Virtual power plant joint scheduling method, device, equipment and medium

Abstract

The invention discloses a virtual power plant combined dispatching method, device and equipment and a medium, and relates to the technical field of virtual power plant combined dispatching, the combined dispatching method comprises the steps that a combined dispatching model of a virtual power plant is constructed, and the combined dispatching model is used for taking the minimum dispatching cost as an optimization target; and scheduling the cooperative operation process among the photovoltaic power generation equipment, the energy storage equipment, the electric vehicle charging station, the electric vehicle battery swap station, the air conditioning load and the power grid under the constraint of the constraint condition. The constraint conditions comprise an electric power balance constraint, an energy storage equipment constraint, an electric vehicle charging station constraint, an electric vehicle battery swap station constraint, an air conditioner load constraint and a power grid interaction power constraint, and then solving the combined scheduling model to obtain a combined scheduling scheme of the virtual power plant. The operation risk of the virtual power plant can be reduced, and the operation efficiency, stability, flexibility and reliability of the virtual power plant can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of virtual power plant combined dispatching, and in particular, to a virtual power plant combined dispatching method, device, equipment and medium. Background Art

[0002] With the rapid development of renewable energy, virtual power plants (VPPs) have gradually become an important dispatching mode in the power system. By integrating various distributed energy resources, such as photovoltaic power generation equipment, energy storage equipment, electric vehicle charging and swapping stations (including electric vehicle charging stations and electric vehicle swapping stations), air-conditioning loads, etc., virtual power plants can provide flexible power supply solutions. By optimizing the dispatching of these distributed energy resources, virtual power plants can balance power supply and demand, optimize grid load, and provide stability for the power market. However, virtual power plants face many technical challenges in practical applications, especially the lack of effective combined dispatching schemes.

[0003] As a key regulation means in virtual power plants, energy storage equipment can effectively balance the volatility of photovoltaic power generation. However, most traditional virtual power plant dispatching methods focus on the coordination between energy storage equipment and the power grid, while ignoring the combined dispatching of energy storage equipment with renewable energy and electric vehicle charging and swapping stations. This single dispatching method cannot fully utilize the synergistic effect of various resources, easily leads to low system operation efficiency, and fails to achieve the optimal allocation of resources.

[0004] As an important controllable load in virtual power plants, air-conditioning loads have significant flexibility and adjustability, which provides potential for dispatching optimization. By adjusting the operating power, start-stop state and set temperature, air-conditioning loads can quickly participate in demand response. At the same time, air-conditioning loads need to balance between user comfort and dispatching economy, and excessive dispatching may cause user dissatisfaction. Nevertheless, air-conditioning loads can participate in demand response and collaborative optimization through flexibility, form combined dispatching with photovoltaic power generation equipment, electric vehicle charging and swapping stations, and energy storage equipment, improve system operation efficiency and resource utilization rate, and enhance stability.

[0005] Most current virtual power plant dispatching methods still lack effective combined dispatching schemes and fail to synergistically optimize multiple resources such as photovoltaic power generation equipment, electric vehicle charging and swapping stations, energy storage equipment, and air-conditioning loads. Related virtual power plant dispatching methods usually regard various resources as independent dispatching objects and fail to perform global optimization within a unified framework. This dispatching method lacking coordination between resources increases the operation risk of the system. In a complex power market environment, the dispatching strategy of a single resource is difficult to cope with the rapidly changing power demand and supply situation, resulting in a significant reduction in the flexibility and reliability of the system. Summary of the Invention

[0006] The purpose of this application is to provide a virtual power plant joint scheduling method, device, equipment and medium, which can reduce the operation risk of the virtual power plant and improve the operation efficiency, stability, flexibility and reliability of the virtual power plant.

[0007] To achieve the above object, the present application provides the following solutions.

[0008] In a first aspect, the present application provides a virtual power plant joint scheduling method, and the virtual power plant joint scheduling method includes: Construct a joint scheduling model of the virtual power plant; the joint scheduling model is used to optimize the objective of minimizing the scheduling cost, and under the constraints of the constraint conditions, schedule the coordinated operation process among the photovoltaic power generation equipment, energy storage equipment, electric vehicle charging station, electric vehicle swapping station, air conditioning load and the power grid. The constraint conditions include power balance constraint, energy storage equipment constraint, electric vehicle charging station constraint, electric vehicle swapping station constraint, air conditioning load constraint and grid interaction power constraint; Solve the joint scheduling model to obtain a joint scheduling plan for the virtual power plant; the joint scheduling plan includes the charging power and discharging power of the energy storage equipment, the charging power and discharging power of the electric vehicles, the charging power and discharging power of each battery in the electric vehicle swapping station, the reducible power of the air conditioning load, and the interaction power between the virtual power plant and the power grid. The interaction power is the power purchase power of the virtual power plant from the power grid or the power selling power of the virtual power plant to the power grid.

[0009] In a second aspect, the present application provides a virtual power plant joint scheduling device, and the virtual power plant joint scheduling device includes: A model construction module, configured to construct a joint scheduling model of the virtual power plant; the joint scheduling model is used to optimize the objective of minimizing the scheduling cost, and under the constraints of the constraint conditions, schedule the coordinated operation process among the photovoltaic power generation equipment, energy storage equipment, electric vehicle charging station, electric vehicle swapping station, air conditioning load and the power grid. The constraint conditions include power balance constraint, energy storage equipment constraint, electric vehicle charging station constraint, electric vehicle swapping station constraint, air conditioning load constraint and grid interaction power constraint; A model solving module, configured to solve the joint scheduling model to obtain a joint scheduling plan for the virtual power plant; the joint scheduling plan includes the charging power and discharging power of the energy storage equipment, the charging power and discharging power of the electric vehicles, the charging power and discharging power of each battery in the electric vehicle swapping station, the reducible power of the air conditioning load, and the interaction power between the virtual power plant and the power grid. The interaction power is the power purchase power of the virtual power plant from the power grid or the power selling power of the virtual power plant to the power grid.

[0010] In a third aspect, the present application provides a computer device, including: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor executes the computer program to implement the above-mentioned virtual power plant joint scheduling method.

[0011] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the above-mentioned virtual power plant joint scheduling method is implemented.

[0012] According to the specific embodiments provided by the present application, the present application has the following technical effects: The present application provides a virtual power plant joint scheduling method, device, equipment and medium, constructs a joint scheduling model of the virtual power plant, and the joint scheduling model is used to optimize the collaborative operation process among photovoltaic power generation equipment, energy storage equipment, electric vehicle charging stations, electric vehicle battery swapping stations, air-conditioning loads and the power grid with the minimum scheduling cost as the optimization goal, and the constraints include power balance constraints, energy storage equipment constraints, electric vehicle charging station constraints, electric vehicle battery swapping station constraints, air-conditioning load constraints and grid interaction power constraints, and then solves the joint scheduling model to obtain the joint scheduling plan of the virtual power plant. By constructing and solving the joint scheduling model, the present application can globally optimize various resources in the virtual power plant within a unified framework, reduce the operation risk of the virtual power plant, improve the resource utilization rate, and thus improve the operation efficiency, stability, flexibility and reliability of the virtual power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and those of ordinary skill in the art can also obtain other drawings based on these drawings without creative efforts.

[0014] Figure 1 It is an application environment diagram of a virtual power plant joint scheduling method provided in Embodiment 1 of the present application.

[0015] Figure 2 It is a schematic flowchart of a virtual power plant joint scheduling method provided in Embodiment 1 of the present application.

[0016] Figure 3 It is a schematic diagram of an optimization strategy framework provided in Embodiment 1 of the present application.

[0017] Figure 4 It is a schematic diagram of a rolling horizon optimization algorithm provided in Embodiment 1 of the present application.

[0018] Figure 5Schematic diagram of the virtual power plant multi-factor joint scheduling strategy based on the rolling horizon optimization algorithm and the mixture density neural network provided in Embodiment 1 of this application.

[0019] Figure 6 Schematic diagram of the functional modules of a virtual power plant joint scheduling device provided in Embodiment 2 of this application.

[0020] Figure 7 Schematic diagram of the structure of a computer device provided in Embodiment 3 of this application. Detailed implementation manners

[0021] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0022] Embodiment 1.

[0023] The virtual power plant joint scheduling method provided in the embodiment of this application can be applied to the application environment as Figure 1 shown. Among them, the terminal communicates with the server through the network. The data storage system can store the data that the server needs to process. The data storage system can be set up separately, integrated on the server, or placed on the cloud or other servers. The terminal can send the joint scheduling request to be processed to the server. After receiving the joint scheduling request to be processed, for the joint scheduling request to be processed, the server constructs a joint scheduling model of the virtual power plant; solves the joint scheduling model to obtain a joint scheduling plan for the virtual power plant. The server can feedback the joint scheduling result, that is, the joint scheduling plan for the joint scheduling request, to the terminal.

[0024] In addition, in some embodiments, the virtual power plant joint scheduling method can also be implemented by the server or the terminal alone. For example, the terminal can directly process the joint scheduling request to be processed, or the server can obtain the joint scheduling request to be processed from the data storage system and process the joint scheduling request to be processed.

[0025] Among them, the terminal can be, but is not limited to, various desktop computers, laptop computers, smart phones, tablet computers, Internet of Things devices, and portable wearable devices. The Internet of Things devices can be smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, etc. The portable wearable devices can be smart watches, smart bracelets, head-mounted devices, etc. The server can be implemented by an independent server or a server cluster composed of multiple servers, and can also be a cloud server.

[0026] In an exemplary embodiment, as Figure 2 shown, a virtual power plant joint dispatching method is provided. This method is executed by a computer device, specifically, it can be executed alone by a computer device such as a terminal or a server, or jointly executed by a terminal and a server. In the embodiments of the present application, taking the case where this method is applied to Figure 1 the server as an example for illustration, it includes the following steps.

[0027] Step S1, construct a joint dispatching model of the virtual power plant; the joint dispatching model is used to take the minimum dispatching cost as the optimization goal, and under the constraints of the constraint conditions, dispatch the collaborative operation process among photovoltaic power generation equipment, energy storage equipment, electric vehicle charging stations, electric vehicle swapping stations, air-conditioning loads, and the power grid. The constraint conditions include power balance constraints, energy storage equipment constraints, electric vehicle charging station constraints, electric vehicle swapping station constraints, air-conditioning load constraints, and power grid interaction power constraints.

[0028] Step S2, solve the joint dispatching model to obtain a joint dispatching plan for the virtual power plant; the joint dispatching plan includes the charging power and discharging power of the energy storage equipment, the charging power and discharging power of the electric vehicles, the charging power and discharging power of each battery in the electric vehicle swapping station, the reducible power of the air-conditioning load, and the interaction power between the virtual power plant and the power grid. The interaction power is the power purchase power of the virtual power plant from the power grid or the power selling power of the virtual power plant to the power grid.

[0029] Implementing the above steps S1 to S2, this embodiment provides a virtual power plant joint dispatching method based on the coordination of photovoltaic - energy storage - charging - swapping - air - conditioning load. By constructing a joint dispatching model of the virtual power plant and solving the joint dispatching model, a joint dispatching plan for the virtual power plant is obtained. Since various resources in the virtual power plant are considered when constructing the joint dispatching model, solving the joint dispatching model can globally optimize various resources in the virtual power plant within a unified framework. Subsequently, operating according to the joint dispatching plan can reduce the operation risk of the virtual power plant and improve the operation efficiency, stability, flexibility, and reliability of the virtual power plant.

[0030] As Figure 3As shown in the figure, in the virtual power plant of this embodiment, the photovoltaic power generation equipment, energy storage equipment, electric vehicle charging station, electric vehicle battery swapping station, air-conditioning load, and power grid are all connected to the DC bus of the virtual power plant. In this embodiment, a joint scheduling model considering components such as electric vehicles is built. According to the uncertainty of photovoltaic power generation, the charging and discharging constraints of energy storage equipment, the reducible characteristics of air-conditioning load, and the charging and swapping behaviors of electric vehicles, a comprehensive joint scheduling model is specifically constructed in this embodiment. This joint scheduling model considers the coordination and optimization among multiple resources (photovoltaic power generation equipment, energy storage equipment, electric vehicle charging and swapping stations, air-conditioning load), and the goal is to minimize the scheduling cost of the virtual power plant. In the joint scheduling model, the charging and discharging strategies of energy storage equipment are optimized according to the fluctuations of grid load and photovoltaic power generation. While ensuring power balance and system stability, the joint scheduling model also considers the operating costs and resource limitations of various resources to ensure the efficient and economic scheduling of various resources.

[0031] The joint scheduling model considering multi-scenario uncertainty built in this embodiment includes an objective function and constraint conditions. The objective function is to minimize the scheduling cost. This joint scheduling model is used to optimize the collaborative operation process among the photovoltaic power generation equipment, energy storage equipment, electric vehicle charging station, electric vehicle battery swapping station, air-conditioning load, and power grid with the minimum scheduling cost as the optimization goal under the constraints of the constraint conditions.

[0032] For the objective function, due to the involvement of scenario sampling and probability problems, the minimum expected value of the scheduling cost can be used as the optimization goal at this time. The expression of the objective function is as follows: ; Among them, is the scheduling period; is the number of scenarios; is the probability of scenario, is the centroid of the cluster corresponding to the scenario; is the income obtained from electric vehicle charging in the scenario; is the income obtained from electric vehicle battery swapping in the scenario; is the cost of trading with the power grid in the power market in the scenario; is the compensation expenditure cost of electric vehicles in the scenario; is the operating depreciation cost of energy storage equipment in the scenario.

[0033] ; Among them, is the The number of electric vehicles participating in charging in the scenario; is the state of charge of the battery when the th electric vehicle finishes charging in the scenario; is the state of charge of the battery when the th electric vehicle starts charging in the scenario; is the scheduling step size, which can be 1h in this embodiment; is the energy capacity of the th electric vehicle in the scenario; is the charging unit price of the electric vehicle, which can be billed according to time-of-use electricity price to improve the ability of vehicle owners to respond to demand; is the number of batteries in the electric vehicle swapping station; is the state of charge of the new battery after the electric vehicle swapping station in the th electric vehicle in the scenario replaces the battery; is the state of charge of the old battery (i.e., the discharged battery being replaced) after the electric vehicle swapping station in the th electric vehicle in the scenario replaces the battery; is the energy capacity of the th battery; is the unit price of electric vehicle battery swapping; is the interactive power between the virtual power plant and the power grid in the time period in the scenario; is the time-of-use electricity price of the power grid in the is the on-grid electricity price of the power grid in the is the incentive expenditure cost in the scenario; is the coefficient of the compensation cost for electric vehicles; is the discharge power of the energy storage device in the time period; is the charging power of the energy storage device in the time period;

[0034] This embodiment designs an incentive mechanism for electric vehicles. To encourage electric vehicle charging and swapping stations to participate in virtual power plant dispatching, this embodiment designs a comprehensive incentive mechanism. For electric vehicles, the incentive mechanism combines factors such as charging time periods, battery health status, and charging demands of electric vehicles to set dynamic electricity prices, giving certain electricity price incentives to electric vehicle charging and swapping stations to guide electric vehicles to charge according to grid demands. This mechanism can not only reduce the impact of charging on the grid load but also delay battery degradation and improve the service life of electric vehicles.

[0035] Specifically, the electric vehicle incentive plan designed in this embodiment is as follows: To encourage electric vehicle owners to participate in VPP dispatching, an innovative compensation plan is proposed, aiming to adjust incentive measures and maximize the use of available storage capacity within the VPP and reduce the pressure brought by dispatching uncertainties. As long as the vehicle owner agrees to participate in VPP dispatching and enables the VPP operator to discharge electricity using its battery, the owner can obtain benefits without considering the parking time of the electric vehicle, enabling the participation degree of electric vehicles to be dynamically adjusted through the incentive mechanism.

[0036] Establish a battery aging model to calculate the battery loss compensation for electric vehicle owners participating in the VPP discharging process. Its expression is as follows: ; Where, is the battery loss compensation; is the attenuation rate of the battery for each charge and discharge cycle; is the battery at the battery discharge energy during the time period; is the maximum energy of the battery when fully charged; is the battery at the depth of discharge during the time period; is the total cost of the battery.

[0037] Based on this, the incentive plan is designed to compensate for the battery wear of electric vehicles and encourage their participation in grid services. The incentive plan can be set as: ; Where, is the incentive expenditure cost under the scenario; is the battery loss compensation coefficient; is the battery participation in grid service compensation coefficient; is at the incentive coefficient for participating in grid service discharging (such as frequency support, load balancing, etc.) during the time period; is the energy for the VPP to dispatch electric vehicles at during the time period under the scenario; is the total capacity of the electric vehicle battery.

[0038] For the constraints, the constraints include power balance constraint, energy storage device constraint, electric vehicle charging station constraint, electric vehicle battery swapping station constraint, air-conditioning load constraint, and grid interaction power constraint.

[0039] The expression of the power balance constraint inside the virtual power plant is as follows: ; Among them, is the photovoltaic power generation of the photovoltaic power generation equipment in the time period; is the discharge power of the energy storage device in the time period; is the number of electric vehicles participating in charging in the is the th electric vehicle's discharge power in the is the interaction power between the virtual power plant and the grid in the time period, including the power purchase power (positive value) for purchasing electricity from the grid or the power selling power (negative value) for selling electricity to the grid; is the number of batteries in the electric vehicle battery swapping station; is the th battery's discharge power in the is the basic load demand of the virtual power plant in the time period, and the basic load demand can include ordinary load and air-conditioning load; is the charging power of the energy storage device in the time period; is the th electric vehicle's charging power in the is the air-conditioning load's reducible power in the is the th battery's charging power in the

[0040] The energy storage device constraints include charge and discharge power limits, SOC (State of Charge) update, and SOC constraints, and the expressions are as follows: ; Among them, is the maximum charging power of the energy storage device; is the first 0-1 state variable, representing the charging state of the energy storage device during the period. Being 1 represents charging, and being 0 represents discharging, ensuring that the energy storage device cannot charge and discharge simultaneously; is the maximum discharge power of the energy storage device; is the state of charge of the energy storage device during the period; is the state of charge of the energy storage device during the period; is the charging efficiency of the energy storage device; is the discharging efficiency of the energy storage device; is the minimum state of charge of the energy storage device; is the maximum state of charge of the energy storage device; is the state of charge of the energy storage device at the 1st period; is the state of charge of the energy storage device during the period, is the scheduling period.

[0041] The constraints of the electric vehicle charging station include charging power limit, discharging power limit, SOC update, SOC constraint, energy demand satisfaction, charging window, etc., and the expressions are as follows: ; Among them, is the maximum charging power of the th electric vehicle; is the second 0-1 state variable, representing the charging state of the th electric vehicle during the period. Being 1 represents charging, and being 0 represents discharging, ensuring that the electric vehicle cannot charge and discharge simultaneously; is the maximum discharging power of the th electric vehicle. If the owner of this electric vehicle does not accept the unified VPP scheduling, then is 0; is the third 0-1 state variable, representing whether during the scenario, the period is within the charging window of the th electric vehicle, that is, whether it is parked at the electric vehicle charging station. Being 1 represents being there, and being 0 represents not being there; is the start period of the charging window of the th electric vehicle, that is, the start period of parking; is the end period of the charging window of the th electric vehicle, that is, the end period of parking; is the state of the th electric vehicle during the The state of charge of the battery during a period; is the state of charge of the battery of the th electric vehicle during a period in the scenario; is the charging efficiency of the th electric vehicle; is the discharging efficiency of the th electric vehicle; is the minimum state of charge of the battery of the th electric vehicle; is the maximum state of charge of the battery of the th electric vehicle; is the state of charge of the battery of the th electric vehicle during a period in the scenario; is the target state of charge of the battery of the th electric vehicle in the scenario.

[0042] The constraints of an electric vehicle swapping station include swapping requirement constraints and energy storage battery constraints, and the expressions are as follows: ; Among them, is the maximum charging power of the th battery; is the fourth 0-1 state variable, representing the charging state of the th battery during a period in , 1 represents charging, and 0 represents discharging; is the fifth 0-1 state variable, representing whether the th battery is assigned a charger during a period in , 1 represents being assigned a charger, and 0 represents not being assigned a charger; is the maximum discharging power of the th battery. In engineering practice, usually the maximum charging power and the maximum discharging power of the th battery are equal, and the specific value is related to the battery model; is the number of batteries in the electric vehicle swapping station; is the number of chargers; is the sixth 0-1 state variable, representing whether the th battery can provide swapping service during a period in the scenario, 1 represents being able to provide swapping service, and 0 represents not being able to provide swapping service. The swapping service can be provided only when the state of charge of the battery is greater than 95%; is The state of charge of the th battery in the time period; is the seventh 0-1 state variable, representing whether battery swapping service is provided for the th battery in the time period in the scenario. 1 represents that battery swapping service is provided, and 0 represents that no battery swapping service is provided; is the eighth 0-1 state variable, representing whether battery swapping service is provided to the th electric vehicle in the time period in the scenario. 1 represents that battery swapping service is provided, and 0 represents that no battery swapping service is provided; is the number of electric vehicles arriving at the electric vehicle battery swapping station in the time period in the scenario; is the state of charge of the th battery in the time period in the scenario; is the initial state of charge of the th electric vehicle coming for battery swapping in the time period in the scenario; is the state of charge of the th battery in the time period in the scenario; is the charging efficiency of the th battery; is the discharging efficiency of the th battery; is the minimum state of charge of the th battery; is the maximum state of charge of the

[0043] For the air conditioning load constraint, since the air conditioning load can be slightly reduced in a short time without affecting user usage, it is modeled as an air conditioning type of load that can be reduced, including the air conditioning reduction amount limit and the total air conditioning reduction amount limit during the scheduling period. The expressions are as follows: ; ; Among them, is the maximum reducible power of the air conditioning load in the time period; is the maximum cumulative reducible power of the air conditioning load.

[0044] The grid interaction power constraint includes the grid interaction power limit, and the expression is as follows: ; where is the absolute value of the maximum value of the interaction power between the virtual power plant and the grid, that is, the maximum value of the power of the connection line between the grid and the virtual power plant.

[0045] Before solving the joint dispatching model, this embodiment first needs to collect the basic data related to the virtual power plant, including photovoltaic power generation data, meteorological data, electric vehicle charging and swapping station data, energy storage device operation data, and air-conditioning load data. The photovoltaic power generation data includes historical photovoltaic power generation, meteorological conditions (such as sunshine intensity), etc. The electric vehicle charging and swapping station data includes the charging and swapping power demand and charging and swapping time periods of the charging and swapping station. All data needs to be preprocessed to ensure the integrity and consistency of the data. In addition, all data also needs to be normalized to eliminate the scale differences of different data sources and ensure the comparability and effectiveness of various data in the model.

[0046] Solve the joint dispatching model to obtain the joint dispatching plan of the virtual power plant. The joint dispatching plan includes the charging power and discharging power of the energy storage device, the charging power and discharging power of the electric vehicle, the charging power and discharging power of each battery in the electric vehicle swapping station, the reducible power of the air-conditioning load and the interaction power between the virtual power plant and the grid . The interaction power is the power purchase power of the virtual power plant from the grid or the power selling power of the virtual power plant to the grid.

[0047] In practical applications, virtual power plants also face the technical challenge of source-load uncertainty. The volatility and uncertainty of renewable energy sources such as photovoltaic power generation are one of the biggest problems in the dispatching of virtual power plants. Photovoltaic power generation is affected by multiple factors such as weather, time, and season, and its output has significant uncertainty. At present, many virtual power plant dispatching models still adopt prediction methods based on historical data, assuming that the photovoltaic power generation can be accurately predicted. However, in reality, the volatility and sudden changes in photovoltaic power generation make this assumption often not hold. The lack of full consideration of uncertainty will lead to poor robustness of the virtual power plant dispatching strategy, and it is easy to over-rely on energy storage devices or the power grid, thus affecting the overall revenue and system stability. The charging and discharging behaviors of electric vehicles also bring source-load uncertainty. The charging and swapping demands of electric vehicles are not only related to the real-time usage status of the vehicles and the load of charging facilities, but also affected by individual factors such as the charging and swapping habits of vehicle owners, the selection of charging and swapping time periods, and the initial state of charge of the electric vehicle batteries when they arrive at the charging and swapping stations. At present, most dispatching models do not effectively consider the dynamic demand characteristics of electric vehicles, and lack the optimal management of the charging time periods and charging powers of electric vehicles, making it impossible for virtual power plants to achieve the best revenue when dispatching electric vehicles, and may even have an adverse impact on electric vehicle batteries.

[0048] To solve the above problems, in this embodiment, a prediction model for the uncertainty of photovoltaic power generation based on a mixture density network (MDN) is built. In order to accurately capture the volatility and uncertainty of photovoltaic power generation, this embodiment uses a mixture density network for the prediction of photovoltaic power generation. Different from traditional single prediction methods, MDN can output the probability distribution of photovoltaic power generation, rather than a simple point estimate, which is particularly important for subsequent dispatching optimization. Through the neural network structure, MDN learns the complex relationship between photovoltaic power generation and meteorological data and its time series relationship, and generates the parameters (mean, variance, and weight) of multiple Gaussian distributions, so as to describe the possible change range of photovoltaic power generation. During the model training process, the input data includes historical photovoltaic power generation data, meteorological data, and other relevant factors. MDN is trained through the backpropagation algorithm to ensure the high accuracy and robustness of the prediction results. Similarly, a prediction model for the uncertainty of electric vehicle charging and a prediction model for the uncertainty of electric vehicle swapping are established to use MDN to predict the electric vehicle load data (i.e., electric vehicle charging data and electric vehicle swapping data), so that the volatility of photovoltaic power generation and electric vehicle load data can be modeled through the MDN prediction results.

[0049] At this time, in this embodiment, the joint dispatching model is solved to obtain the joint dispatching scheme of the virtual power plant, which specifically includes the following steps.

[0050] (1) Using the meteorological data of the scheduling period as input, the photovoltaic power generation data distribution of the scheduling period is predicted by using the photovoltaic power generation uncertainty prediction model.

[0051] (2) Using the electric vehicle charging data of the historical scheduling period as input, the electric vehicle charging data distribution of the scheduling period is predicted by using the electric vehicle charging uncertainty prediction model.

[0052] (3) Using the electric vehicle battery swapping data of the historical scheduling period as input, the electric vehicle battery swapping data distribution of the scheduling period is predicted by using the electric vehicle battery swapping uncertainty prediction model.

[0053] (4) Random sampling is respectively performed on the photovoltaic power generation data distribution, the electric vehicle charging data distribution, and the electric vehicle battery swapping data distribution to obtain the photovoltaic power generation, the electric vehicle charging data, and the electric vehicle battery swapping data of the scheduling period under multiple scenarios. The number of scenarios is the same as the number of random sampling times. The electric vehicle charging data includes the initial state of charge of each electric vehicle participating in charging and the charging window, and the charging window includes the start time and the end time , and the electric vehicle battery swapping data includes the initial state of charge of each electric vehicle participating in battery swapping and the battery swapping window, and the battery swapping window includes the start time and the end time, determining .

[0054] (5) Using the photovoltaic power generation, the electric vehicle charging data, and the electric vehicle battery swapping data of the scheduling period under multiple scenarios as input, the joint scheduling model is optimized and solved to obtain the joint scheduling scheme of the virtual power plant.

[0055] In this embodiment, the photovoltaic power generation uncertainty prediction model, the electric vehicle charging uncertainty prediction model, and the electric vehicle battery swapping uncertainty prediction model all adopt the mixture density neural network. The mixture density neural network adopts the negative log-likelihood loss function and the Adam (Adaptive Moment Estimation) optimizer during training.

[0056] Specifically, in this embodiment, data acquisition and cleaning are first performed. Since it involves the prediction tasks of photovoltaic power generation and electric vehicle load data, before deployment, the power generation data of the photovoltaic power generation equipment of the virtual power plant to be scheduled, the electric vehicle load data, and the corresponding meteorological data should be collected first, and cleaning steps such as missing value and outlier processing are performed. Optionally, all data can also be normalized to eliminate the scale differences of different data sources.

[0057] This embodiment then designs the mixture density neural network, which specifically includes the following steps.

[0058] (1) Design the neural network structure. First, the MDN should include an input layer that inputs the relevant features of photovoltaic power generation and the relevant features of electric vehicle load data. Second, the MDN should include a hidden layer, and the hidden layer may need to be designed in detail according to the data features. In this embodiment, a three-layer hidden layer is adopted, and the number of neurons in each layer decreases. There are 128 neurons in the first layer, 64 neurons in the second layer, and 32 neurons in the third layer. The ReLU (Rectified Linear Unit) is used as the activation function of the hidden layer, which has the advantages of high computational efficiency and avoiding the problem of gradient disappearance. The expression is as follows: ; Among them, is the input value.

[0059] The connection between its layers uses a standard fully connected layer (Dense Layer), that is, each neuron in the hidden layer is connected to all neurons in the previous layer. Finally, the MDN should include an output layer. The goal of the output layer of the MDN is to predict the mean, standard deviation, and weight of each Gaussian distribution component in the Gaussian Mixture Model (GMM). To output these parameters, the conditions that the standard deviation is positive, the weight is non-negative, and the sum of the weights is 1 should be ensured. Therefore, the Softplus and Softmax activation functions are used respectively, and the mean is directly output. The mathematical expressions are as follows: ; Among them, is the number of Gaussian distribution components, then the number of nodes in the output layer is , and the means , standard deviations , and weights of each Gaussian distribution component are output respectively; is the raw output value of the th node; is the raw output value of the th node; is the raw output value of the th node; is the raw output value of the th node.

[0060] Therefore, the mixture density neural network includes an input layer, a hidden layer, and an output layer connected in sequence. If you want to capture the time series relationship between data, a long short-term memory neural network layer or a bidirectional long short-term memory neural network layer can be added for time series memory. Specifically, it can be placed between the input layer and the hidden layer.

[0061] (2) Design the loss function. In this embodiment, the negative log-likelihood loss function is adopted. For each training sample, it is necessary to calculate its log-likelihood under the current mixture Gaussian model. The specific expression is as follows: ; where, is the loss; is the number of training samples; is the original output value of the -th node; is the true label value of the -th sample (the true value of photovoltaic power generation or the true value of electric vehicle load data); is the probability density formula of the -th Gaussian distribution component, and its expression is as follows: ; This loss function minimizes the difference between the true label value and the mixture Gaussian distribution predicted by the model, thereby optimizing the network parameters.

[0062] (3) Design the hyperparameter tuning algorithm. In this embodiment, the commonly used Adam optimizer is adopted for optimization.

[0063] In this embodiment, the designed MDN model is finally used to predict the distribution of photovoltaic power generation data and the distribution of electric vehicle load data (including the distribution of electric vehicle charging data and the distribution of electric vehicle battery swapping data) for each time period in the scheduling cycle.

[0064] In this embodiment, random sampling is respectively performed on the distribution of photovoltaic power generation data, the distribution of electric vehicle charging data, and the distribution of electric vehicle battery swapping data, specifically including: using the Monte Carlo sampling method to respectively perform random sampling on the distribution of photovoltaic power generation data, the distribution of electric vehicle charging data, and the distribution of electric vehicle battery swapping data.

[0065] Several random scenarios are obtained through Monte Carlo sampling, specifically including the following steps.

[0066] (1) Extract components from the Gaussian distribution components. Select one component from these components according to the weight of each Gaussian distribution component, which can be completed by means of random numbers.

[0067] Let the cumulative distribution function of the weights be: ; where, is the cumulative value of the -th Gaussian distribution component; is the weight of the -th Gaussian distribution component.

[0068] Generate a random number with a uniform distribution , select the -th Gaussian distribution component such that: ; where is the cumulative value of the -th Gaussian distribution component.

[0069] (2) Sample from the normal distribution of the selected component. Once the -th Gaussian distribution component is selected, draw a sample from the Gaussian distribution corresponding to this component.

[0070] It can be obtained by generating a standard normal random variable and using a linear transformation, i.e.: ; The obtained is the predicted value for the current period.

[0071] (3) Repeat the above process for each period in the scheduling cycle to obtain a sufficient number of samples to form a scenario matrix.

[0072] Assume the number of sampled scenarios is , then the obtained scenario matrix is: ; where is the scheduling cycle and also the window length in the stochastic rolling horizon convex optimization. In this embodiment, = 1h×24, the scheduling step is 1h, the scheduling cycle is 24 scheduling steps, that is, the scheduling cycle is 1 day, and each row of the obtained scenario matrix is a sampled scenario, and each column is the sampling result for that period.

[0073] Through the above sampling process, the distributions of photovoltaic power generation and electric vehicle load data can be converted into specific values of photovoltaic power generation and electric vehicle load data under different scenarios.

[0074] In this embodiment, taking the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data in the scheduling cycle under multiple scenarios as inputs, the joint scheduling model is solved to obtain the joint scheduling scheme of the virtual power plant, which specifically includes: using the clustering algorithm to cluster the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data in the scheduling cycle under multiple scenarios to obtain the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data in the scheduling cycle under multiple clusters; taking the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data in the scheduling cycle under multiple clusters as inputs, and solving the joint scheduling model to obtain the joint scheduling scheme of the virtual power plant.

[0075] In this embodiment, the sampled scenarios are clustered. If the uncertainty of the virtual power plant is too high, resulting in a relatively scattered sampling function distribution (i.e., distribution), it is necessary to increase the number of sampled scenarios and then use a clustering algorithm for scenario reduction. The available clustering algorithms include: K-means algorithm and Density-Based Spatial Clustering of Applications with Noise (abbreviated as DBSCAN), etc. In this embodiment, taking the K-means algorithm as an example, the clustering process includes the following steps.

[0076] (1) First, perform initialization. Select initial cluster centers (centroids), and these initial centroids can be obtained by randomly selecting data points in the scenario matrix or other heuristic methods (such as K-means++).

[0077] (2) Assign each data point to the nearest cluster center, that is, minimize the distance from the data point to the cluster center (usually using the Euclidean distance).

[0078] The expression for the distance is as follows: ; where is the distance between the th row vector in scenario and the centroid of the th cluster; is the total number of features, that is, the number of columns of ; is the th feature of scenario ; is the th feature of the centroid of the th cluster. Then assign the data point

[0079] to the nearest cluster, that is: ; where is the index of the cluster to which scenario belongs; is the centroid of the th cluster.

[0080]

[0080] (3) Update the centroid.

[0081] For each cluster, calculate the new centroid of the cluster. The new centroid is the mean of all scenarios in the cluster, that is: ; Among them, is the number of data points in the th cluster; represents all data points in the th cluster.

[0082] (4) Repeat (2) and (3) until the position of the centroid no longer changes or changes very little, and consider the algorithm to converge, obtaining clusters, corresponding to scenarios.

[0083] (5) Assign probabilities to each scenario.

[0084] The probability of each scenario (the centroid of the cluster) is proportional to the size of the cluster, and its expression is as follows: ; Among them, is the probability of the th cluster.

[0085] This embodiment can also predict the air-conditioning load. Taking the air-conditioning load in the historical scheduling period and the temperature in the current scheduling period as inputs, use the MDN to predict the air-conditioning load distribution in the current scheduling period, and determine the value of the air-conditioning load through sampling and clustering methods. Of course, the air-conditioning load can also be a known value.

[0086] In this embodiment, the joint scheduling model is solved to obtain the joint scheduling plan of the virtual power plant, specifically including: using the RHC (Rolling Horizon Control) algorithm to optimize and solve the joint scheduling model to obtain the joint scheduling plan of the virtual power plant.

[0087] Specifically, this embodiment uses a scheduling algorithm based on RHC for solving. To solve the virtual power plant scheduling problem, this embodiment adopts the RHC algorithm for solving. The RHC algorithm introduces a rollback mechanism in multiple time steps, combines different prediction scenarios of photovoltaic power generation, and performs dynamic optimization. At each time period, the RHC algorithm selects the optimal scheduling decision according to the current prediction information and real-time data, and continuously updates and adjusts the scheduling strategy in the future rolling optimization process. The RHC algorithm can handle multiple uncertainty factors (such as the uncertainty of photovoltaic power generation, electric vehicle charging and discharging, etc.), and performs robust optimization by considering the probability distribution of future scenarios. Through iterative solution, the RHC algorithm can provide a relatively stable and efficient scheduling plan under uncertain conditions, ensuring the maximization of the overall benefit of the virtual power plant.

[0088] This embodiment uses the RHC algorithm to solve the joint scheduling model, such as Figure 4 and Figure 5As shown in the figure, in order to effectively cope with the uncertainties of photovoltaic power generation, electric vehicle charging demand, and load curtailment, the RHC algorithm dynamically updates the scheduling plan within each rolling time step to ensure the maximization of the virtual power plant's profit. Specifically, the following steps will be used to solve the joint scheduling model: First, set the scheduling period and scheduling step size. In this embodiment, the scheduling period is 24 hours and the scheduling step size is 1 hour. Second, call the established MDN model to predict the distribution of photovoltaic power generation and electric vehicle load data in this time horizon. Then, call the proposed sampling method for scenario reduction to obtain typical scenarios and their distribution probabilities. Finally, call the solver to solve the established mixed-integer linear programming model. In this embodiment, the Cplex solver is specifically used for solving. After obtaining the optimal joint scheduling plan for dispatchable components such as energy storage and electric vehicles, execute the joint scheduling plan corresponding to one scheduling step size, and then repeat the above process.

[0089] Considering the obvious deficiencies in the source-load uncertainty modeling and joint scheduling plan design of related technologies, in order to improve the overall performance and reliability of virtual power plant scheduling, it is urgent to develop an optimal scheduling method that can comprehensively consider the synergistic effects of various resources and effectively cope with source-load uncertainties. This can not only improve the economic benefits of virtual power plants but also enhance the system's adaptability to external disturbances. Based on this, this embodiment proposes an optimal scheduling method for virtual power plants based on rolling horizon optimization and mixed density neural network prediction. The core technology uses RHC to optimize the scheduling decisions of virtual power plants with photovoltaic power generation equipment, energy storage equipment, air-conditioning loads, and electric vehicle charging and swapping stations, uses MDN to predict the uncertainty of photovoltaic power generation, and designs an incentive mechanism for electric vehicle charging and a swapping station model to achieve the joint scheduling of multiple energy resources and minimize the scheduling cost of virtual power plants. This optimization method based on RHC and MDN prediction can effectively handle the volatility and uncertainty of photovoltaic power generation, fully coordinate various resources such as photovoltaic power generation equipment, energy storage equipment, electric vehicle charging and swapping stations, and air-conditioning loads, and improve the operating efficiency and economy of virtual power plants.

[0090] Compared with the prior art, this embodiment has significant advantages in the construction and solution of the virtual power plant scheduling model. In particular, the innovations in uncertainty processing, multi-component joint scheduling, and incentive mechanism design have greatly improved the scheduling efficiency and resource optimization ability of virtual power plants.

[0091] In terms of the joint scheduling of multiple components, in this embodiment, multiple energy resources such as photovoltaic power generation equipment, energy storage equipment, electric vehicle charging and swapping stations, and air-conditioning loads are jointly scheduled. Traditional virtual power plant scheduling methods usually focus on the optimization of a single component, ignoring the synergy effects between resources, resulting in less than ideal scheduling results. However, in this embodiment, by introducing an incentive mechanism for electric vehicles, dynamic coordination among various components is achieved. Electric vehicles participate in load scheduling through the charging incentive mechanism, the energy storage equipment performs charge and discharge operations according to the uncertainty of photovoltaic power generation, and the air-conditioning load reduces power consumption behavior through the demand response mechanism. Through this scheduling method of multi-component coordination, not only can the power supply and demand be balanced, but also the overall economic benefits of the virtual power plant can be improved.

[0092] In terms of uncertainty handling, in this embodiment, the probability distribution of photovoltaic power generation is predicted by using MDN, effectively solving the problem of the uncertainty of photovoltaic power generation. Traditional photovoltaic power generation prediction usually can only give a single point estimate, while MDN can generate multiple Gaussian distribution parameters, reflecting the possible fluctuation range of photovoltaic power generation, which provides a more accurate input for subsequent scheduling decisions. In addition, this embodiment combines the RHC algorithm. During the solution process, it not only relies on the current prediction information but also fully considers the probability distributions of different future scenarios, making the scheduling plan more robust. The RHC algorithm dynamically optimizes the scheduling decision based on considering the volatility of photovoltaic power generation, the charge and discharge constraints of energy storage equipment, the charging demands of electric vehicles, and the air-conditioning load, and can effectively cope with various uncertainty factors to ensure the efficient operation of the virtual power plant.

[0093] In terms of incentive mechanism design, this embodiment proposes a flexible incentive mechanism aimed at encouraging flexible load resources such as electric vehicles to participate in the scheduling optimization of the virtual power plant. For electric vehicles, this embodiment designs a dynamic electricity price incentive scheme based on battery health status and charging demand, encouraging electric vehicles to reasonably arrange the charging time and charging amount according to the grid load conditions. This mechanism can not only avoid the pressure on the grid caused by concentrated electric vehicle charging, but also extend the battery service life and reduce the risk of battery degradation. Specifically, when electric vehicles participate in scheduling, according to their charging demands, battery status, and charging periods, the system will provide corresponding incentive electricity prices to ensure that the charging process not only meets the load demands of the grid but also takes into account the economic interests of electric vehicle users.

[0094] In summary, in this embodiment, by using MDN for photovoltaic power generation uncertainty prediction, the RHC algorithm for scheduling optimization, and jointly scheduling multiple energy components, the problems of uncertainty handling and resource optimal scheduling are effectively solved. In addition, by designing a reasonable incentive mechanism, this embodiment can promote the participation of flexible load resources such as electric vehicles and air-conditioning loads, further improving the scheduling efficiency and economy of the virtual power plant. These innovations make the application of this embodiment in the context of the virtual power plant have high practical value, and can effectively improve the robustness, flexibility, and overall benefits of virtual power plant scheduling.

[0095] The present application also provides an application scenario that applies the above virtual power plant joint scheduling method. Specifically, the virtual power plant joint scheduling method provided in this embodiment can be applied in a virtual power plant joint scheduling scenario. The virtual power plant joint scheduling scenario includes a scenario generation link and an operation link. The scenario generation link is used to establish and solve the joint scheduling model of the virtual power plant to obtain a joint scheduling plan, and the operation link is used to control the virtual power plant to operate according to the joint scheduling plan. The virtual power plant joint scheduling method provided in this embodiment belongs to the scenario generation link.

[0096] Embodiment 2.

[0097] Based on the same inventive concept, the embodiment of the present application also provides a virtual power plant joint scheduling device for implementing the above-mentioned virtual power plant joint scheduling method. The solution provided by this device to solve the problem is similar to the solution described in the above method. Therefore, the specific limitations in one or more of the following virtual power plant joint scheduling device embodiments can refer to the limitations on the virtual power plant joint scheduling method in the above text, and will not be repeated here.

[0098] In an exemplary embodiment, as Figure 6 shown, a virtual power plant joint scheduling device is provided, and the virtual power plant joint scheduling device includes the following modules.

[0099] A model construction module M1, configured to construct a joint scheduling model of the virtual power plant; the joint scheduling model is used to schedule the collaborative operation process among photovoltaic power generation equipment, energy storage equipment, electric vehicle charging stations, electric vehicle swapping stations, air-conditioning loads, and the power grid with the minimum scheduling cost as the optimization goal, and the constraint conditions include power balance constraints, energy storage equipment constraints, electric vehicle charging station constraints, electric vehicle swapping station constraints, air-conditioning load constraints, and power grid interaction power constraints.

[0100] The model solution module M2 is used to solve the joint scheduling model to obtain the joint scheduling plan of the virtual power plant; the joint scheduling plan includes the charging power and discharging power of the energy storage device, the charging power and discharging power of the electric vehicle, the charging power and discharging power of each battery in the electric vehicle swapping station, the reducible power of the air-conditioning load, and the interaction power between the virtual power plant and the power grid, where the interaction power is the power purchase power of the virtual power plant from the power grid or the power selling power of the virtual power plant to the power grid.

[0101] Embodiment 3.

[0102] In an exemplary embodiment, a computer device is provided. The computer device can be a server or a terminal, and its internal structure diagram can be as Figure 7 shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O), and a communication interface. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. 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, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals through a network connection. When the computer program is executed by the processor, it implements a virtual power plant joint scheduling method.

[0103] Those skilled in the art can understand that Figure 7 the structure shown in

[0104] is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0105] Embodiment 4.

[0106] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, it implements the virtual power plant joint scheduling method in Embodiment 1.

[0107] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0108] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0109] Specific examples are used in this article to elaborate on the principles and implementation manners of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, based on the idea of this application, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A virtual power plant combined dispatching method, characterized in that The virtual power plant joint dispatching method includes: Constructing a joint dispatching model for the virtual power plant; the joint dispatching model is used to optimize the objective of minimizing the dispatching cost, and under the constraints of the constraint conditions, schedule the coordinated operation process among photovoltaic power generation equipment, energy storage equipment, electric vehicle charging stations, electric vehicle battery swapping stations, air-conditioning loads and the power grid. The constraint conditions include power balance constraints, energy storage equipment constraints, electric vehicle charging station constraints, electric vehicle battery swapping station constraints, air-conditioning load constraints and power grid interaction power constraints; Solving the joint dispatching model to obtain a joint dispatching plan for the virtual power plant; the joint dispatching plan includes the charging power and discharging power of the energy storage equipment, the charging power and discharging power of the electric vehicles, the charging power and discharging power of each battery in the electric vehicle battery swapping station, the reducible power of the air-conditioning load, and the interaction power between the virtual power plant and the power grid. The interaction power is the power purchase power for the virtual power plant to purchase electricity from the power grid or the power selling power for the virtual power plant to sell electricity to the power grid.

2. The virtual power plant joint dispatching method according to claim 1, wherein The power balance constraint is: ； wherein, is the photovoltaic power generation of the photovoltaic power generation equipment in the time period; is the discharge power of the energy storage equipment in the is the number of electric vehicles participating in charging in the is the th electric vehicle's discharge power in the is the interaction power between the virtual power plant and the power grid in the time period; is the number of batteries in the electric vehicle swapping station; is the th battery's discharge power in the is the basic load demand of the virtual power plant in the is the charging power of the energy storage equipment in the is the th electric vehicle's charging power in the is the reducible power of the air-conditioning load in the is the th battery's charging power in the 3. The virtual power plant joint dispatching method according to claim 1, wherein Solving the joint dispatching model to obtain a joint dispatching plan for the virtual power plant, specifically including: Taking the meteorological data of the dispatching period as input, and using the photovoltaic power generation uncertainty prediction model to predict the photovoltaic power generation data distribution of the dispatching period; Taking the electric vehicle charging data of the historical dispatching period as input, and using the electric vehicle charging uncertainty prediction model to predict the electric vehicle charging data distribution of the dispatching period; Taking the electric vehicle battery swapping data of the historical dispatching period as input, and using the electric vehicle battery swapping uncertainty prediction model to predict the electric vehicle battery swapping data distribution of the dispatching period; Randomly sampling the photovoltaic power generation data distribution, the electric vehicle charging data distribution and the electric vehicle battery swapping data distribution respectively to obtain the photovoltaic power generation, electric vehicle charging data and electric vehicle battery swapping data of the dispatching period under multiple scenarios; the number of scenarios is the same as the number of random sampling times. The electric vehicle charging data includes the initial state of charge of each electric vehicle participating in charging and the charging window, and the electric vehicle battery swapping data includes the initial state of charge of each electric vehicle participating in battery swapping and the battery swapping window; Taking the photovoltaic power generation, electric vehicle charging data and electric vehicle battery swapping data of the dispatching period under multiple scenarios as input, and solving the joint dispatching model to obtain a joint dispatching plan for the virtual power plant.

4. The virtual power plant combined dispatching method according to claim 3, characterized in that The photovoltaic power generation uncertainty prediction model, the electric vehicle charging uncertainty prediction model and the electric vehicle battery swapping uncertainty prediction model all adopt a mixture density neural network, and the mixture density neural network uses a negative log-likelihood loss function and an Adam optimizer during training.

5. The virtual power plant combined dispatching method according to claim 3, wherein Randomly sampling the photovoltaic power generation data distribution, the electric vehicle charging data distribution and the electric vehicle battery swapping data distribution respectively, specifically including: using the Monte Carlo sampling method to randomly sample the photovoltaic power generation data distribution, the electric vehicle charging data distribution and the electric vehicle battery swapping data distribution respectively.

6. The virtual power plant combined dispatching method according to claim 3, wherein Taking the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data during the scheduling period in multiple scenarios as inputs, solving the joint scheduling model to obtain the joint scheduling plan of the virtual power plant, specifically including: Using a clustering algorithm to cluster the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data during the scheduling period in multiple scenarios to obtain the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data during the scheduling period under multiple clusters; Taking the photovoltaic power generation, electric vehicle charging data, and electric vehicle battery swapping data during the scheduling period under multiple clusters as inputs, solving the joint scheduling model to obtain the joint scheduling plan of the virtual power plant.

7. The virtual power plant combined dispatching method according to claim 6, characterized in that, Solving the joint scheduling model to obtain the joint scheduling plan of the virtual power plant, specifically including: using the RHC algorithm to optimize and solve the joint scheduling model to obtain the joint scheduling plan of the virtual power plant.

8. A virtual power plant combined dispatching device, characterized in that, The virtual power plant joint scheduling device includes: A model construction module for constructing a joint scheduling model of the virtual power plant; the joint scheduling model is used to optimize the scheduling cost as the optimization goal, and under the constraints of the constraint conditions, schedule the coordinated operation process among photovoltaic power generation equipment, energy storage equipment, electric vehicle charging stations, electric vehicle battery swapping stations, air-conditioning loads, and the power grid. The constraint conditions include power balance constraints, energy storage equipment constraints, electric vehicle charging station constraints, electric vehicle battery swapping station constraints, air-conditioning load constraints, and grid interaction power constraints; A model solving module for solving the joint scheduling model to obtain the joint scheduling plan of the virtual power plant; the joint scheduling plan includes the charging power and discharging power of the energy storage equipment, the charging power and discharging power of the electric vehicles, the charging power and discharging power of each battery in the electric vehicle battery swapping station, the reducible power of the air-conditioning load, and the interaction power between the virtual power plant and the power grid. The interaction power is the power purchase power of the virtual power plant from the power grid or the power selling power of the virtual power plant to the power grid.

9. A computer device, comprising: A memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the computer program to implement the virtual power plant joint scheduling method according to any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the virtual power plant joint scheduling method according to any one of claims 1-7.

Images