A virtual power plant peak shaving optimization scheduling method and system, electronic device and medium

Abstract

The application discloses a virtual power plant peak regulation optimization scheduling method and system, electronic equipment and medium, and relates to the field of virtual power plant peak regulation. The method comprises the following steps: a virtual energy storage model of distributed resources of a virtual power plant is established; the power feasible region of the decoupled and dimension-reduced virtual energy storage model of the distributed resources is determined by using the Minkowski sum method and the embedded hyperbox method; a double-objective scheduling model for peak regulation optimization scheduling is established based on the power feasible region of the decoupled and dimension-reduced virtual energy storage model of the distributed resources; the double-objective scheduling model for peak regulation optimization scheduling is solved by using the McCormick envelope method, and the optimal solution of the distributed resources is determined, so that the virtual power plant is optimized and scheduled according to the optimal solution. The virtual power plant is optimized and scheduled by using the distributed resources, and the flexibility of the virtual power plant peak regulation optimization scheduling is improved.

Description

Technical Field

[0001] This invention relates to the field of peak shaving in virtual power plants, and in particular to a method, system, electronic device, and medium for optimizing peak shaving scheduling in virtual power plants. Background Technology

[0002] The application of renewable energy (RES) in power systems is becoming increasingly widespread. However, its inherent uncertainties and variability pose significant challenges to the safe and efficient operation of power systems. To address the widening gap between peak and off-peak loads, it is necessary to improve the flexibility of peak-shaving resources to achieve supply-demand balance. Furthermore, hydrogen (H2) possesses characteristics such as long-term storage, large capacity, sustainability, and zero carbon emissions, making it a promising candidate for adapting to the high penetration rate of distributed renewable energy in power systems. Utilizing water electrolysis technology to convert renewable energy power generation into distributed hydrogen energy is a sustainable and low-cost hydrogen production pathway.

[0003] Virtual Power Plants (VPPs) are a business model for the integrated management of distributed energy resources. VPPs can reduce the impact of uncertainties in renewable energy generation on the power system by coordinating dispersed renewable energy sources and other distributed energy sources within the distribution network, and provide energy and ancillary services for the power system's operation. Meanwhile, with the successful application of hydrogen-powered electric vehicles, distributed hydrogen production and storage devices, and hydrogen fuel cells in transportation and distribution networks, electro-hydrogen interaction is becoming a new trend in the development of VPPs.

[0004] Currently, to address the peak-valley load differences caused by the uncertainty and variability of renewable energy generation, thermal power units with load-following capabilities are typically used as the primary peak-shaving resource. However, thermal power units cannot always operate at their ideal optimal operating point during deep peak shaving, resulting in low power generation efficiency, increased fuel costs, and increased carbon emissions. Furthermore, due to limitations in operational economics and thermal efficiency, the regulation capacity of thermal power units is usually limited, especially under operating conditions with large load fluctuations, where their regulation capacity struggles to meet the operational requirements of the power system. The long investment payback period and high production costs of hydrogen energy projects constrain the development of distributed hydrogen energy projects. During off-peak periods of hydrogen load, the superior energy storage capacity of these hydrogen power plants is not fully utilized, and their aggregation flexibility for peak-shaving services is often overlooked by VPPs (Virtual Power Plants).

[0005] Peak-shaving power markets (PRMs) typically have minimum bid requirements, which small-scale distributed energy sources often struggle to meet. Virtual power plants (VPPs) can aggregate large amounts of distributed energy to meet market demands. The flexibility of VPPs can be modeled as equivalent virtual energy storage or virtual synchronous machines, with virtual parameters including charge / discharge power constraints, energy constraints, slope, and self-discharge rate. However, these studies primarily focus on the power flexibility of VPPs, neglecting the impact of distributed hydrogen resources and failing to address the flexibility assessment of electric-hydrogen interactive virtual power plants (EH-VPPs). Summary of the Invention

[0006] The purpose of this invention is to provide a method, system, electronic device and medium for peak shaving and optimization scheduling of virtual power plants, so as to improve the flexibility of peak shaving and optimization scheduling of virtual power plants using distributed resources.

[0007] To achieve the above objectives, the present invention provides the following solution:

[0008] A virtual power plant peak-shaving optimization scheduling method includes:

[0009] A virtual energy storage model for distributed resources of a virtual power plant is established; the virtual energy storage model for distributed resources includes a virtual energy storage model for electric vehicles, a virtual energy storage model for air conditioning systems, and a virtual energy storage model for hydrogen refueling stations powered by renewable energy.

[0010] Using the Minkowski method and the embedded superbox method, the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource is determined; wherein, the embedded superbox method is to find the largest inscribed cuboid in the high-dimensional aggregated polygon, and use the largest inscribed cuboid to represent the high-dimensional polyhedron.

[0011] Based on the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resources, a dual-objective scheduling model for peak shaving optimization scheduling is established.

[0012] The McCormick envelope method is used to solve the dual-objective scheduling model for peak shaving optimization, and the optimal solution for the distributed resources is determined. The virtual power plant is then optimized and scheduled based on the optimal solution, whereby the optimal solution is the power value of the distributed resources.

[0013] Optionally, the virtual energy storage model for hydrogen refueling stations powered by renewable energy is... in, Let t be the amount of hydrogen stored in the hydrogen storage tank at time t; Let t be the amount of hydrogen stored in the hydrogen storage tank at time t-1; The equivalent external power characteristics of a hydrogen refueling station powered by renewable energy at time t under a set confidence level α; η P2H The hydrogen production conversion efficiency of the electro-hydrogen conversion device; K is the calorific value of hydrogen. HS,min K represents the minimum hydrogen storage capacity of the hydrogen storage tank. HS,max This represents the maximum hydrogen storage capacity of the hydrogen storage tank. This is the lower bound of the predictive power of PVs at time t under a set confidence level α; The upper limit of the predictive power of PVs at time t under a set confidence level α; This represents the lower limit of the predictive capability of hydrogen fuel cell vehicles at time t under a set confidence level α. The upper limit of the predictive ability of hydrogen fuel cell vehicles at time t under a set confidence level α; This represents the average predictive power of PVs at time t. Let be the standard deviation of the predictive power of PVs at time t; This represents the average predictive ability of hydrogen fuel cell vehicles at time t. Let be the standard deviation of the predictive power of hydrogen fuel cell vehicles at time t; α is the set confidence level.

[0014] Optionally, the virtual energy storage model for electric vehicles is... in, The virtual charging state of the electric vehicle at time t; The virtual charging state of the electric vehicle at time t-1; η represents the equivalent charge / discharge power of the electric vehicle at time t. EV For the charging and discharging efficiency of electric vehicles; The upper limit of the virtual charging state of the electric vehicle at time t; The virtual charging state limit for the electric vehicle at time t; This represents the upper limit of the equivalent charging and discharging power of the electric vehicle at time t. This represents the lower limit of the equivalent charge / discharge power of the electric vehicle at time t. Let N be the grid connection state of electric vehicle v in the i-th cluster at time t; i Let i be the number of electric vehicles in cluster i; Let be the maximum virtual charging state of electric vehicle v in the i-th cluster at time t; Let be the minimum virtual charging state of electric vehicle v in the i-th cluster at time t; Let be the maximum equivalent charging and discharging power of electric vehicle v in the i-th cluster at time t; Let be the minimum equivalent charging and discharging power of electric vehicle v in the i-th cluster at time t.

[0015] Optionally, the virtual energy storage model of the air conditioning system is

[0016] N represents the discharge power of the air conditioner at time t; J represents the number of air conditioning system groups; N represents the discharge power of the air conditioner at time t. jLet j be the number of rooms in the j-th group of air conditioning systems; To determine the on / off state of the air conditioner in the a-th room of the j-th air conditioning system at time t, This indicates that the air conditioner is on. This indicates that the air conditioner is off; Let be the discharge power of the air conditioner in the a-th room of the j-th air conditioning system at time t; Let t be the charging power of the air conditioner at time t; Let be the charging power of the air conditioner in the a-th room of the j-th air conditioning system at time t; Let be the power of the air conditioner at time t; Let be the maximum steady-state charging power of the air conditioner in the a-th room of the j-th air conditioning system at time t. Let be the outdoor temperature at time t; Let be the equivalent thermal resistance of the a-th room in the j-th air conditioning system; Let be the cooling capacity of the a-th room in the j-th air conditioning system; Let be the energy efficiency ratio of the air conditioner in the a-th room of the j-th air conditioning system; Let T be the power of the air conditioner in the a-th room of the j-th air conditioning system at time t, when the temperature is most comfortable for the human body. set ,min The lower limit of the human body's comfortable temperature; T comf The most comfortable temperature for the human body; Let be the maximum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; Let be the minimum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; The duration of the air conditioner's maximum cooling power operation during the charging cycle; Δt represents the minimum cooling power operating time of the air conditioner during the charging cycle; Δt is the charging / discharging cycle. Let be the maximum steady-state discharge power of the air conditioner in the a-th room of the j-th air conditioning system at time t. This refers to the operating time of the air conditioner at its maximum cooling power during the discharge cycle. This refers to the operating time of the air conditioner at its minimum cooling power during the discharge cycle.

[0017] Optionally, the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource is determined using the Minkowski sum method and the embedded superbox method, specifically including:

[0018] Calculate the time-coupled power feasible region of the virtual energy storage model of the distributed resources;

[0019] Based on the time-coupled power feasible region, the aggregated high-dimensional feasible region is determined using the Minkowski sum method.

[0020] The high-dimensional feasible region of the aggregate is reduced and decoupled using the embedded superbox method to obtain the decoupled and reduced power feasible region.

[0021] Optionally, the peak-shaving optimization scheduling dual-objective scheduling model is: in, is the discrete expression for the conditional risk value; b is the threshold of the loss function; L is the number of segments for discretizing the random variable; α is the set confidence level; The upward regulation capability of a virtual power plant with electro-hydrogen interaction after participating in the peak-shaving electricity market at time t; The total operating power of the virtual power plant with pre-dispatch hydrogen-electric interaction at a set confidence level α; The maximum estimated power of the virtual power plant at time t under confidence level α; The minimum estimated power of the virtual power plant at time t under confidence level α; The downward regulation capability of a virtual power plant participating in the peak-shaving electricity market and its interaction with hydrogen at time t; z l =[f(x,y l )-b] + ,(l=1,2,…,L);f(x,y l ) is the discretized loss function.

[0022] A virtual power plant peak-shaving optimization scheduling system includes:

[0023] The energy storage model building module is used to build a virtual energy storage model for the distributed resources of the virtual power plant; the virtual energy storage model for the distributed resources includes a virtual energy storage model for electric vehicles, a virtual energy storage model for air conditioning systems, and a virtual energy storage model for hydrogen refueling stations powered by renewable energy.

[0024] The feasible region determination module is used to determine the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource using the Minkowski sum method and the embedded superbox method; wherein, the embedded superbox method is to find the largest inscribed cuboid in the high-dimensional aggregated polygon and use the largest inscribed cuboid to represent the high-dimensional polyhedron.

[0025] The scheduling model establishment module is used to establish a peak-shaving optimization scheduling dual-objective scheduling model based on the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resources.

[0026] The scheduling module is used to solve the peak-shaving optimization scheduling bi-objective scheduling model using the McCormick envelope method to determine the optimal solution for the distributed resources, so as to optimize the scheduling of the virtual power plant based on the optimal solution; the optimal solution is the power value of the distributed resources.

[0027] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor runs the computer program to enable the electronic device to perform the virtual power plant peak shaving optimization scheduling method described above.

[0028] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described virtual power plant peak-shaving optimization scheduling method.

[0029] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0030] This invention provides a virtual power plant peak-shaving optimization scheduling method, system, electronic equipment, and medium. It establishes a virtual energy storage model of distributed resources in a virtual power plant; utilizes the Minkowski sum method and the embedded superbox method to determine the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model; based on the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model, it establishes a dual-objective scheduling model for peak-shaving optimization; and uses the McCormick envelope method to solve the dual-objective scheduling model for peak-shaving optimization to determine the optimal solution for the distributed resources, thereby optimizing the scheduling of the virtual power plant according to the optimal solution. This invention provides a compact and concise scheduling region by establishing a virtual energy storage model of distributed resources in a virtual power plant and using the Minkowski sum and embedded superbox methods to reduce and decouple the virtual energy storage model of distributed resources. Furthermore, utilizing distributed resources for peak-shaving optimization scheduling of the virtual power plant improves the flexibility of peak-shaving optimization scheduling of the virtual power plant. Attached Figure Description

[0031] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0032] Figure 1 The flowchart of the virtual power plant peak-shaving optimization scheduling method provided by the present invention is shown below.

[0033] Figure 2 A graph showing the power variation of the AC load during a charge-discharge cycle;

[0034] Figure 3 A diagram illustrating Minkowski's rules;

[0035] Figure 4 This is an approximate schematic diagram of the embedded superbox of the polymerization power polyhedron;

[0036] Figure 5 The graph shows the predictive power and power output fluctuation range of photovoltaic PVs at time t for R-HRSs with a confidence level of α of 90%.

[0037] Figure 6 The predictive power and hydrogen load fluctuation range curves of hydrogen fuel cell vehicle HVs at time t for R-HRSs with a confidence level α of 90%.

[0038] Figure 7 The graphs show the polymerization power curves for schemes 1, 2, and 3 at confidence levels α of 90% and 50%, respectively.

[0039] Figure 8 The graph shows the random aggregation power curves of Scheme 1 at different confidence levels. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] The purpose of this invention is to provide a method, system, electronic device and medium for peak shaving and optimization scheduling of virtual power plants, so as to improve the flexibility of peak shaving and optimization scheduling of virtual power plants using distributed resources.

[0042] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] Example 1

[0044] This invention models various distributed resources, such as electric vehicles (EVs), air conditioning systems (ACs), and renewable energy-powered hydrogen refueling stations (R-HRSs), as virtual energy storage (VS) models with charge and discharge response parameter characteristics, and uses spectral clustering to adjust the aggregation parameters in the model. In addition, the time-coupled power curve of VS can be abstracted into a polyhedron from a geometric perspective. The aggregation high-dimensional flexible region of VPP is composed of the Minkowski sum of these polyhedra, and the dimension of the aggregation polyhedron is further reduced by using the embedded superbox method. For electric-hydrogen integrated virtual power plants participating in the peak-shaving electricity market, a day-ahead peak-shaving optimization scheduling model (peak-shaving optimization scheduling bi-objective scheduling model) is established, which considers the dimensionality reduction prediction of the flexible region and conditional risk value. The dimension of the aggregation polygon is reduced by the embedded superbox approximation, providing a compact and concise scheduling region.

[0045] like Figure 1As shown, the virtual power plant peak-shaving optimization scheduling method provided by the present invention includes:

[0046] Step 101: Establish a virtual energy storage model for the distributed resources of the virtual power plant; the virtual energy storage model for the distributed resources includes a virtual energy storage model for electric vehicles, a virtual energy storage model for air conditioning systems, and a virtual energy storage model for hydrogen refueling stations powered by renewable energy.

[0047] Considering the response characteristics of distributed resources (DERs) and the interaction between electricity and hydrogen, heterogeneous distributed resources, such as electric vehicles, air conditioning systems, and hydrogen refueling stations (hydrogen refueling stations powered by renewable energy), are uniformly modeled as a first-order virtual energy storage (VS) model with charge and discharge response characteristics to capture the flexibility of heterogeneous DERs. Combined with spectral clustering, the aggregation parameters of heterogeneous distributed resources are obtained.

[0048] Step 101 is as follows:

[0049] Step 1011: Establish the VS model. DERs are small in capacity, numerous, and heterogeneous. To simplify the operation of these DERs under nonlinear and high-order constraints in EH-VPP scheduling, a reduced-order VS model is introduced. This model considers both the internal parameters of the DERs (such as state of charge (SOC), impedance, and energy usage habits) and external parameters (such as temperature and radiation intensity). The VS model is as follows:

[0050]

[0051] in, and The virtual memory (SOVC) states at time t and t-1 are respectively. It is the charging and discharging power; and Indicates the lower and upper bounds of the charging and discharging power; and This reflects the lower and upper bounds of the virtual storage state; η VS It refers to charge / discharge efficiency; and It is determined by the individual response characteristics of each DER.

[0052] 1) Determine the VS model for renewable energy-powered hydrogen refueling stations (R-HRSs). Resources within a renewable energy-powered hydrogen refueling station (R-HRS) comprise four categories: hydrogen storage units, hydrogen supply units, hydrogen demand units, and renewable energy generation units. All electricity generated by photovoltaic (PV) power generation is consumed by the hydrogen-to-electricity (P2H) conversion unit. The hydrogen produced by the P2H (hydrogen supply unit) is first supplied to hydrogen fuel cell vehicles (HVs) (hydrogen demand units), with the remainder stored in hydrogen storage tanks (HSs). When the hydrogen supply from the P2H is insufficient to meet the needs of the HVs, the HSs will replace the P2H as the hydrogen supply unit until the hydrogen storage is depleted. After this, the refueling station must purchase electricity from the local grid to meet the needs of the HVs.

[0053] Based on the operating mode of hydrogen refueling stations, the process of transferring and converting hydrogen electricity within renewable energy-powered hydrogen refueling stations (R-HRSs) can be described as follows:

[0054]

[0055] P P2H,min ≤P t P2H ≤P P2H,max .

[0056] P t P2H =P t PV +P t M .

[0057] H t P2H =H t HV +H t HS .

[0058]

[0059] K t HS,min ≤K t HS ≤K t HS,max .

[0060] It is the electrical power consumed by the P2Hs electro-hydrogen conversion device; It refers to the hydrogen flow rate at the input and output of the P2Hs electro-hydrogen conversion device; Let t be the amount of hydrogen stored in the hydrogen storage tank at time t; Let t be the amount of hydrogen stored in the hydrogen storage tank at time t-1; The hydrogen flow rate is the input and output of the hydrogen storage tank HSs; This is the maximum input power of the P2Hs electro-hydrogen conversion device; It is the minimum input power of the P2Hs electro-hydrogen conversion device; This indicates the minimum hydrogen storage capacity of the hydrogen storage tank HSs; Indicates the maximum hydrogen storage capacity of the hydrogen storage tank HSs; η P2H The hydrogen production conversion efficiency of the electro-hydrogen conversion device; This refers to the calorific value of hydrogen. This represents the hydrogen exchange demand of HV at time t; The electricity was purchased from the power grid.

[0061] The temporal coupling of energy storage in R-HRSs can be restated as follows:

[0062]

[0063] in, and These are the predictive capabilities of photovoltaic (PV) and hydrogen fuel cell vehicle (HV) at time t.

[0064] Furthermore, due to the uncertainties in PV power generation and HV demand, R-HRSs should comply with the following operational constraints:

[0065]

[0066] in, This represents the equivalent external power characteristic of R-HRSs at time t; and The minimum and maximum power of R-HRSs at a set confidence level α; This represents the lower limit of the predictive power of photovoltaic PVs at time t under a set confidence level α; This represents the upper limit of the predictive power of photovoltaic PVs at time t under a set confidence level α. This represents the lower limit of the predictive capability of hydrogen fuel cell vehicles at time t under a set confidence level α. This represents the upper limit of the predictive capability of hydrogen fuel cell vehicles at time t under a set confidence level α.

[0067] Due to various anthropogenic and natural uncertainties, the actual values ​​of hydrogen consumption and photovoltaic power generation fluctuate around the predicted values ​​and can be treated as random variables. Using a conversion method between optimistic (sup) and pessimistic (inf) values, the output / input limits of PV and HV at a set confidence level α are further determined.

[0068]

[0069] Among them, (Φ PV ) -1 and (ΦHV ) -1 Representing random variables respectively and The inverse cumulative distribution function (ICDF).

[0070] The amount of hydrogen consumed by HVs and PVs within each time interval typically follows a normal distribution. For any normally distributed N(e,σ), its ICDF can be expressed as:

[0071]

[0072] The upper and lower bounds of a random variable at a given confidence level α, i.e., the confidence interval, can be obtained by the following formula:

[0073]

[0074] in, This represents the average predictive power of PVs at time t. Let be the standard deviation of the predictive power of PVs at time t; This represents the average predictive ability of hydrogen fuel cell vehicles at time t. Let be the standard deviation of the predictive ability of hydrogen fuel cell vehicles at time t; α is the set confidence level, which is in the range of [0.5,1) to ensure the feasibility of the confidence interval.

[0075] In summary, the VS model of R-HRS can be described as follows:

[0076]

[0077] 2) Determine the VS model for the electric vehicle. Assuming the electric vehicle starts charging when connected to the grid, the upper limit of the virtual state of charge (SOVC) at time t can be calculated; under the condition of meeting the minimum charging amount when the vehicle is disconnected from the grid, the lower limit of the virtual state of charge at time t can be calculated. Through analysis, the VS model for the electric vehicle is established as follows:

[0078]

[0079] Where the subscript i represents the number of clusters; Let be the maximum virtual charging state of electric vehicle v in the i-th cluster at time t; Let be the minimum virtual charging state of electric vehicle v in the i-th cluster at time t; Indicates the maximum energy storage allowed by the technology; This indicates the initial energy storage when EVs are disconnected from the grid; Indicates the minimum energy storage when EVs are off-grid; η EV This refers to the charge and discharge efficiency of EVs; The maximum charging power of electric vehicle v in cluster i; Let V be the off-grid time of electric vehicle v in cluster i; Let be the grid connection time of electric vehicle v in cluster i. Electric vehicles are clustered using the Ng-Jordan-Weiss (NJW) spectral clustering algorithm based on their demand characteristics.

[0080] Using a bottom-up aggregation method, the equivalent aggregation parameters for electric vehicle VS were obtained based on the unique driving modes and characteristics of electric vehicles in different clusters, as follows:

[0081]

[0082]

[0083]

[0084]

[0085] in, The upper limit of the virtual charging state of the electric vehicle at time t; The virtual charging state limit for the electric vehicle at time t; This represents the upper limit of the equivalent charging and discharging power of the electric vehicle at time t. This represents the lower limit of the equivalent charge / discharge power of the electric vehicle at time t. Let be the maximum equivalent charging and discharging power of electric vehicle v in the i-th cluster at time t; Let be the minimum equivalent charging and discharging power of electric vehicle v in the i-th cluster at time t. N represents the grid connection state of electric vehicle v in cluster i at time t, where i is the cluster number of EVs; i It represents the number of electric vehicles in cluster i.

[0086] In summary, the VS model for electric vehicles is as follows:

[0087]

[0088] in, and Let SOVC and equivalent charge / discharge power be the electric vehicle at time t.

[0089] 3) Determine the VS model of the air conditioning system. Considering the heat transfer between indoors and outdoors, the classic lumped equivalent thermal parameter (LETP) modeling method is used to simulate the heat transfer process.

[0090]

[0091] in, and These represent the cooling capacity and operating power of the air conditioner in room a of group j at time t; Let be the indoor temperature of room a in group j at time t; Let be the outdoor temperature at time t; and These are the equivalent thermal resistance and cooling capacity of room a in group j, respectively. Let be the energy efficiency ratio of the air conditioner in room a of group j. Meanwhile, based on thermodynamic parameters... and and the operating parameters of the air conditioner (including maximum cooling capacity) and energy efficiency ratio The NJW spectrum clustering algorithm was used to cluster air-conditioned rooms.

[0092] When the air conditioner in room a is in a steady state, Balanced with the heat coming in from the outside, therefore and It remains unchanged. Therefore, the steady-state operating power of the air conditioner can be calculated as follows:

[0093]

[0094] Figure 2 The power variation curve of the AC load during the charge-discharge cycle is shown in the figure. Figure 2 As shown, the time-varying characteristics of the air conditioner's operating power under different set temperatures are illustrated.

[0095] Taking a charging cycle as an example, during the transient state of charging, the indoor ACs operate at maximum cooling power until a set lower limit is reached. Then, the room temperature and the ACs' power... The following stable state is achieved:

[0096]

[0097] Let be the maximum steady-state charging power of the air conditioner in the a-th room of the j-th air conditioning system at time t. Furthermore, to ensure user comfort, the AC operates at its minimum cooling power at the end of the charging period. Run until Drop to the most comfortable temperature T comf Corresponding cooling power It can be represented as:

[0098]

[0099] Let be the power of the air conditioner in the a-th room of the j-th air conditioning system at time t, at the temperature most comfortable for the human body; consider the charging power of the equivalent AC as the average cooling power during the charging period, when the room temperature is set to the lower limit of the human comfort temperature T. set,min At that time, the maximum charging power can be obtained. as follows:

[0100]

[0101]

[0102]

[0103] in, Let be the maximum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; Let be the minimum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; The duration of the air conditioner's maximum cooling power operation during the charging cycle; Δt represents the minimum cooling power operating time of the air conditioner during the charging cycle; Δt represents the charging / discharging cycle.

[0104] Similarly, when the room temperature is set to T set,max At that time, the maximum discharge power at time a can be obtained.

[0105]

[0106] in, Let be the maximum steady-state discharge power of the air conditioner in the a-th room of the j-th air conditioning system at time t. This refers to the operating time of the air conditioner at its maximum cooling power during the discharge cycle. This refers to the operating time of the air conditioner at its minimum cooling power during the discharge cycle.

[0107] Therefore, the VS model of ACs is expressed as follows:

[0108]

[0109] Let be a two-dimensional variable, representing the on / off state of the type j air conditioner in the a-th room at time t. This indicates that the air conditioner is on. This indicates that the air conditioner is off.

[0110] A2. VS Parameter Determination. Based on the VS model established in A1, and considering the characteristics of hydrogen refueling stations powered by renewable energy, electric vehicles, and air conditioning, the VS parameters for each distributed energy storage (DER) are determined. The VS parameters include the minimum virtual energy storage state at time t. and maximum virtual energy storage state Minimum virtual charging / discharging power at time t and maximum virtual charging / discharging power In step 102 below, the power feasible region will be calculated using the VS model with the VS parameters substituted.

[0111] Step 102: Using the Minkowski method and the inscribed hyperbox method, determine the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource. The inscribed hyperbox method involves finding the largest inscribed cuboid within the high-dimensional aggregated polygon, using this largest inscribed cuboid to represent the high-dimensional polyhedron, thereby achieving dimensionality reduction.

[0112] Using the first-order VS model proposed in step A, the feasible region of the VS with time-coupled power characteristics is calculated, and the feasible region is geometrically abstracted as a polyhedron. The feasible region considering distributed resource DERs is plotted using the Minkowski sum method to create a high-dimensional polyhedron of total adjustable power. Then, the dimensionality of the aggregated polyhedron is approximately reduced using the embedded superbox method, providing a compact and concise scheduling region, thus decoupling and reducing the dimensionality of the power feasibility region of time-coupled DERs.

[0113] Step 102 specifically includes:

[0114] Calculate the time-coupled power feasible region of the virtual energy storage model of the distributed resources.

[0115] Based on the time-coupled power feasible region, the aggregated high-dimensional feasible region is determined using the Minkowski sum method.

[0116] The high-dimensional feasible region of the aggregate is reduced and decoupled using the embedded superbox method to obtain the decoupled and reduced power feasible region.

[0117] In practical applications, the adjustable power of distributed resources exhibits time coupling and uncertainty. Therefore, under a certain confidence level, the adjustable power range is visualized as a convex polyhedron with power at each time period as the coordinate axis. Since the adjustable power of distributed resources is additive, the Minkowski sum (symbol ) is introduced. This is used to calculate the aggregated adjustable power range (aggregated high-dimensional feasible region) of the virtual power plant. The power feasibility curves of the individual DERs mentioned above are mapped to the aggregated power curves:

[0118]

[0119] Where D is the number of DER types in set Δ.

[0120] Taking the Minkowski sums of two types of DER in two consecutive time intervals t and t+1 as examples, like Figure 3 As shown, p t and p t+1 Polymerization power can be expressed as:

[0121]

[0122] p 1,t p 1,t+1 p 2,t and p 2,t+1 These represent the operating power of the two DERs during the two time periods t and t+1, respectively.

[0123] In the VS model, energy dynamics exist, thus the power feasibility curve for each DER is time-coupled. With increasing time intervals and distributed resources, the solution dimension for the aggregated adjustable power range increases dramatically, causing the corresponding Minkowski sum to become an NP-hard problem that cannot be directly solved. Therefore, an approximate solution for the aggregated adjustable power range is provided using the embedded superbox method. Considering the demand for flexibility in the peak-shaving electricity market, such as... Figure 4 As shown, an approximate flexibility evaluation is performed using an embedded superbox. By solving for the largest embedded high-dimensional rectangle in the irregularly aggregated adjustable power range polyhedron, the maximum feasible power range of EH-VPP in each time interval is determined.

[0124]

[0125] and These represent the side lengths of the embedded superboxes. and These represent the minimum and maximum computational power at time t, respectively, under confidence level α.

[0126] The embedded superbox has independent upper and lower power limits at different time periods, achieving time decoupling of the aggregated adjustable power range. Specifically, the power range for each time period... It is assembled into a box-shaped feasible region, namely the approximately random aggregation power flexibility of EH-VPP.

[0127]

[0128] Step 103: Based on the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resources, establish a dual-objective scheduling model for peak shaving optimization.

[0129] Based on the power feasible region of DERs after decoupling and dimensionality reduction proposed in step 102, a dual-objective scheduling model for peak-shaving optimization scheduling considering the predictive flexibility region and conditional risk value is established for EH-VPPs participating in the peak-shaving power market. This model maximizes the expected economic benefits of the peak-shaving market and reduces the revenue risk of EH-VPPs.

[0130] Step 103 is as follows:

[0131] First, the objective function is determined. The uncertainty of renewable energy output and load energy demand introduces stochastic aggregated power regions, leading to risk-benefit dynamics (PRM) in EH-VPP. EH-VPP aims to maximize benefits while reducing risk; therefore, the optimal peak-shaving scheduling model of EH-VPP can be expressed as a risk-conditional (CVaR) optimization model based on multi-objective optimization theory. The loss function f(x,y) of the CVaR optimization model is defined as follows:

[0132]

[0133]

[0134]

[0135]

[0136] Among them, C t and R t These are the costs and benefits of peak-shaving scheduling, respectively. and The output / input power of PVs, ACs, EVs, and HVs at time t are respectively at confidence level α; The total operating power of EH-VPP at confidence level α before scheduling; It is a decision vector and and This indicates the downward and upward regulatory capacity of EH-VPP at time t after participation in PRM; It is a random vector that takes into account the uncertainty of input and output; and EH-VPP provides unit scheduling costs and benefits for peak shaving and valley filling.

[0137] Given a CVaR value V for x at confidence level α. CVaR,α (x) is defined as the expected value b when the loss function exceeds the pessimistic value. α (x):

[0138]

[0139] In the formula, y represents the probability density function (PDF).

[0140] Because of b α The analytical expression for (x) is difficult to derive, therefore F is introduced. α The function (x,b) yields the value of CVaR, which is then discretized. As shown below:

[0141]

[0142]

[0143] Where, [f(x,y)-b] + Defined as max{f(x,y)-b,0}, y l (l=1,2,…,L) are the sample values ​​obtained in a given step within the confidence interval.

[0144] Then determine the constraints.

[0145] Let dummy variable z k (k = 1, 2, ..., q), let z k =[f(x,y k )-b] + Then z k The constraints are as follows:

[0146]

[0147] The constraints for the flexible aggregated power down-shaving model are:

[0148]

[0149]

[0150] In summary, the peak-shaving market optimization scheduling problem that minimizes CVaR is transformed into a linear programming problem (peak-shaving optimization scheduling bi-objective scheduling model):

[0151]

[0152] in, is the discrete expression for the conditional risk value; b is the threshold of the loss function; L is the number of segments for discretizing the random variable; α is the set confidence level; The upward regulation capability of a virtual power plant with electro-hydrogen interaction after participating in the peak-shaving electricity market at time t; The total operating power of the virtual power plant with pre-dispatch hydrogen-electric interaction at a set confidence level α; The maximum estimated power of the virtual power plant at time t under confidence level α; The minimum estimated power of the virtual power plant at time t under confidence level α; The downward regulation capability of a virtual power plant participating in the peak-shaving electricity market and its interaction with hydrogen at time t; z l =[f(x,y l )-b] + ,(l=1,2,…,L);f(x,y l ) is the discretized loss function.

[0153] Step 104: Solve the peak-shaving optimization scheduling bi-objective scheduling model using the McCormick envelope method to determine the optimal solution for the distributed resources, and then optimize the scheduling of the virtual power plant based on the optimal solution. The optimal solution is any feasible solution that makes the objective function reach its optimal value (maximum or minimum value). In this invention, it refers to the power values ​​of various distributed resources when the objective function of maximizing revenue while minimizing risk is achieved.

[0154] The peak-shaving scheduling model in step C is discretized and transformed into a convex optimization problem. To address the nonlinearity caused by the bilinear product terms of bid quantity and price in the equation, the McCormick envelope method is used to relax the bilinear product terms in the continuous space to prevent local optima and ensure that the solution is the global optimum.

[0155] To address the nonlinearity issue caused by the bilinear product term in the dual-objective scheduling model for peak shaving optimization, the McCormick envelope method is used to relax the bilinear product term in continuous space. The key to relaxation using the McCormick envelope method is finding the upper and lower bounds of each bilinear product term; in this embodiment, this mainly involves the boundaries of unit cost and unit revenue.

[0156] In terms of unit cost limits, balancing the benefits of EH-VPP and DER is crucial. On one hand, considering DER profits, the unit cost of EH-VPP should be greater than the average operating cost of DER. On the other hand, the unit cost of EH-VPP should not exceed the lowest bid price for EH-VPP in PRM.

[0157]

[0158] In summary, the relaxed linear form of the equation is as follows:

[0159]

[0160] To analyze the performance of the method of this invention in processing a large number of heterogeneous DERs, unique operating parameters for different DERs were randomly generated using the Latin hypercube sampling method. It is assumed that the operating parameters of R-HRSs, EVs, and ACs follow the probability distribution functions in Tables 1, 2, and 3 (nd represents a normal distribution).

[0161] Table 1. Probability distribution of R-HRSs parameters

[0162]

[0163] Table 2 Probability distribution of EVs parameters

[0164]

[0165] Table 3. Probability distribution of ACs parameters

[0166] <![CDATA[R AC / (℃ / kW)]]> nd 7.12 1.00 <![CDATA[C AC / (kJ / ℃)]]> nd 690.87 1.00 <![CDATA[λ AC ]]> nd 3.40 1.00 <![CDATA[P AC,max / kW]]> nd 2.80 1.00 Opening time / hour nd 12:00 3.00 Closure time / hour nd 21:00 3.00

[0167] Figure 5 The graph shows the predictive power of photovoltaic PVs and the fluctuation range of photovoltaic output at time t for R-HRSs with a confidence level α of 90%. Figure 6 The graph shows the predictive power of hydrogen fuel cell vehicle HVs and the range of hydrogen load fluctuations at time t for R-HRSs with a confidence level α of 90%.

[0168] Numerical analysis was conducted to examine the effects of incorporating electro-hydrogen interactions and the NJW spectral clustering process. Figure 7 A comparison of multi-time period clustering results under different conditions is presented.

[0169] Option 1: EH-VPP, which includes cluster electric vehicles, cluster air-conditioned rooms, and R-HRSs for electro-hydrogen interactions.

[0170] Option 2: EH-VPP, which includes cluster electric vehicles, cluster air-conditioned rooms, and R-HRSs, but does not consider the interaction of electro-hydrogen.

[0171] Option 3: EH-VPP, which includes uniform electric vehicles, uniformly air-conditioned rooms, and R-HRSs, but does not consider the spectral clustering process.

[0172] Figure 7 (a), (c), and (e) in the figure show the optimized scheduling results of schemes 1, 2, and 3 at a confidence level of 90%, respectively. Figure 7 (b), (d), and (f) in the diagram show the optimized scheduling results of schemes 1, 2, and 3 at a confidence level of 50%, respectively. Figure 7 As shown in (b) and (f), incorporating the DERs clustering process improves the power range within and within the power range, thus enhancing the aggregation results. Numerical simulations demonstrate that the overall schedulable power range at α=50% and α=90% increases by 3.42% and 4.29% respectively compared to Scheme 3, reflecting the stable enhancement of the clustering process.

[0173] contrast Figure 7As shown in (b) and (d), the dispatchable power range is significantly expanded after the addition of the electro-hydrogen interaction. The interaction increases by 4.80% and 15.04% when α = 50% and α = 90%, respectively. Benefiting from the complementarity of electrical and hydrogen energy, the flexibility range is enhanced, especially during peak hydrogen load periods. Furthermore, the increase in flexibility gradually increases due to the greater fluctuations in hydrogen load at higher confidence levels. In summary, simulation results demonstrate that, compared to traditional VPPs, EH-VPPs with clustering processes and electro-hydrogen interactions can utilize DERs more flexibly, enhancing their flexibility.

[0174] The implementation effect of EH-VPP economic dispatch.

[0175] The peak-shaving scheduling problem is formulated as a CVaR optimization model to quantitatively demonstrate the practical value of this method. Figure 8 It can be seen that the power range gradually increases with the value of α, and the increase is particularly significant during the period from 8:00 to 16:00. This indicates that higher α results in higher economic efficiency, which is proportional to the level of market risk. The peak-shaving economic scheduling model of the EH-VPP system aims to maximize overall operating profit while considering market risk. The decision-making aspects of the problem include the number of bids and the bid price.

[0176] Table IV shows the economic dispatch plan and net revenue of EH-VPP in PRM. The dispatch plan for active power and regulatory services of EH-VPP is time-varying, and bidding decisions are comprehensively affected by the upper and lower aggregated power ranges within a certain PRM price range. Therefore, the number of bids for Schemes 2 and 3 decreases compared to Scheme 1. In Table 4, the net revenue of Schemes 2 and 3 decreases by approximately 11% and 3% respectively compared to Scheme 1, indicating that, at a given confidence level, EH-VPP with clustering processes and electricity-hydrogen interactions achieves higher economic efficiency in PRM.

[0177] The results show that the number of bids decreases as the confidence level increases. Correspondingly, the net income in Scenario 1 with α=90% is 12.08% and 16.71% higher than that in Table 4 with α=70% and α=50%, respectively. Due to the greater volatility at higher confidence levels, regulatory capacity and net income tend to decline when EH-VPPs are risk-averse, and vice versa.

[0178] Table 4 Comparison of Net Income under Different Circumstances

[0179] α＝90％(￥) 35912.03 31199.01 35261.57 α＝70％(￥) 32042.70 29011.05 31087.89 α＝50％(￥) 30771.61 27240.06 29347.52

[0180] This invention utilizes an enhanced mechanism for the flexibility of electricity-hydrogen interaction to improve the aggregated flexibility and dispatchable range of VPPs for peak-shaving ancillary services, thereby increasing VPP operating revenue and promoting the consumption of renewable energy generation.

[0181] This invention designs different types of VS models with time coupling and energy / power constraints for electric vehicles, air conditioning systems, and hydrogen refueling stations that primarily utilize renewable energy and are equipped with distributed hydrogen production and storage facilities, in order to capture the flexibility of various DER clusters. The aggregation parameters of the VS models are adjusted through spectral clustering to reflect the different response characteristics of DERs and the electro-hydrogen interaction.

[0182] This invention proposes a flexible projection method based on embedded superboxes (embedded superbox method) to estimate the available power and energy space of an electro-hydrogen interactive virtual power plant (EH-VPP) from a geometric perspective. The temporally coupled power distribution of the DER cluster is treated as a polyhedron, and the Minkowski sum of multiple polyhedra is used to represent the high-dimensional flexible region of the VPP. The dimensionality of the aggregated polyhedron is further reduced through the embedded superbox approximation method.

[0183] For EH-VPP, a day-ahead peak shaving optimization scheduling model considering the dimensionality reduction prediction flexibility region and conditional risk value (CVaR) is established to maximize the expected economic benefits of the peak shaving market (PRM). The peak shaving scheduling model is discretized and transformed into a convex optimization model, and the McCrmick envelope method is used to handle the bilinear terms of bid quantity and bid price.

[0184] Example 2

[0185] In order to implement the method corresponding to Embodiment 1 above and achieve the corresponding functions and technical effects, a virtual power plant peak-shaving optimization scheduling system is provided below, including:

[0186] The energy storage model building module is used to build a virtual energy storage model for the distributed resources of the virtual power plant; the virtual energy storage model for the distributed resources includes a virtual energy storage model for electric vehicles, a virtual energy storage model for air conditioning systems, and a virtual energy storage model for hydrogen refueling stations powered by renewable energy.

[0187] The feasible region determination module is used to determine the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource using the Minkowski sum method and the embedded superbox method.

[0188] The scheduling model establishment module is used to establish a dual-objective scheduling model for peak shaving optimization based on the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resources.

[0189] The scheduling module is used to solve the peak-shaving optimization scheduling bi-objective scheduling model using the McCormick envelope method to determine the optimal solution for the distributed resources, so as to optimize the scheduling of the virtual power plant based on the optimal solution.

[0190] Example 3

[0191] The present invention provides an electronic device, including: a memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to enable the electronic device to execute the virtual power plant peak shaving optimization scheduling method of Embodiment 1.

[0192] Example 4

[0193] This invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the virtual power plant peak-shaving optimization scheduling method of Embodiment 1.

[0194] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0195] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A virtual power plant peak-shaving optimization scheduling method, characterized in that, include: A virtual energy storage model for distributed resources of a virtual power plant is established; the virtual energy storage model for distributed resources includes a virtual energy storage model for electric vehicles, a virtual energy storage model for air conditioning systems, and a virtual energy storage model for hydrogen refueling stations powered by renewable energy. The virtual energy storage model of the air conditioning system is ;in, N represents the discharge power of the air conditioner at time t; J represents the number of air conditioning system groups; N represents the discharge power of the air conditioner at time t. j Let j be the number of rooms in the j-th group of air conditioning systems; To determine the on / off state of the air conditioner in the a-th room of the j-th air conditioning system at time t, This indicates that the air conditioner is on. This indicates that the air conditioner is off; Let be the discharge power of the air conditioner in the a-th room of the j-th air conditioning system at time t; Let t be the charging power of the air conditioner at time t; Let be the charging power of the air conditioner in the a-th room of the j-th air conditioning system at time t; Let be the power of the air conditioner at time t; Let be the maximum steady-state charging power of the air conditioner in the a-th room of the j-th air conditioning system at time t. Let be the outdoor temperature at time t; Let be the equivalent thermal resistance of the a-th room in the j-th air conditioning system; Let be the cooling capacity of the a-th room in the j-th air conditioning system; Let be the energy efficiency ratio of the air conditioner in the a-th room of the j-th air conditioning system; Let be the power of the air conditioner in the a-th room of the j-th air conditioning system at time t when the temperature is most comfortable for the human body; The lower limit of the human body's comfortable temperature; The most comfortable temperature for the human body; Let be the maximum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; Let be the minimum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; The duration of the air conditioner's maximum cooling power operation during the charging cycle; The duration of the air conditioner's minimum cooling power operation during the charging cycle; This is a charge / discharge cycle; Let be the maximum steady-state discharge power of the air conditioner in the a-th room of the j-th air conditioning system at time t. This refers to the operating time of the air conditioner at its maximum cooling power during the discharge cycle. This refers to the operating time of the air conditioner at its minimum cooling power during the discharge cycle. Using the Minkowski method and the embedded superbox method, the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource is determined; wherein, the embedded superbox method is to find the largest inscribed cuboid in the high-dimensional aggregated polygon, and use the largest inscribed cuboid to represent the high-dimensional polyhedron. Using the Minkowski sum method and the embedded superbox method, the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource is determined, specifically including: Calculate the time-coupled power feasible region of the virtual energy storage model of the distributed resources; Based on the time-coupled power feasible region, the aggregated high-dimensional feasible region is determined using the Minkowski sum method. The high-dimensional feasible region of the aggregated system is decoupled and reduced in dimension using the embedded superbox method to obtain the decoupled and reduced power feasible region. Based on the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resources, a dual-objective scheduling model for peak shaving optimization scheduling is established. The McCormick envelope method is used to solve the dual-objective scheduling model for peak shaving optimization, and the optimal solution for the distributed resources is determined. The virtual power plant is then optimized and scheduled based on the optimal solution, whereby the optimal solution is the power value of the distributed resources.

2. The virtual power plant peak-shaving optimization scheduling method according to claim 1, characterized in that, The virtual energy storage model for hydrogen refueling stations powered by renewable energy is as follows: ;in, Let t be the amount of hydrogen stored in the hydrogen storage tank at time t; Let t be the amount of hydrogen stored in the hydrogen storage tank at time t-1; A hydrogen refueling station powered by renewable energy at time t at a set confidence level The equivalent external power characteristics under the following conditions; The hydrogen production conversion efficiency of the electro-hydrogen conversion device; This refers to the calorific value of hydrogen. This represents the minimum hydrogen storage capacity of the hydrogen storage tank. This represents the maximum hydrogen storage capacity of the hydrogen storage tank. Set the confidence level for PVs at time t. The lower limit of predictive power; Set the confidence level for PVs at time t. The upper limit of predictive power; For hydrogen fuel cell vehicles at time t, a set confidence level is set. The lower limit of predictive power; For hydrogen fuel cell vehicles at time t, a set confidence level is set. The upper limit of predictive power; This represents the average predictive power of PVs at time t. Let be the standard deviation of the predictive power of PVs at time t; This represents the average predictive ability of hydrogen fuel cell vehicles at time t. Let be the standard deviation of the predictive power of hydrogen fuel cell vehicles at time t; To set the confidence level.

3. The virtual power plant peak-shaving optimization scheduling method according to claim 1, characterized in that, The virtual energy storage model for electric vehicles is ;in, The virtual charging state of the electric vehicle at time t; The virtual charging state of the electric vehicle at time t-1; Let be the equivalent charge / discharge power of the electric vehicle at time t; For the charging and discharging efficiency of electric vehicles; The upper limit of the virtual charging state of the electric vehicle at time t; The virtual charging state limit for the electric vehicle at time t; This represents the upper limit of the equivalent charging and discharging power of the electric vehicle at time t. This represents the lower limit of the equivalent charge / discharge power of the electric vehicle at time t. For the first i The grid connection status of electric vehicle v in the cluster at time t; N i Let i be the number of electric vehicles in cluster i; For the first i The maximum virtual charging state of electric vehicle v in the cluster at time t; For the first i The minimum virtual charging state of electric vehicle v in the cluster at time t; For the first i The maximum equivalent charge / discharge power of electric vehicle v in the cluster at time t; For the first i The minimum equivalent charge / discharge power of electric vehicle v in the cluster at time t.

4. The virtual power plant peak-shaving optimization scheduling method according to claim 1, characterized in that, The peak-shaving optimization scheduling bi-objective scheduling model is as follows: ;in, This is a discrete expression for the conditional risk value; b The threshold for the loss function; L The number of segments into which the random variable is discretized; α To set the confidence level; The upward regulation capability of a virtual power plant with electro-hydrogen interaction after participating in the peak-shaving electricity market at time t; For virtual power plants that enable pre-dispatch electro-hydrogen interaction, a set confidence level is required. α Total operating power; To be at confidence level α Down t The maximum evaluated power of the virtual power plant at any given time; To be at confidence level α Down t Minimum estimated power of the virtual power plant at any given time; The downward regulation capability of a virtual power plant participating in the peak-shaving electricity market and its hydrogen-electric interaction at time t; ; f ( x,y l ) is the discretized loss function.

5. A virtual power plant peak-shaving optimization dispatching system, characterized in that, include: The energy storage model building module is used to build a virtual energy storage model for the distributed resources of the virtual power plant; the virtual energy storage model for the distributed resources includes a virtual energy storage model for electric vehicles, a virtual energy storage model for air conditioning systems, and a virtual energy storage model for hydrogen refueling stations powered by renewable energy. The virtual energy storage model of the air conditioning system is ;in, N represents the discharge power of the air conditioner at time t; J represents the number of air conditioning system groups; N represents the discharge power of the air conditioner at time t. j Let j be the number of rooms in the j-th group of air conditioning systems; To determine the on / off state of the air conditioner in the a-th room of the j-th air conditioning system at time t, This indicates that the air conditioner is on. This indicates that the air conditioner is off; Let be the discharge power of the air conditioner in the a-th room of the j-th air conditioning system at time t; Let t be the charging power of the air conditioner. Let be the charging power of the air conditioner in the a-th room of the j-th air conditioning system at time t; Let be the power of the air conditioner at time t; Let be the maximum steady-state charging power of the air conditioner in the a-th room of the j-th air conditioning system at time t. Let be the outdoor temperature at time t; Let be the equivalent thermal resistance of the a-th room in the j-th air conditioning system; Let be the cooling capacity of the a-th room in the j-th air conditioning system; Let be the energy efficiency ratio of the air conditioner in the a-th room of the j-th air conditioning system; Let be the power of the air conditioner in the a-th room of the j-th air conditioning system at time t when the temperature is most comfortable for the human body; The lower limit of the human body's comfortable temperature; The most comfortable temperature for the human body; Let be the maximum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; Let be the minimum cooling power of the air conditioner in the a-th room of the j-th air conditioning system; The duration of the air conditioner's maximum cooling power operation during the charging cycle; The duration of the air conditioner's minimum cooling power operation during the charging cycle; This is a charge / discharge cycle; Let t be the maximum steady-state discharge power of the air conditioner in the a-th room of the j-th air conditioning system; This refers to the operating time of the air conditioner at its maximum cooling power during the discharge cycle. This refers to the operating time of the air conditioner at its minimum cooling power during the discharge cycle. The feasible region determination module is used to determine the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource using the Minkowski sum method and the embedded superbox method; wherein, the embedded superbox method is to find the largest inscribed cuboid in the high-dimensional aggregated polygon and use the largest inscribed cuboid to represent the high-dimensional polyhedron. Using the Minkowski sum method and the embedded superbox method, the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resource is determined, specifically including: Calculate the time-coupled power feasible region of the virtual energy storage model of the distributed resources; Based on the time-coupled power feasible region, the aggregated high-dimensional feasible region is determined using the Minkowski sum method. The high-dimensional feasible region of the aggregated system is decoupled and reduced in dimension using the embedded superbox method to obtain the decoupled and reduced power feasible region. The scheduling model establishment module is used to establish a peak-shaving optimization scheduling dual-objective scheduling model based on the decoupled and dimensionality-reduced power feasible region of the virtual energy storage model of the distributed resources. The scheduling module is used to solve the peak-shaving optimization scheduling bi-objective scheduling model using the McCormick envelope method to determine the optimal solution for the distributed resources, so as to optimize the scheduling of the virtual power plant based on the optimal solution; the optimal solution is the power value of the distributed resources.

6. An electronic device, characterized in that, include: A memory and a processor, wherein the memory is used to store a computer program, and the processor runs the computer program to cause the electronic device to perform the virtual power plant peak-shaving optimization scheduling method according to any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the virtual power plant peak-shaving optimization scheduling method according to any one of claims 1-4.

Images