virtual thermal energy storage model that is basically consistent with the conventional energy storage expression and the building i at time t-1 is used as an initial condition to integrate the above equation to obtain the following special solution form:

[0064] in, is the indoor temperature of building i at time t-1.

[0065] The heat load at the lowest allowable heating temperature (16°C) of the building is set as the benchmark heat load. If the actual heat load of the building deviates positively from the benchmark heat load, the building can be regarded as being in a heat storage state; otherwise, the building is regarded as being in a heat release state. Assume that when the indoor temperature of the building is at the lowest allowable heating temperature, the heat storage of the building is 0; when the indoor temperature of the building is at the highest allowable heating temperature, the heat storage of the building is 100%. The heat storage state of building i at time t can be expressed as follows:

[0066] in, is the heat storage state of building i at time t, is the indoor temperature of building i at time t, is the minimum allowable heating temperature, The maximum allowable heating temperature.

[0067] Next, the current execution end determines the building heat storage state at each moment in the scheduling cycle according to the set upper and lower limits of the allowable heating temperature and the heat load. That is, the above formula is further sorted out to obtain the final building virtual energy storage model formula, as follows:

[0068] in, is the heat storage state of building i at time t+1, is the heat storage state of building i at time t, is the self-heating coefficient, is the heat storage / release efficiency, is the heat load of building i at time t+1.

[0069] 403. A generalized energy storage model for the air conditioning cluster is established based on the upper and lower limits of the temperature set by the air conditioner.

[0070] In the embodiment of the present invention, the current execution end first uses the fuzziness of human body's temperature perception to establish a mechanism model of a single air conditioner:

[0071] in, , are the heat storage and electrical power of the air conditioner respectively; , are the equivalent heat capacity and thermoelectric conversion coefficient, respectively; It is the virtual power of air conditioning load.

[0072] When increasing the indoor temperature To the maximum allowable temperature of the user When the discharge limit is reached, the available virtual energy storage capacity is 0, so the virtual energy storage capacity of the air conditioning load is:

[0073] When the virtual storage capacity of air conditioning load is at indoor temperature The lowest temperature acceptable to the user Raise to maximum temperature The virtual energy storage capacity reaches its maximum value when It can be expressed as: .

[0074] Further, as a refinement and extension of the specific implementation of the above embodiment, in order to comprehensively consider the economic and comfort requirements of the virtual power plant operation, another non-cooperative game scheduling method for a virtual power plant based on generalized energy storage is provided, wherein a non-cooperative game model of virtual players is established based on the generalized energy storage model and the prediction results of the new energy output, including: Construct an economic objective function based on the operation and maintenance costs and energy purchase costs of the virtual power plant; In the embodiment of the present invention, the economic objective function can be expressed by the following formula:

[0075]

[0076]

[0077] in, is the economic target value; , They are the virtual power plant operation and maintenance costs and energy purchase costs; is the equipment operation and maintenance cost coefficient; Provides power for equipment operation; and They are time-of-use electricity price and electricity sales price; To buy and sell electrical power.

[0078] Based on the generalized energy storage model, the start and stop states of the air conditioner and the electric vehicle at each time in the scheduling cycle are determined, and the comfort objective function is constructed in combination with the thermal comfort index and the time comfort; In the embodiment of the present invention, the comfort objective function can be expressed by the following formula:

[0079]

[0080]

[0081] in, For thermal comfort; For time comfort; It is the thermal comfort index; is the time comfort index; Set a value for comfort time; The start and stop status of the air conditioner; It is the start-stop state of the electric vehicle.

[0082] In the embodiment of the present invention, the current execution end determines the economic objective function and the comfort objective function as the profit function of the virtual game player, and optimally establishes the non-cooperative game model of the virtual game player by solving the minimum value of the profit function.

[0083] It should be noted that due to the existence of absolute value terms and square terms, the profit function formula (36) is not differentiable and does not meet the condition that there is a unique Nash equilibrium solution for non-cooperative games. Therefore, it is necessary to replace the original non-differentiable non-convex profit function with a differentiable quasi-convex function by adding constraints and replacing variables. The embodiment of the present invention converts the above profit function formula (36) into the following formula:

[0084]

[0085]

[0086]

[0087]

[0088]

[0089] in, , is the original formula (35) New related auxiliary variables; , In the original formula and in the original formula (38) The new variables of the above related variables must meet the following constraints:

[0090]

[0091]

[0092]

[0093] In the embodiment of the present invention, when the current execution end optimizes and solves the non-cooperative game model of the virtual players, the profit function is first reconstructed by the Nikaido-Isoda method. The specific formula is as follows:

[0094] in, is the profit function of the virtual player v; and They are the game strategies of the virtual player v and the game strategies of the game participants other than v respectively; is the changed strategy of the virtual player v. In the embodiment of the present invention, when v=1, it is a virtual player with an economic objective function, and when v=2, it is a virtual player with a comfort objective function. The Nikaido-Isoda function introduces a series of variables and parameters to describe the preferences, constraints, and mutual strategic influences of these participants. By solving this function, the equilibrium point that makes all participants reach the optimal state, namely the Nash equilibrium, can be found.

[0095] Next, the current execution end regularizes the reconstructed profit function. When the strategy of the virtual game player v changes while the strategies of other players remain unchanged, the change in total profit is regularized to obtain:

[0096] in, is a constant coefficient greater than 0.

[0097] Then, the current execution end converts the solution of Nash equilibrium into the problem of finding the global optimal solution based on the regularized function. The specific process is to define the function as follows:

[0098] If and only if ( is the optimal strategy) is a Nash equilibrium, Therefore, the global optimal problem is as follows:

[0099] in, is the policy set of x.

[0100] After obtaining the global optimal problem, the current execution end also needs to establish a non-cooperative game model strategy space including distribution network constraints, power purchase constraints, photovoltaic constraints, upper and lower limits of equipment output, energy bus balance, etc. The strategy space of the economic player is composed of distribution network constraints, power purchase constraints, photovoltaic constraints, wind power generation constraints, power storage equipment constraints, and power balance constraints; the strategy space of the comfort player is composed of the remaining equipment constraints. Then, MATLAB, Yalmip, and Ipopt are used to solve the global quasi-convex function to obtain the optimal scheduling result. The embodiment of the present invention provides a non-cooperative game scheduling method for a virtual power plant based on generalized energy storage. Compared with the prior art, the present invention collects environmental parameters that affect the output of new energy, and screens and processes the environmental parameters to obtain environmental influencing factors that meet the correlation requirements; predicts and processes the output of new energy based on the environmental influencing factors to obtain the prediction results of the output of new energy; establishes a generalized energy storage model, and establishes a virtual player non-cooperative game model based on the generalized energy storage model and the prediction results of the output of new energy; optimizes and solves the virtual player non-cooperative game model to obtain the scheduling results of the virtual power plant, and completes the power scheduling based on generalized energy storage. The present invention comprehensively considers the economy, user thermal comfort, and time comfort, and proposes a virtual player non-cooperative game model with economy and multiple comfort as the goals, so that the virtual power plant considers the economy and comfort of each subject in the virtual power plant while supplying energy to the regional power load, maintains a stable, continuous and predictable power supply state and power supply efficiency, and maintains power supply safety.

[0101] As the above Figure 1 The implementation of the method shown in the figure, the embodiment of the present invention provides a non-cooperative game scheduling device for a virtual power plant based on generalized energy storage, such as Figure 5 As shown, the device comprises: The factor screening module 51 is used to collect environmental parameters that affect the output of new energy sources, and screen the environmental parameters to obtain environmental influencing factors that meet the relevance requirements; A prediction module 52, configured to perform prediction processing on the output of the new energy source based on the environmental influencing factors to obtain a prediction result of the output of the new energy source; A model building module 53 is used to build a generalized energy storage model, and to build a virtual player non-cooperative game model based on the generalized energy storage model and the prediction result of the new energy output; The solution module 54 is used to optimize and solve the non-cooperative game model of the virtual players to obtain the scheduling result of the virtual power plant.

[0102] Furthermore, the factor screening module 51 is also used for: Setting each of the environmental parameters as node construction graph structure data; Using a multivariate correlation analysis method to calculate the correlation of each of the environmental parameters in the graph structure data, and obtaining a correlation value corresponding to each of the environmental parameters; Based on the correlation values, environmental parameters that do not meet the correlation requirements are eliminated to obtain the environmental influencing factors that meet the correlation requirements.

[0103] Furthermore, the prediction result includes a day-ahead prediction result and an intraday prediction result, and the prediction module 52 is further used for: The gated recurrent unit is used to improve the graph neural network to obtain an improved graph neural network model; Based on the improved graph neural network model and the environmental influencing factors, the new energy output is predicted and processed in the day-ahead stage to obtain the day-ahead prediction result; The intraday stage is divided into multiple prediction cycles, and the new energy output of the current prediction cycle is predicted based on the actual output of new energy in the previous prediction cycle, the improved graph neural network model and the environmental influencing factors until all prediction cycles are processed, and the intraday prediction results corresponding to each prediction cycle are obtained.

[0104] Further, the generalized energy storage model includes an electric vehicle generalized energy storage model, an air conditioning cluster generalized energy storage model, and an air conditioning cluster generalized energy storage model, and the model building module 53 includes a first building unit, a second building unit, and a third building unit; The first establishing unit is used to establish the generalized energy storage model of the electric vehicle according to the state characteristics of the electric vehicle; The second establishing unit is used to establish the generalized energy storage model of the building according to the indoor temperature of the building and the set upper and lower limits of the allowable heating temperature; The third establishing unit is used to establish the generalized energy storage model of the air-conditioning cluster according to the upper and lower limits of the temperature set by the air-conditioning.

[0105] Furthermore, the first establishing unit is further used for: Sampling each electric vehicle to obtain state characteristics of each electric vehicle; and classifying each electric vehicle based on the state characteristics to obtain a vehicle classification result; Based on the automobile classification results and the state characteristics of each type of automobile, determine the probability density function of automobile off-grid and on-grid and the probability density function of automobile driving distance; The dispatchable power of the vehicle at each moment in the dispatch cycle is determined according to the probability density function of the vehicle being off-grid and on-grid and the probability density function of the vehicle's travel distance.

[0106] Furthermore, the second establishing unit is further used for: Determine the heat load of the building based on the indoor temperature of the building, the equivalent heat capacity coefficient, the equivalent heat loss coefficient, and the ambient temperature around the building; The heat storage state of the building at each moment in the scheduling cycle is determined according to the set upper and lower limits of the allowable heating temperature and the heat load.

[0107] Furthermore, the model building module 53 further includes a fourth building unit, which is used to: Construct an economic objective function based on the operation and maintenance costs and energy purchase costs of the virtual power plant; Based on the generalized energy storage model, the start and stop states of the air conditioner and the electric vehicle at each time in the scheduling cycle are determined, and the comfort objective function is constructed in combination with the thermal comfort index and the time comfort; The economic objective function and the comfort objective function are determined as the profit functions of the virtual game players, and the non-cooperative game model of the virtual game players is optimally established by solving the minimum value of the profit function.

[0108] The embodiment of the present invention provides a non-cooperative game scheduling device for a virtual power plant based on generalized energy storage. Compared with the prior art, the present invention collects environmental parameters that affect the output of new energy, and screens and processes the environmental parameters to obtain environmental influencing factors that meet the correlation requirements; predicts and processes the output of new energy based on the environmental influencing factors to obtain the predicted results of the output of new energy; establishes a generalized energy storage model, and establishes a virtual player non-cooperative game model based on the generalized energy storage model and the predicted results of the output of new energy; optimizes and solves the virtual player non-cooperative game model to obtain the scheduling results of the virtual power plant, and completes the power scheduling based on generalized energy storage. The present invention comprehensively considers the economy, user thermal comfort, and time comfort, and proposes a virtual player non-cooperative game model with economy and multiple comfort as the goals, so that the virtual power plant considers the economy and comfort of each subject in the virtual power plant while supplying energy to the regional electric load, maintains a stable, continuous and predictable power supply state and power supply efficiency, and maintains power supply safety.

[0109] According to one embodiment of the present invention, a storage medium is provided, wherein the storage medium stores at least one executable instruction, and the computer executable instruction can execute the non-cooperative game scheduling method for a virtual power plant based on generalized energy storage in any of the above method embodiments.

[0110] Figure 6 A schematic diagram of the structure of a device provided according to an embodiment of the present invention is shown, and the specific embodiment of the present invention does not limit the specific implementation of the device.

[0111] like Figure 6 As shown, the device may include: a processor (processor) 602 , a communication interface (Communications Interface) 604 , a memory (memory) 606 , and a communication bus 608 .

[0112] The processor 602 , the communication interface 604 , and the memory 606 communicate with each other via a communication bus 608 .

[0113] The communication interface 604 is used to communicate with other devices such as clients or other servers.

[0114] Processor 602 is used to execute program 610, and specifically can execute relevant steps of the above-mentioned virtual power plant non-cooperative game scheduling method based on generalized energy storage.

[0115] Specifically, the program 610 may include program codes, which include computer operation instructions.

[0116] The processor 602 may be a central processing unit (CPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiment of the present invention. The one or more processors included in the device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0117] The memory 606 is used to store the program 610. The memory 606 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0118] The program 610 may be specifically configured to enable the processor 602 to perform the following operations: Collect environmental parameters that affect the output of new energy, and screen and process the environmental parameters to obtain environmental influencing factors that meet the relevance requirements; Performing prediction processing on the output of the new energy source based on the environmental influencing factors to obtain a prediction result of the output of the new energy source; Establishing a generalized energy storage model, and establishing a virtual player non-cooperative game model based on the generalized energy storage model and the prediction results of the new energy output; The non-cooperative game model of the virtual players is optimized and solved to obtain the dispatch result of the virtual power plant.

[0119] Obviously, those skilled in the art should understand that the above modules or steps of the present invention can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present invention is not limited to any specific combination of hardware and software.

[0120] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A non-cooperative game scheduling method for virtual power plants based on generalized energy storage, characterized in that: include: Collect environmental parameters that affect the output of new energy, and screen and process the environmental parameters to obtain environmental influencing factors that meet the relevance requirements; Performing prediction processing on the output of the new energy source based on the environmental influencing factors to obtain a prediction result of the output of the new energy source; Establishing a generalized energy storage model, and establishing a virtual player non-cooperative game model based on the generalized energy storage model and the prediction results of the new energy output; The non-cooperative game model of the virtual players is optimized and solved to obtain the dispatch result of the virtual power plant.

2. The method according to claim 1, characterized in that The environmental parameters are screened to obtain environmental influencing factors that meet the relevance requirements, including: Setting each of the environmental parameters as node construction graph structure data; Using a multivariate correlation analysis method to calculate the correlation of each of the environmental parameters in the graph structure data, and obtaining a correlation value corresponding to each of the environmental parameters; Based on the correlation values, environmental parameters that do not meet the correlation requirements are eliminated to obtain the environmental influencing factors that meet the correlation requirements.

3. The method according to claim 1, characterized in that The prediction result includes a day-ahead prediction result and a day-intraday prediction result. The prediction processing of the new energy output based on the environmental influencing factors to obtain the prediction result of the new energy output includes: The gated recurrent unit is used to improve the graph neural network to obtain an improved graph neural network model; Based on the improved graph neural network model and the environmental influencing factors, the new energy output is predicted and processed in the day-ahead stage to obtain the day-ahead prediction result; The intraday stage is divided into multiple prediction cycles, and the new energy output of the current prediction cycle is predicted based on the actual output of new energy in the previous prediction cycle, the improved graph neural network model and the environmental influencing factors until all prediction cycles are processed, and the intraday prediction results corresponding to each prediction cycle are obtained.

4. The method according to claim 1, characterized in that: The generalized energy storage model includes an electric vehicle generalized energy storage model, an air conditioning cluster generalized energy storage model, and an air conditioning cluster generalized energy storage model. The establishment of the generalized energy storage model includes: Establishing a generalized energy storage model of the electric vehicle according to the state characteristics of the electric vehicle; Establishing a generalized energy storage model of the building according to the indoor temperature of the building and the set upper and lower limits of the allowable heating temperature; The generalized energy storage model of the air-conditioning cluster is established according to the upper and lower limits of the temperature set by the air-conditioning.

5. The method according to claim 4, characterized in that The generalized energy storage model of the electric vehicle is established according to the state characteristics of the electric vehicle, including: Sampling each electric vehicle to obtain state characteristics of each electric vehicle; and classifying each electric vehicle based on the state characteristics to obtain a vehicle classification result; Based on the automobile classification results and the state characteristics of each type of automobile, determine the probability density function of automobile off-grid and on-grid and the probability density function of automobile driving distance; The dispatchable power of the vehicle at each moment in the dispatch cycle is determined according to the probability density function of the vehicle being off-grid and on-grid and the probability density function of the vehicle's travel distance.

6. The method according to claim 4, characterized in that The generalized energy storage model of the building is established according to the indoor temperature of the building and the set upper and lower limits of the allowable heating temperature, including: Determine the heat load of the building based on the indoor temperature of the building, the equivalent heat capacity coefficient, the equivalent heat loss coefficient, and the ambient temperature around the building; The heat storage state of the building at each moment in the scheduling cycle is determined according to the set upper and lower limits of the allowable heating temperature and the heat load.

7. The method according to claim 1, characterized in that The non-cooperative game model of virtual players is established based on the generalized energy storage model and the prediction results of the new energy output, including: Construct an economic objective function based on the operation and maintenance costs and energy purchase costs of the virtual power plant; Based on the generalized energy storage model, the start and stop states of the air conditioner and the electric vehicle at each time in the scheduling cycle are determined, and the comfort objective function is constructed in combination with the thermal comfort index and the time comfort; The economic objective function and the comfort objective function are determined as the profit functions of the virtual game players, and the non-cooperative game model of the virtual game players is optimally established by solving the minimum value of the profit function.

8. A non-cooperative game scheduling device for a virtual power plant based on generalized energy storage, characterized in that: include: A factor screening module is used to collect environmental parameters that affect the output of new energy, and screen the environmental parameters to obtain environmental influencing factors that meet the relevance requirements; A prediction module, used for predicting the output of the new energy source based on the environmental influencing factors to obtain a prediction result of the output of the new energy source; A model building module, used to build a generalized energy storage model, and to build a virtual player non-cooperative game model based on the generalized energy storage model and the prediction result of the new energy output; The solution module is used to optimize and solve the non-cooperative game model of the virtual players to obtain the scheduling result of the virtual power plant.

9. A storage medium storing at least one executable instruction, wherein the executable instruction executes operations corresponding to the non-cooperative game scheduling method for a virtual power plant based on generalized energy storage as described in any one of claims 1 to 7.

10. A device comprising a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform operations corresponding to the non-cooperative game scheduling method for a virtual power plant based on generalized energy storage as described in any one of claims 1-7.

Images