A method, device, equipment and medium for joint scheduling of feeder lines and charging clusters

Abstract

The application discloses a kind of feeder and charging cluster combined scheduling method, device, equipment and medium, belong to the field of equipment monitoring, the method is: based on the characteristic information of charging station constructs migration learning prediction model, determines demand prediction parameter using migration learning prediction model;With demand prediction parameter as constraint condition, determine the safety power envelope of feeder group full period;Based on safety power envelope, determine the reservation allocation constraint information of charging, determine site charging window time information using charging reservation request and reservation allocation constraint information, determine station side optical storage scheduling constraint information according to site charging window time information;According to reservation allocation constraint information, site charging window time information and station side optical storage scheduling constraint information, construct objective function and solve to obtain target scheduling scheme, based on target scheduling scheme, the site of charging cluster and feeder are scheduled processing.This application can fit the actual demand of charging station and charging vehicle, to improve scheduling accuracy.

Description

Technical Field

[0001] This invention relates to the technical field of energy dispatching, and in particular to a method, apparatus, equipment, and medium for the joint dispatching of feeders and charging clusters. Background Technology

[0002] With the increasing severity of carbon pollution, renewable energy generation and smart transportation have become key elements in the construction and development of urban power distribution networks. Electric vehicles are increasingly being integrated into the power system through charging facilities, generating a large amount of uncertain charging demand. Together with the uncertainties brought by renewable energy, this poses a significant challenge to new power distribution systems. Furthermore, the connection of a large number of new power devices, such as distributed power sources, energy storage, and electric vehicles, to the power distribution network has led to significant changes in the traditional power distribution network structure, making it difficult for existing planning technologies to meet the requirements of power supply security and stability.

[0003] To balance electricity consumption and supply, one common method for power distribution network scheduling is to construct a predictive model of electricity demand based on historical electricity consumption data, use the predictive model to predict the charging demand of new power stations, adjust the prediction results according to the type of new power stations, and finally perform power scheduling and allocation according to the adjusted electricity demand.

[0004] However, the above method has the following technical problems: the impact of different electric vehicle loads on different feeders and different equipment varies greatly. As the types of electric vehicles increase, it is often difficult to adapt to a large number of different types of electric vehicles using historical electricity consumption data for prediction. Therefore, the existing technology for predicting electricity demand and power dispatching has deviations and low dispatching accuracy. Summary of the Invention

[0005] This invention provides a method, apparatus, equipment, and medium for joint scheduling of feeders and charging clusters, which can solve one or more technical problems existing in the prior art.

[0006] A first aspect of this invention provides a joint scheduling method for feeders and charging clusters, the method comprising: A transfer learning prediction model is constructed based on the feature information of the charging station, and the demand prediction parameters are determined using the transfer learning prediction model. The demand prediction parameters include: charging power parameters and feeder group load parameters. The safe power envelope of the feeder group is determined with the demand forecast parameters as constraints, wherein the safe power envelope is the upper limit of the maximum total charging power allowed by the distribution network. Based on the safety power envelope, the charging reservation allocation constraint information is determined. The charging reservation request and the reservation allocation constraint information are used to determine the site charging window time information. The site-side photovoltaic-storage scheduling constraint information is determined based on the site charging window time information. Based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information, an objective function is constructed and solved to obtain the target scheduling scheme. Based on the target scheduling scheme, the sites and feeders of the charging cluster are scheduled.

[0007] In conjunction with the first aspect, in one implementation, the step of constructing a transfer learning prediction model based on the feature information of the charging station, and using the transfer learning prediction model to determine the demand prediction parameters, includes: Obtain feature information of the charging station, and construct a transfer learning prediction model based on the feature information; The transfer learning prediction model is invoked to predict charging power parameters, wherein the charging power parameters include: reserved power demand parameters and immediate power demand parameters; The charging power parameters are equivalently converted based on the preset time window length to obtain the feeder group load parameters.

[0008] In conjunction with the first aspect, in one implementation, determining the safe power envelope of the feeder group throughout the entire time period using the demand forecast parameters as constraints includes: The transformer margin, current margin, and voltage margin of the feeder group are determined respectively using the aforementioned demand forecast parameters as constraints. The safe power envelope to be released by the feeder group at a future scheduled time is determined using the transformer margin, the current margin, and the voltage margin.

[0009] In conjunction with the first aspect, in one implementation, the step of determining the charging reservation allocation constraint information based on the safe power envelope, determining the site charging window time information using the charging reservation request and the reservation allocation constraint information, and determining the site-side photovoltaic-storage scheduling constraint information based on the site charging window time information includes: The available capacity for reservation is calculated using the safe power envelope and the preset reservation release coefficient. The upper limit of the reservation power is determined based on the available capacity for reservation and the preset station-level weight to obtain reservation allocation constraint information. The preset station-level weight is determined based on the charging power parameters. Obtain charging reservation requests and candidate charging stations respectively. Construct a reservation time window set based on the charging reservation requests and candidate charging stations. Divide the reservation time window set according to a preset reservation cost function with the reservation allocation constraint information as the constraint condition to obtain the station charging window time information. Based on the charging reservation request, the range of charging power and the number of reservations within the charging window time information of the site are determined, and the site-side photovoltaic-storage scheduling constraint information is obtained.

[0010] In conjunction with the first aspect, in one implementation, the step of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, includes: The site charging window time information and the site-side photovoltaic-storage scheduling constraint information are integrated into an optimization decision variable set, and the objective function is constructed using the optimization decision variable set. Using the reservation allocation constraint information as constraints, a preset solution algorithm is used to solve the objective function to obtain the target scheduling scheme. The preset solution algorithm includes branch and bound, cutting plane, sequential linearization, or mixed integer quadratic programming methods. The charging cluster's sites and feeders are scheduled according to the target scheduling scheme.

[0011] In conjunction with the first aspect, in one implementation, after the steps of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, the method further includes: Extract the feeder congestion coefficient and capacity shadow price from the target scheduling scheme, and construct the reservation price model using the feeder congestion coefficient and the capacity shadow price; The reservation price model is invoked to calculate the dynamic price, and the preset reservation cost function is updated using the dynamic price.

[0012] In conjunction with the first aspect, in one implementation, after the steps of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, the method further includes: Obtain real-time arrival data and adjust the charging power parameters based on the real-time arrival data.

[0013] A second aspect of the present invention provides a joint scheduling device for feeders and charging clusters, the device comprising: The prediction determination module is used to construct a transfer learning prediction model based on the feature information of the charging station, and to determine the demand prediction parameters using the transfer learning prediction model. The demand prediction parameters include: charging power parameters and feeder group load parameters. The power determination module is used to determine the safe power envelope of the feeder group for the entire time period based on the demand forecast parameters as constraints, wherein the safe power envelope is the upper limit of the maximum total charging power allowed by the distribution network. The information determination module is used to determine the charging reservation allocation constraint information based on the safe power envelope, determine the site charging window time information using the charging reservation request and the reservation allocation constraint information, and determine the site-side photovoltaic-storage scheduling constraint information based on the site charging window time information. The scheduling module is used to construct an objective function and solve it to obtain a target scheduling scheme based on the reservation allocation constraint information, the site charging window time information and the site-side optical-storage scheduling constraint information, and to perform scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme.

[0014] Compared to existing technologies, the present invention provides a method, apparatus, device, and medium for joint scheduling of feeders and charging clusters, which offers the following advantages: The present invention can construct a transfer learning prediction model based on the characteristic information of charging stations, and use this model to determine demand prediction parameters; determine the safe power envelope of the feeder group throughout the entire time period using the demand prediction parameters as constraints; determine the reservation allocation constraint information for charging based on the safe power envelope; determine the station charging window time information using charging reservation requests and reservation allocation constraint information; determine the station-side photovoltaic and energy storage scheduling constraint information based on the station charging window time information; construct an objective function based on the reservation allocation constraint information, station charging window time information, and station-side photovoltaic and energy storage scheduling constraint information, and solve for the target scheduling scheme; and perform scheduling processing on the stations and feeders of the charging cluster based on the target scheduling scheme. The present invention determines demand prediction parameters through the characteristic information of charging stations, and then combines the demand prediction parameters and charging reservation requests to construct an objective function to solve for the scheduling scheme, thereby aligning with the actual needs of charging stations and charging vehicles and improving scheduling accuracy. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating a joint scheduling method for feeders and charging clusters according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a joint scheduling device for feeders and charging clusters provided in an embodiment of the present invention. Detailed Implementation

[0016] 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.

[0017] To address the aforementioned issues, the following detailed description and explanation will be provided through specific embodiments of a feeder and charging cluster joint scheduling method, apparatus, equipment, and medium provided in this application.

[0018] To address the technical problem of low scheduling accuracy in existing technologies, which makes it difficult to meet application requirements, this paper refers to... Figure 1 The diagram shows a flowchart of a joint scheduling method for feeders and charging clusters provided by an embodiment of the present invention.

[0019] As an example, the joint scheduling method for the feeder and the charging cluster may include: S11. Construct a transfer learning prediction model based on the feature information of the charging station, and use the transfer learning prediction model to determine the demand prediction parameters, wherein the demand prediction parameters include: charging power parameters and feeder group load parameters.

[0020] In one embodiment, feature information of the charging station can be obtained, including features such as the charging station's capacity, location, and power. Then, a transfer learning prediction model is constructed based on the feature information of the charging station, and the transfer learning prediction model is used to determine demand prediction parameters. The demand prediction parameters include charging power parameters and feeder group load parameters.

[0021] Charging power parameters can be power-related parameters of the charging station. Feeder group load parameters can be relevant parameters of the distribution network feeder load.

[0022] In one embodiment, constructing a transfer learning prediction model based on the feature information of the charging station and using the transfer learning prediction model to determine the demand prediction parameters may include the following sub-steps: S111. Obtain the feature information of the charging station and construct a transfer learning prediction model based on the feature information.

[0023] S112. Call the transfer learning prediction model to predict the charging power parameters, wherein the charging power parameters include: the reserved power demand parameters and the immediate power demand parameters.

[0024] S113. Based on the preset time window length, the charging power parameters are equivalently converted to obtain the feeder group load parameters.

[0025] Assume the set of charging stations is The source site set is The target station set is For any charging station It can obtain the feature information of charging stations and use the feature information of charging stations to construct site feature vectors, as shown in the following formula: ; in, Indicates geographical location characteristics, Indicates road accessibility characteristics, Indicates the characteristics of surrounding land use or business formats. This indicates the configuration characteristics of the charging gun. Indicates the characteristics of station-side energy storage. Indicates the characteristics of photovoltaic power generation at the station. Indicates the identifier of the feeder or feeder group to which it belongs. This indicates the characteristics of historical utilization rate.

[0026] Based on the characteristics of charging stations, charging stations are divided into several prototype classes, and prototype mappings are defined. Then, a transfer learning prediction model is built. Its training objective can be expressed as follows: ; in, Indicates site The input feature sequence, Indicates the corresponding label sequence; and This indicates a small number of historical samples from the target station; To share model parameters; Configure parameters for the target station; These are the prior parameters of the corresponding prototype; For prediction loss function; and These are the weighting coefficients.

[0027] The transfer learning model employs a structure of "shared feature extractor + prototype prior output layer + target station few-sample regularized adaptation layer". The shared feature extractor can be trained using historical data from multiple source stations. The target station adaptation layer uses a regularized least squares form with prototype and global priors to provide a closed-form solution, ensuring that a prediction model can be quickly obtained even for newly built stations or stations with few samples, even with limited samples.

[0028] Let the set of charging stations be: ; in, This is a collection of source sites with relatively abundant historical data. This refers to the set of target stations that are newly established, have a small sample size, or have undergone significant changes in operation. For any given station... Construct the static feature vector of the site: ; in, Indicates geographical location characteristics, Indicates road accessibility characteristics, Indicates the characteristics of the surrounding businesses. This indicates the configuration characteristics of the charging gun. This indicates the characteristics of the station-side energy storage configuration. This indicates the characteristics of the photovoltaic configuration on the station side. Indicates the characteristics of the feeder or feeder group to which it belongs. This indicates the characteristics of historical utilization rate.

[0029] Standardize site features: ; in, This is the mean vector of site features. It is a diagonal matrix formed by the characteristic standard deviations.

[0030] Let the set of site prototypes be: ; No. Class prototype center is Site Belongs to the prototype The distance is defined as: ; in, For the first Site feature covariance matrix. Site prototype classification results are as follows: ; If soft categorization is required, then the site Belongs to the prototype The probability can be expressed as: ; in, This is the temperature coefficient. The corresponding hard classification can be derived from: ; get.

[0031] For transfer learning prediction models, let the site be... At any moment The input features are: ; in, Represents the historical charging load sequence. Indicates weather characteristics, Indicates time characteristics such as holidays, days of the week, and time periods. This indicates the static characteristics of the site.

[0032] Define a shared feature extraction function: ; in, To enable cross-site parameter sharing, the forecast output includes both pre-booking demand and on-demand top-up demand. ; in, For the site Output layer adaptation parameters.

[0033] For the source site, the pre-training objective is: ; in, Indicates the first Output layer parameters corresponding to the site prototype. Tag based on actual demand.

[0034] For the target site If a small number of historical samples are already available, construct the feature matrix and the label matrix: ; ; The target site adaptation issue can be described as: ; in, These are the prior parameters of the prototype to which the target station belongs. The global average prior parameters for all source sites. and This is the regularization coefficient.

[0035] The target station adaptation problem has a closed-form solution: ; Therefore, new stations or stations with few samples can quickly obtain a station adaptation prediction model with a small number of samples based on a shared feature extractor and prototype priors. If the shared feature extractor uses a nonlinear neural network, the overall pre-training process is generally solved by gradient descent; however, the last layer adaptation parameters of the target station can be given by the above formula as a closed-form solution.

[0036] For each station and time period, forecast both reservation demand and immediate arrival demand.

[0037] For each site and time period The transfer learning prediction model outputs the reserved power demand parameter (i.e., the predicted value of the reserved demand) and the immediate power demand parameter (i.e., the predicted value of the charging demand) respectively, thus obtaining the charging power parameter.

[0038] Specifically, the reserved power demand parameter (i.e., the reserved demand forecast value) and the immediate power demand parameter (i.e., the charging demand forecast value) can be represented by the following formula: ; ; in, Indicates site During the period The system expects reservation requests to be submitted in advance through the reservation platform, which will allow for the allocation of stations, time windows, and charging power plans; these are the reservation power demand parameters. Indicates site During the period The on-demand charging needs of those who arrive at the station without prior reservation, and the on-demand power requirements parameters.

[0039] The difference between the two is that reservation demand is a controllable load that can be arranged in advance, and is subsequently used for reservation capacity release, site-time window joint allocation, and calculation of the power of the accepted reservation plan; on-demand charging demand is a load with strong randomness that cannot be fully allocated in advance, and is subsequently used for feeder group risk margin reservation, site-level reservation safety constraints, and site-side opportunity constraints.

[0040] Site During the period The total demand forecast can be expressed as follows: ; To describe the uncertainty of pre-booking demand and on-demand top-up demand, quantile forecasts for both are further provided. Let... For quantile levels, it is often taken as , or ,in This represents the median demand forecast. or Let the upper bound of demand at a high confidence level be: ; ; in, Represents the conditional quantile function. This indicates that the demand for appointments is at the quantile level. The predicted value below, This indicates that the demand for instant charging is at the quantile level. The predicted value is below.

[0041] To ensure compatibility with the safety capacity of subsequent feeder groups and station-side power constraints, if and To indicate the electricity demand within a time period, you can use the length of the time window. Converting to equivalent power, the feeder group load parameters are obtained. The feeder group load parameters can be expressed as follows: ; ; ; If the model directly outputs the power demand, then let , , .

[0042] Furthermore, set up stations The feeder group to which it belongs is feeder group The set of sites within can be represented as follows: ; Then feeder group During the period The high-quantile forecast value of instant charging demand can be obtained by aggregating the high-quantile forecast values ​​of instant charging at each station within the feeder group: ; This amount is used in subsequent steps as part of the feeder group risk reserve margin to prevent the system from releasing all remaining feeder group capacity to reserved users. Accordingly, the on-demand charging uncertainty term in the feeder group risk reserve margin can be expressed as: ; in, Indicates feeder group During the period This is a risk reserve capacity for random loads that require immediate charging. This risk reserve capacity, together with the base load forecast error and the photovoltaic output fluctuation margin, constitutes the risk reserve margin in the safety capacity envelope of the subsequent feeder group.

[0043] In subsequent steps, the site During the period The power of the accepted reservation plan The station-level reservation security constraints should be met in conjunction with the high demand for instant top-up. ; in, Indicates site During the period The available operating power limit is obtained by comprehensively considering the capacity of charging stations within the station, the decomposition results of the safety capacity of the feeder group, and the station-side energy storage and photovoltaic support capabilities. This constraint indicates that when accepting reservations, the system cannot simply rely on the planned power of reserved users to fully utilize the station or feeder group capacity; instead, it should reserve corresponding safety capacity for users who do not make reservations and charge on demand.

[0044] Therefore, the output of this step , , , and their corresponding power forms , , These will serve as inputs for subsequent reservation capacity release, feeder group safety capacity envelope, site-time window joint allocation, site-side opportunity constraint scheduling, and rolling correction, thereby achieving a closed-loop connection between the prediction results and subsequent control steps.

[0045] in, The quantile level is usually set to 0.5, 0.9, or 0.95 to allow for the construction of a safety margin in subsequent calculations.

[0046] It should be noted that the reservation demand and the immediate arrival demand are predicted separately because the two have different controllable attributes in the reservation linkage control of the charging station group-feeder group.

[0047] Reservation demand originates from reservation requests submitted in advance or expected by users. It features the characteristics of being allocable, sortable, transferable, and plannable. It is mainly used for subsequent steps such as releasing reservation capacity, joint allocation of site-time windows, and reservation linkage optimization goals.

[0048] The demand comes from users who arrive at the station to charge without prior reservation. It is random and cannot be fully scheduled. It cannot be directly allocated as a deterministic reservation plan. Instead, it should be considered as a risk margin for uncertain loads entering the feeder group, a station-level reservation safety constraint, and a station-side opportunity constraint.

[0049] Specifically, the reservation demand forecast is used to estimate the reservation request scale for each site and time window. In subsequent steps, it, along with the available capacity of the feeder group, determines the number of reservation slots and the upper limit of reservation power for each site. After a user actually submits a reservation, this portion of the demand is converted into the accepted reservation planned power, which is then used in subsequent steps to determine the site's charging window time information and the site-side photovoltaic and energy storage scheduling constraints. The on-demand demand forecast describes the random arrival load of unreserved vehicles. Its mean, standard deviation, and high quantile forecast values ​​are respectively used in the risk reserve margin, site-level reservation safety constraints, and feeder group opportunity constraints, ensuring that the system does not release all feeder margin to reserved users when accepting reservations, but instead reserves necessary safety capacity for random on-demand users. Therefore, these two forecasts are not redundant, but correspond to "planned load" and "randomly reserved load" respectively, jointly constituting the inputs for subsequent safety capacity envelope and reservation linkage optimization.

[0050] S12. Determine the safe power envelope of the feeder group for the entire time period using the demand forecast parameters as constraints, wherein the safe power envelope is the upper limit of the maximum total charging power allowed by the distribution network.

[0051] In one embodiment, the safe power envelope of the feeder group throughout the entire time period can be determined using demand forecast parameters as constraints, wherein the safe power envelope is the upper limit of the maximum total charging power allowed by the distribution network.

[0052] In one operating mode, different margins of the charging station can be determined by using demand forecast parameters as constraints. By superimposing the different margins, the safe power envelope of the feeder group at all times (the upper limit of the maximum total charging power allowed by the distribution network) can be generated.

[0053] In an optional embodiment, determining the safe power envelope of the feeder group throughout the entire time period using the demand forecast parameters as constraints may include the following sub-steps: S121. Determine the transformer margin, current margin, and voltage margin of the feeder group respectively, using the aforementioned demand forecast parameters as constraints.

[0054] S122. The safe power envelope to be released by the feeder group at a future scheduled time is determined using the transformer margin, the current margin, and the voltage margin.

[0055] Let the set of feeder groups be Site The feeder group to which it belongs is denoted as For feeder groups Zhongren First Branch Road And parting The three-phase power flow relationship can be established as follows: ; ; ; in, and These are the contributing and non-contributing tidal currents of the branch lines, respectively. and Basic load; and For charging load; and Injecting into photovoltaics; and Branch parameters; This represents the node voltage.

[0056] Let the set of feeder groups be: The charging station collection is Define the station-feeder group mapping function: ; If the site Connect to feeder group Then there is Accordingly, the station-feeder mapping matrix is ​​defined as follows: ; in: ; Then feeder group The collection of sites within is: ; Further define the site-node-phase mapping. Let the set of distribution network nodes be... The separate sets are If the site Access Node The set of phases Then define the mapping matrix: ; Therefore, the site The change in charging power can be mapped to the change in injected power at the corresponding node phase.

[0057] In one operating method, the station-feeder mapping operation of the present invention is specifically as follows: Enter the charging station access node, its feeder, transformer area or feeder group number, and the distribution network topology data.

[0058] The first step is to read each site. access node Access phase and its feeder group .

[0059] The second step is to generate the station-feeder mapping matrix: ; The third step is to generate the site-node-phase mapping matrix. .

[0060] The fourth step is to obtain the set of stations within the feeder group based on the mapping matrix: ; Output , and .

[0061] In one embodiment, the three-phase AC power flow equation can be written as: ; in, For nodes Parting Complex voltage, These are elements of the three-phase node admittance matrix. Equivalently: ; ; At the reference operating point In the vicinity, the three-phase power flow equations are linearized. Let the state variables be: ; in, Let n be the phase angle vector. This is the voltage magnitude vector. The power injection vector is: ; Then we have: ; in, For the three-phase current Jacobian matrix: ; like If not singular, then: ; If the charging load at the site is calculated based on the power factor... When the system is running, the relationship between its reactive power change and active power change can be expressed as: ; ; Therefore, the power injection disturbance corresponding to the change in charging power at all stations can be written as: ; in, The vector of active power change for charging stations. , and These are the mapping matrices from the changes in active and reactive power at the station to the injection vectors of the node phases, respectively.

[0062] Therefore, the sensitivity matrix of node voltage amplitude to station charging power is: ; in, From state variables The selection matrix for extracting voltage amplitude components. Matrix elements: ; Indicates site The effect of changes in charging power on nodes Separation The effect of voltage amplitude.

[0063] For branch currents, let the branch current vector be: ; in, The matrix for calculating branch currents. Linearizing with respect to the reference point, we have: ; The change in branch current amplitude can be approximated as: ; Therefore, the sensitivity matrix of branch current to station charging power is: ; in, This represents a matrix formed by the linearized relationships of branch current magnitudes. Matrix elements: ; Indicates site The change in charging power affects the branch Separation The effect of current amplitude.

[0064] Accordingly, the charging power plan for a given site The node voltages and branch currents can be approximately updated as follows: ; ; The aforementioned sensitivity matrix is ​​used for subsequent steps such as feeder group safety margin calculation, safety constraints, and emergency re-optimization.

[0065] In one embodiment, the feeder group can be defined with demand forecast parameters as constraints. At any moment The transformer margin, current margin, and voltage margin can be specifically represented by the following formula: ; ; ; in, The rated capacity of the feeder group transformer, Predicted apparent power under base load This is the upper limit of the branch current carrying capacity. Predicting current in the basic scenario To inject sensitivity to branch currents into the feeder group. This represents the upper limit of node voltage or the boundary corresponding to the safety margin. To predict voltage, To inject sensitivity to node voltage, It is a very small positive number.

[0066] In one operating method, the execution flow of sensitivity calculation in this invention is as follows: Input the three-phase node admittance matrix, branch parameters, base load forecast, photovoltaic forecast, and site access location.

[0067] The first step is to solve for the baseline three-phase power flow and obtain... , , and .

[0068] The second step is to calculate the three-phase power flow Jacobian matrix at the reference operating point: ; The third step is to generate based on the site access location and power factor. , and .

[0069] Step 4: Calculate the voltage sensitivity matrix: ; Step 5: Calculate the linearization matrix of branch current amplitude. And obtain the current sensitivity matrix: ; Output and .

[0070] If energy storage is configured within the feeder group, the net energy storage support capacity can be defined as follows: ; Therefore, the safe power envelope definition for the feeder group, which can be used in the future planned release, can be expressed as follows: ; in, To reserve a margin for risk, the following formula can be used: ; in, , , These are the uncertainties of base load, immediate charging demand, and photovoltaic fluctuations, respectively. This represents the risk coefficient.

[0071] S13. Determine the charging reservation allocation constraint information based on the safe power envelope, determine the site charging window time information using the charging reservation request and the reservation allocation constraint information, and determine the site-side photovoltaic-storage scheduling constraint information based on the site charging window time information.

[0072] In one embodiment, the charging reservation allocation constraint information and the site charging window time information can be determined based on the safety power envelope.

[0073] The reservation allocation constraints can include the maximum power capacity and the number of reservation slots at each station. The station charging window time information can refer to multiple candidate charging stations that offer charging and multiple acceptable reservation time windows.

[0074] Next, the station-side photovoltaic and energy storage scheduling constraints can be determined based on the station's charging window time information. These constraints can include planned charging power and the station's cumulative reserved occupancy.

[0075] In one embodiment, determining the charging reservation allocation constraint information and the site charging window time information based on the safe power envelope, and determining the site-side photovoltaic-storage scheduling constraint information based on the site charging window time information, may include the following sub-steps: S131. Calculate the available reservation capacity using the safe power envelope and the preset reservation release coefficient, determine the upper limit of reservation power based on the available reservation capacity and the preset station-level weight, and obtain reservation allocation constraint information, wherein the preset station-level weight is determined based on the charging power parameters.

[0076] S132. Obtain charging reservation requests and candidate charging stations respectively. Construct a reservation time window set based on the charging reservation requests and candidate charging stations. Divide the reservation time window set according to a preset reservation cost function with the reservation allocation constraint information as the constraint condition to obtain the station charging window time information.

[0077] S133. Based on the charging reservation request, determine the range of charging power and the number of reservations within the charging window time information of the site, and obtain the station-side photovoltaic-storage scheduling constraint information.

[0078] In one embodiment, the total reservation power that the feeder group can release to reservation users during the corresponding time period can be obtained by multiplying the safe power envelope of the feeder group obtained in the above steps by the reservation release coefficient, thus obtaining the reservation release coefficient; secondly, a station-level allocation weight is constructed based on factors such as the station's gun position capacity, the station-side energy storage and photovoltaic buffer capacity, and the predicted intensity of reservation demand, and the reservation power of the feeder group is decomposed to each charging station; thirdly, the station's own gun position capacity, the station-side power limit, and the demand for immediate charging are combined to form the station's reservation power limit; finally, based on the average reservation power per vehicle, the average reservation electricity per vehicle, the time window length, and the number of available gun positions, the station's reservation power limit is converted into the number of reservation slots that can be released, thus obtaining reservation allocation constraint information.

[0079] Meanwhile, the obtained site reservation power limit and reservation quota are not defined in isolation, but serve as hard constraints for determining the site charging window time information in subsequent steps. User reservation requests will only be accepted if they do not exceed the site reservation quota, the site reservation power limit, or the available capacity of the feeder group to which the user belongs.

[0080] Specifically, feeder group At any moment The available capacity for reservations can be represented by the following formula: ; in, This is the reservation release coefficient, used to control the ratio between release capacity and security capacity.

[0081] The reserved capacity of the feeder group is allocated to each site, and a preset site-level weight is defined, as shown in the following formula: ; in, This indicates the station's gun position capability factor. This indicates the station-side energy storage / photovoltaic buffer capacity factor. These are the weighting coefficients.

[0082] Then the site At any moment The upper limit of the reserved power can be expressed as follows: ; The following conditions must be met simultaneously: ; To handle the overlap of immediate arrival users and reserved users, a site-level reservation security constraint is defined, as shown in the following formula: ; in, The planned power corresponding to the reservations already accepted at the site. The 95th percentile power forecast for the imminent demand.

[0083] In one operating mode, the operations of reservation capacity decomposition, cross-site reassignment priority, and reservation pricing model are as follows: Based on the above-obtained feeder group safe power envelope The feeder group can publish the following reservation capacity: ; in, This is the reservation release coefficient.

[0084] For the site Define the site comprehensive allocation factor: ; in, Indicates the available gun position power capacity of the site. Indicates the energy storage and photovoltaic buffering capacity. This represents the predicted power of reservation demand, and: ; The site reservation capacity allocation weight is as follows: ; The initial reservation power limit for the site is: ; Considering the available operational capacity of the sites and the risk of instant charging, the final maximum power capacity for site reservations is: ; in, This represents the maximum available operating power for the site. High-resolution power prediction to meet immediate charging demand.

[0085] The number of reservation slots that can be published on the site is: ; in, The average scheduled charging power per vehicle, Number of available guns, This represents the average time spent.

[0086] The number of reservations accepted and the power are as follows: ; ; And satisfy: ; ; ; When there are sudden changes in feeder group capacity, equipment failures, or increased arrival deviations, the system prioritizes reservation requests that have not yet been executed. Let the request... The priority rating is: ; in, Indicates whether it is a high-priority vehicle. Indicates the SOC margin of the travel chain. Indicates time flexibility. Indicates a request A set of feasible alternative sites—time windows This indicates the minimum amount of electricity needed to replenish the supply chain. to This is a weighting factor. The higher the score, the less likely the request is to be reassigned or delayed; the system will prioritize this. Higher requests, for Lower-value requests with greater time or site flexibility are reassigned.

[0087] ask The set of feasible alternatives is defined as: ; in, and These represent the remaining reservation slots and remaining reservation capacity for alternative sites and time windows, respectively.

[0088] The cross-site reassignment target is: ; in, In order to take a detour, As a result of the delay, To replace the shadow price of feeder group capacity.

[0089] The reservation price model is as follows: ; in, For site reservation occupancy rate, This is the feeder congestion coefficient. To mitigate the risk of users breaking their promises, It is flexible in terms of time.

[0090] The reservation fee and capacity reservation fee are as follows: ; ; The final payment amount is: ; in, This is to provide incentives for off-peak periods.

[0091] Next, charging reservation requests and candidate charging stations can be obtained separately. A set of reservation time windows can be constructed based on the charging reservation requests and candidate charging stations. Then, the set of reservation time windows can be divided according to the preset reservation cost function with reservation allocation constraint information as the constraint condition to obtain the station charging window time information.

[0092] It should be noted that the charging reservation request can include the user's departure point, destination, earliest arrival time, latest departure time, current SOC, required power, candidate stations, and time flexibility. Specifically, the information in the charging reservation request can be transformed into optimizable mathematical variables.

[0093] The purpose of obtaining the site charging window time information is to simultaneously determine whether a user is accepted, which site they are assigned to, in which time window they charge, and the planned charging power or planned charging amount among multiple candidate charging stations and multiple acceptable reservation time windows. The time window variable is not defined in isolation, but rather serves as a unified time index for subsequent reservation capacity constraints, site gun position constraints, feeder group capacity constraints, station-side scheduling plans, reservation prices, and cross-site reassignment judgments.

[0094] Assume the set of charging reservation requests is The set of candidate charging stations is The appointment time window is set as follows Among them, time window A continuous time interval that can be booked can be represented as: ; in, and Representing time windows The start and end times. If station-side scheduling uses a finer time scale, then the time window... It can be further divided into several station-side scheduling time periods, denoted as: ; in, This indicates the segmented time periods used for power scheduling of station-side energy storage, photovoltaic power, and charging piles. Therefore, the time window for the reservation layer... Scheduling time slots of the execution layer Establish corresponding relationships.

[0095] Any appointment request It can be represented as: ; in, and These represent the user's departure point and destination, respectively. Indicates the earliest time a user can reach the charging station; Indicates the user's latest departure time; Indicates the vehicle's battery capacity; This indicates the current charge status of the vehicle when the reservation was submitted; This indicates that the user wants to replenish the battery. Indicates a request A set of candidate charging stations.

[0096] For the request and candidate sites Based on the user's arrival time Earliest arrival time Latest departure time And the site's opening hours, constructing its acceptable time window set: ; in, Indicates a request On the site An acceptable set of reservation time windows. If a time window does not meet the user's arrive time or latest departure time constraints, that time window will not participate in subsequent allocations.

[0097] Define binary decision variables: ; And define the request acceptance variables: ; Each reservation request will be assigned to at most one site and one time window, provided that: ; ; This constraint indicates that if Then request It must be assigned to one candidate site and one acceptable time window; if If not, the request will not be accepted.

[0098] Suppose a request On the site Time window The planned charging power within is The planned replenishment of electricity is Then the two satisfy: ; in, Indicates vehicle The charging efficiency. If requested... Not assigned to a site Time window Therefore, the planned charging power within that time window should be zero, hence: ; in, Indicates vehicle On the site The maximum acceptable charging power is determined by both the vehicle's charging interface capability and the charging gun's power at the charging station.

[0099] As can be seen from step S131 above, the site In the time window The number of reservations accepted is: ; It should meet the site's limit on the number of reservation slots that can be published: ; in, This indicates the site obtained through the above steps. In the time window You can publish the number of reservation slots.

[0100] Site In the time window The accepted reservation capacity is: ; It should meet the site reservation power limit constraint: ; in, This indicates the site obtained in step S131. In the time window The maximum power limit for reservations.

[0101] Furthermore, if the site The feeder group to which it belongs is feeder group The collection of sites within is Then all stations within the same feeder group will be within the same time window. The sum of the accepted reservation power should satisfy: ; in, This indicates the feeder group obtained in step S131. In the time window The system can publish the available capacity for reservations. This formula indicates that even if a station still has available gun positions, the system cannot accept new reservations within that time window if the available capacity of the feeder group to which that station belongs is insufficient.

[0102] To evaluate the merits of different site-time window combinations, a request is defined. Assigned to a site Time window Generalized reservation cost function: ; in, Indicates a request to station Travel time; Indicates site In the time window Predicted queue waiting time; Indicates site In the time window The reservation price; Indicates site feeder group In the time window The cost of congestion; Indicates the user selects a site The cost of detouring relative to the original route; Indicates time window The cost of deviation relative to user preference time; to These are the weighting coefficients.

[0103] Based on this, the station charging window time information allocated to the reservation request can be represented by the following formula: ; in, This represents the penalty coefficient for not accepting the reservation request. This objective, together with the station-side operating cost, energy storage degradation cost, and feeder group over-limit penalty term in subsequent step six, constitutes the overall objective function for reservation linkage.

[0104] Therefore, time window This step not only represents the time and location of the user's reservation, but also directly enters the site reservation quota constraint, site reservation power constraint, feeder group reservation capacity constraint, and generalized reservation cost function.

[0105] Based on the charging reservation request, the range of charging power and the number of reservations within the charging window time information of the site are determined, and the site-side photovoltaic-storage scheduling constraint information is obtained.

[0106] Optionally, the on-site optical-storage scheduling constraint information may also include: the optimal allocation result. Request Acceptance Status Planned charging power The number of reservations accepted by the site The site has accepted reservations for power. .

[0107] in, This will be used as input for subsequent steps; and This will be used to update the remaining reservation slots and remaining reservation capacity on the site; and This will be coordinated with subsequent reservation pricing and congestion pricing adjustments; if sudden changes in feeder capacity or user arrival deviations occur during operation, those not yet implemented will be affected. This will be the target of adjustments for subsequent steps such as cross-site reassignment, postponement, or reduction of quota.

[0108] Specifically, the available reservation slots Ns,τpub, reservation power limit Ps,τres,max, and the available reservation capacity Pg,τpub of feeder group g in time window τ are obtained for station s. The binary variable xr,s,τ is used to indicate whether reservation request r occupies the reservation capacity of station s in time window τ. Therefore, time window τ is directly used to constrain the number of accepted reservations, the power of accepted reservations, the reservation capacity of feeder group, and the planned charging power of users within the same time window. If a request is assigned to station s and time window τ, its planned power is included in Ps,τacc and used as input for generating the target scheduling scheme in subsequent steps; at the same time, the station occupancy rate and feeder congestion level corresponding to this time window can be used for reservation price adjustment; when feeder capacity changes suddenly or user arrival deviation occurs during the operation phase, the remaining reservation capacity Ps,τrem and remaining reservation slots Ns,τrem in this time window serve as the basis for feasibility judgment of cross-site reassignment and emergency re-optimization in subsequent steps. Therefore, the time window variable runs through the entire process of reservation acceptance, capacity verification, station-side scheduling, price incentives, and emergency reassignment.

[0109] In one embodiment, trip chain feasibility means that after a vehicle completes its scheduled charging, it is not required to be fully charged. Instead, it is required that its State of Charge (SOC) upon leaving the station is sufficient to support the entire subsequent journey from the current charging station to the next available charging opportunity, and that it does not fall below a set minimum SOC safety threshold after each segment of the journey.

[0110] Specifically, the energy consumption for each subsequent leg of the journey can be calculated based on the user's submitted destination, navigation distance, vehicle energy consumption per unit mileage, road congestion correction factor, and safety margin. Then, the minimum SOC required for the vehicle to leave the station is calculated using "journey energy consumption + safety reserve power + minimum SOC reserve". Finally, the minimum amount of electricity needed to replenish within the current reservation window is calculated by combining the vehicle's SOC upon arrival at the charging station. This minimum replenishment power serves as the energy constraint for the station's charging window time information, and together with the station's reservation power limit and reservation quota constraints obtained in step S131, determines whether a reservation request is acceptable. If the feeder network capacity is limited, the system may not fulfill the user's "fully charged" request, but as long as the minimum SOC constraint for the subsequent travel chain is met, the reservation scheme is still considered feasible.

[0111] When considering mobility-aware constraints, the vehicle is not required to be fully charged during the scheduled charging session. Instead, it is required that after leaving the assigned charging station, the vehicle completes the subsequent travel chain from that station to the next available charging opportunity, and maintains a state of charge (SOC) of at least 1% throughout the entire travel chain. Therefore, the charging demand for a scheduled charging request is no longer determined solely by whether the vehicle is fully charged, but rather by the energy consumption of the subsequent journey, the minimum SOC requirement, and the safety reserve power.

[0112] Suppose a request Assigned to a site Time window The vehicle arrives at the station from its current location. The driving energy consumption is: ; in, Indicates vehicle Energy consumption per unit distance, the unit can be ; This indicates that the vehicle has arrived at the station from its current location. The driving distance. If we consider the effects of road congestion, temperature, gradient, and air conditioning load, it can be further written as: ; in, This is the energy consumption correction factor for travel to the destination.

[0113] Vehicle arrives at the station The SOC at that time was: ; in, Indicates a request The vehicle's current SOC at the time of submission. This indicates the vehicle's battery capacity.

[0114] Assume the vehicle leaves the station The subsequent travel chain before the next available charging opportunity is as follows: ; in, Indicates the subsequent Segment of the journey. The first... The distance traveled in this segment is The energy consumption correction factor is The energy consumption for this segment of the journey is: ; From leaving the charging station to completing the first The cumulative energy consumption for this segment of the journey is: ; To ensure that the vehicle meets the minimum SOC requirement at any node in the subsequent travel chain, let the minimum permissible SOC for the vehicle be... The safe backup power is Then the vehicle leaves the station. The minimum departure SOC required to be met is: ; in, This represents the minimum departure SOC required for the vehicle to complete the subsequent travel chain. The formula indicates that the system not only checks the total energy consumption of the entire travel chain, but also checks each intermediate travel node to ensure that the vehicle does not fall below the minimum SOC safety threshold at any subsequent node.

[0115] Therefore, request On the site Time window The minimum amount of electricity that needs to be replenished is: ; If the user also provided a minimum expected battery replenishment... Therefore, the minimum planned charging capacity that should be met for this scheduled charging is: ; Meanwhile, considering the battery SOC limit and the user's expected replenishment capacity, the planned replenishment capacity should meet the following requirements: ; in: ; In the formula, This indicates that the user wants to replenish the battery. This indicates the maximum SOC (State of Charge) that the vehicle is allowed to charge to.

[0116] Therefore, if a request Assigned to a site Time window Then its planned power replenishment needs to meet the following requirements: ; Based on the aforementioned relationship between planned replenishment capacity and planned charging power: ; The lower limit constraint of the planned charging power within the time window can be obtained: ; And upper limit constraints: ; The above constraints indicate that the time window Whether it's acceptable depends not only on whether the station has reservation slots and the feeder network has capacity, but also on whether the minimum battery power required for the subsequent travel chain of vehicles can be replenished within that time window. If a particular station-time window combination cannot meet this requirement: ; Or, if the corresponding charging power cannot be provided within the time window, then the site-time window combination will not support the request. Not feasible.

[0117] Furthermore, the vehicle departure SOC can be expressed as: ; The following The SOC after the end of the segment is: ; And should satisfy: ; This ensures that the vehicle can complete the subsequent travel chain after this scheduled charging, and retains the necessary SOC safety margin after each subsequent trip.

[0118] For example, the battery capacity of a user's vehicle The current SOC is The vehicle travels from its current location to a candidate charging station. The distance is Energy consumption per unit distance is The energy consumption correction factor at the destination is taken as follows: Then the energy consumption at the destination is: ; Vehicle arrival SOC is: ; That is approximately .

[0119] Assuming that after leaving the charging station, the user needs to complete two more trips before the next available charging opportunity, the first trip is... The second paragraph is Energy consumption per unit distance remains at Then the total energy consumption of the subsequent travel chain is: ; If you require the vehicle to remain after the subsequent trip... And set a safe backup power supply. The minimum SOC required for a vehicle to leave the station is: ; That is approximately Therefore, the minimum amount of electricity needed to replenish for this scheduled charging is: ; In other words, if the feeder network experiences capacity shortages during peak periods, the system does not need to fully charge the user or completely meet their high expected power replenishment. It only needs to replenish the user with at least [amount missing] power within the allocated time window. The amount of power is sufficient to ensure that it can complete the subsequent tasks. The travel chain is maintained while preserving a safe SOC. If the time window length is... Charging efficiency Therefore, the corresponding planned charging power is at least: ; Therefore, as long as the site In the time window The remaining reservation power within the area is not less than If the reservation quota at the station and the reservation capacity of the feeder group both meet the constraints, then the reservation scheme can be determined as a feasible travel chain. Conversely, if the remaining power is insufficient within the time window, the system should re-search for feasible schemes in other candidate time windows or candidate stations, or implement postponement, cross-station reassignment, or partial charging strategies in subsequent steps.

[0120] In one embodiment, the evolution of the trip chain SOC and the calculation of the minimum charging energy can be as follows: Set up an appointment request Assigned to a site Time window The vehicle battery capacity is When submitting the appointment, the SOC is The vehicle arrives at the station from its current location. The distance is Energy consumption per unit distance is The energy consumption correction factor is Then the energy consumption at the destination is: ; Vehicle arrival SOC is: ; Let the subsequent travel chain from when the vehicle leaves the station until the next available charging opportunity be: ; No. The distance of the segment is The energy consumption correction factor is The energy consumption for this segment of the journey is: ; To the The cumulative energy consumption at the end of the journey is: ; Let the minimum permissible SOC of the vehicle be The safe backup power is The minimum SOC required for a vehicle to leave the station is: ; Therefore, vehicles at the station The minimum charge replenishment amount for this scheduled charging is: ; If the user also provides the minimum expected power replenishment Therefore, the lower limit of the planned replenishment power is: ; Considering users' expectations for battery replenishment and the highest allowed SOC The planned maximum replenishment capacity is: ; If request Assigned to a site Time window The planned power replenishment should meet the following requirements: ; The planned charging power and the planned replenishment power must satisfy the following: ; in, For charging efficiency. Therefore, the planned charging power should meet the following requirements: ; Vehicle departure SOC is: ; The following The SOC after the end of the segment is: ; The feasibility constraints of the travel chain are: ; The above formula constitutes an executable calculation method for the SOC constraint of the travel chain.

[0121] In addition, reservation allocation constraint information can be used as input to determine the site-side optical storage scheduling constraint information; the site reservation capacity occupancy rate of reservation allocation constraint information can also be used to form reservation prices and congestion price corrections; the site remaining reservation power and remaining reservation slots of reservation allocation constraint information can also be used for cross-site reassignment and emergency re-optimization.

[0122] In one operational method, the travel chain SOC constraint of this invention can be executed according to the following process: Input the user's current SOC, battery capacity, candidate stations, distance to the station, subsequent travel chain, energy consumption per unit distance, minimum SOC, safe reserve power, and candidate time window.

[0123] The first step is to calculate the energy consumption of the vehicle when it reaches the candidate station: ; The second step is to calculate the destination SOC: ; The third step is to calculate the energy consumption for each subsequent segment of the journey and the cumulative energy consumption: ; ; Step 4: Calculate the minimum off-site SOC: ; Step 5: Calculate the minimum replenishment capacity: ; Step 6: Determine if the candidate site and time window are feasible. If they meet the following conditions: ; and: ; and If the candidate site-time window satisfies the travel chain SOC constraint, then the candidate site-time window is deemed infeasible; otherwise, it is deemed infeasible.

[0124] Output feasible candidate set Minimum charge replenishment and minimum planned charging power .

[0125] S14. Construct an objective function and solve it to obtain an objective scheduling scheme based on the reservation allocation constraint information, the site charging window time information and the site-side optical storage scheduling constraint information. Then, perform scheduling processing on the sites and feeders of the charging cluster based on the objective scheduling scheme.

[0126] In one embodiment, an objective function is constructed based on reservation allocation constraint information, site charging window time information, and site-side optical-storage scheduling constraint information. The objective function includes user reservation cost + site-side operating cost + distribution network over-limit penalty + reservation rejection penalty. Then, based on the capacity boundary of the reservation allocation constraint information, the allocation variables of the site charging window time information, and all the constraints of the site-side optical-storage scheduling constraint information, a mixed integer linear or quadratic programming model is formed, and the model is then solved to obtain the target scheduling scheme.

[0127] The target scheduling scheme can include: reservation and allocation schemes; energy storage, photovoltaic, and grid power purchase and operation plans; feeder remaining safety margin; feeder capacity shadow price (marginal congestion cost); and remaining reservation slots and remaining reservation power at each site. Finally, the sites and feeders of the charging cluster can be scheduled based on the target scheduling scheme.

[0128] In one embodiment, the step of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, may include the following sub-steps: S141. Integrate the site charging window time information and the site-side photovoltaic-storage scheduling constraint information into an optimization decision variable set, and use the optimization decision variable set to construct an objective function.

[0129] S142. Using the reservation allocation constraint information as the constraint condition, the objective function is solved using a preset solution algorithm to obtain the target scheduling scheme. The preset solution algorithm includes branch and bound, cutting plane, sequential linearization, or mixed integer quadratic programming methods.

[0130] S143. Perform scheduling processing on the charging cluster sites and feeders according to the target scheduling scheme.

[0131] In one embodiment, the site charging window time information and the site-side photovoltaic-storage scheduling constraint information are integrated into a set of optimization decision variables. Specifically, the set of optimization decision variables includes: reservation request acceptance variable, site charging window time information, planned charging power variable, site-side grid power purchase, energy storage charging and discharging power, photovoltaic power absorption, and feeder group safety constraint relaxation variable.

[0132] Let the set of optimization decision variables be: ; in, Indicates a request Whether assigned to a site Time window ; Indicates a request Whether it is accepted; and These represent the planned charging power and the planned replenishment power, respectively. This indicates the amount of electricity the site purchases from the grid. and These represent the energy storage charging power and discharging power, respectively. Indicates the charge status of the station-side energy storage; Indicates the actual power absorbed by photovoltaic systems; , , and These represent the relaxation variables for the power, current, upper voltage limit, and lower voltage limit constraints of the feeder group, respectively.

[0133] The objective function is constructed using the set of optimization decision variables, and the objective function can be expressed as follows: ; in, Indicates the cost of user reservations. This indicates the station's operating costs. This indicates the cost of power distribution network security. This indicates a penalty for not accepting the appointment request.

[0134] User appointment cost is defined as: ; in, The generalized reservation cost defined in the above steps includes arrival travel time, waiting time, reservation price, feeder congestion cost, detour cost, and time window deviation cost.

[0135] The station-side operating cost is defined as: ; in, Indicates time period The electricity purchase price; Indicates site The energy storage degradation cost coefficient; Indicates the light-wasting penalty coefficient; Indicates the projected output of photovoltaic power; This indicates the time step for station-side scheduling.

[0136] The security cost of a power distribution network is defined as: ; in, , and These are the penalty coefficients for feeder group power exceeding limits, line current exceeding limits, and node voltage exceeding limits, respectively. All of the above relaxation variables satisfy the non-negativity constraint: ; The penalty for not accepting an appointment request is defined as follows: ; in, This is the penalty coefficient for not accepting an appointment request, used to prevent the system from excessively rejecting user requests when unnecessary.

[0137] The above optimization model must satisfy the following constraints.

[0138] First, the reservation request allocation constraints are: ; ; Among them, if Then request They are assigned to a single site and time window; if If not, the request is not accepted.

[0139] Secondly, the planned replenishment amount and the planned charging power must satisfy the following: ; And meet the minimum power replenishment constraints for the vehicle's subsequent travel chain: ; in, The feasibility constraints of the subsequent travel chain SOC are calculated from the station-side optical-storage scheduling constraint information obtained in the above steps. It is determined by the user's expected replenishment capacity, battery capacity, and maximum permissible SOC.

[0140] The planned charging power must also meet the power limits of the vehicle and the charging station: ; Therefore, when hour, and Automatically zero; when At that time, the request will generate a valid charging plan at the corresponding site and within the time window.

[0141] The number of reservation slots at the site is limited as follows: ; in, The number of reservation slots available for the site obtained in step S131 can be published.

[0142] The site reservation power constraint is: ; in, The site reservation power limit obtained in step S131.

[0143] The feeder group reservation capacity constraint is: ; in, Reserve the publishable capacity for the feeder group obtained in step three.

[0144] The station-side power balance constraint is: ; in, Indicates site During the scheduling period Total charging power, This indicates auxiliary load within the station. If the time window... The corresponding set of detailed scheduling periods is Then the planned power of the reservation layer and the charging power of the execution layer satisfy the following: ; in, This indicates the predicted or real-time corrected value of the load that is ready to be supplied immediately.

[0145] The photovoltaic power consumption constraint is: ; Energy storage power constraints are: ; ; The constraints for the evolution of energy storage SOC are: ; And satisfy: ; in, Indicates the energy storage capacity of the site. and These are the energy storage charging efficiency and the discharging efficiency, respectively.

[0146] The chance constraints of a feeder group can be transformed into deterministic safety constraints. Let the feeder group... During the period Basic load forecast is The mean and standard deviation of instant charging demand are respectively and Risk tolerance is Its corresponding quantile coefficient is Then the power safety constraint of the feeder group can be written as: ; in, Indicates feeder group During the period The maximum allowable power. If This indicates that the power constraint of the feeder group is strictly satisfied during this period; if If the power exceeds the limit, it indicates a risk of exceeding the power limit and will be penalized in the objective function.

[0147] If a voltage and current sensitivity model is used, the line current constraint can be written as: ; The node voltage constraint can be written as: ; ; in, Indicates site The impact of grid-side power changes on branch lines Separation Current sensitivity; Indicates site Network-side power changes affect nodes Separation Voltage sensitivity.

[0148] Therefore, after linearizing the voltage and current sensitivity, the above model can be reorganized into a mixed integer linear programming problem or a mixed integer quadratic programming problem. If the complete three-phase AC power flow equation is used, sequential linearization can be performed near the predicted operating point, and the problem can be solved iteratively using a rolling optimization method.

[0149] The specific solution process is as follows.

[0150] The first step is data input. The system reads the set of appointment requests within the forecast period. Site Collection Feeder group collection Time window set The website can publish the number of reservation slots. Site reservation power limit Feeder group reservation can publish capacity Base load forecast, on-demand charging demand forecast, photovoltaic forecast, initial SOC of station-side energy storage, and electricity price information.

[0151] The second step is preliminary screening of candidate solutions. Based on the user's earliest arrival time, latest departure time, accessibility of candidate stations, minimum charging capacity required for subsequent travel chains, and remaining reserved capacity at the stations, charging window time information for stations that do not meet the constraints of time, distance, SOC, or capacity is eliminated, forming a feasible candidate set. .

[0152] The third step is model building. , , , , , Using variables such as , construct the above objective function and constraint set.

[0153] The fourth step is to optimize the solution. The model is solved using branch and bound, cutting plane, sequential linearization, or mixed-integer quadratic programming methods to obtain the optimal solution or a feasible solution that meets engineering accuracy requirements. ; Step 5, Safety Check. Check for over-limit slack variables in the optimal solution. If satisfied: ; This indicates that the reservation allocation result and the station-side scheduling result meet the power, current, and voltage safety constraints of the feeder group. If any slack variable is greater than zero, it means that the current reservation scheme has a risk of exceeding the limit. The system increases the corresponding penalty coefficient, or performs rejection, postponement, reassignment, or reduction processing on low-priority requests, and solves the problem again.

[0154] Step 6: Rolling Execution. The system only executes the scheduled reception and power plan within the current time window or the current control cycle, and updates the model and solves it again in the next cycle based on the actual arriving vehicles, immediate charging demand, photovoltaic output, base load, and feeder operating status.

[0155] After solving the problem, the following calculation results were obtained.

[0156] First, the appointment confirmation result: ; Second, the site-time window allocation results: ; This results in a set of reservation plans: ; Third, the remaining reservation slots and remaining reservation capacity at the site: ; ; Fourth, station operation plan: ; The results are used to guide the station-side energy manager in performing grid purchase, energy storage charging and discharging, photovoltaic consumption, and charging pile power allocation.

[0157] Fifth, safety margin of feeder group: ; in, This indicates that the feeder group still has a remaining safety power margin during this period; if This indicates a risk of exceeding power limits, and is determined by slack variables. Record the level of risk.

[0158] Sixth, the shadow price of feeder group capacity or the marginal price of congestion. If the model uses continuous relaxation or sensitivity approximation, the dual variable corresponding to the feeder group reserved capacity constraint can be used as the shadow price of capacity: ; This value represents the feeder group. In the time window The marginal congestion cost of the unit reservation capacity. If the dual variable is not used directly, it can also be approximated using the finite difference method: ; in, This represents a minor capacity fluctuation. The shadow price will be used in subsequent steps to construct the reservation price and capacity reservation price.

[0159] Therefore, this paper not only presents the overall objective function for reservation linkage but also clarifies the model's decision variables, constraint set, solution process, and calculation results. Its output target scheduling scheme includes: reservation allocation results, planned charging power, station-side operation plan, feeder group safety margin, remaining reserved capacity, and capacity shadow price. Scheduling control is then implemented based on the target scheduling scheme.

[0160] It should be noted that the objective function is essentially a mixed integer rolling optimization problem containing both binary and continuous variables. After determining the overall objective function, we do not merely stop at the definition level. Instead, we unify the reservation allocation constraints (site reservation quotas, site reservation power limits, and feeder group reservation publishable capacity) with the user travel chain SOC constraints and site-side photovoltaic-storage scheduling constraints, and then solve for the reservation request reception results, site-time window allocation results, planned charging power, site-side energy storage charging and discharging plan, photovoltaic consumption plan, feeder group remaining safety margin, and capacity shadow price.

[0161] During the solution process, a rolling time-domain optimization approach can be employed. Within each control cycle, the system first reads the predicted demand, user reservation requests, feeder group safety capacity envelope, remaining reservation slots at each site, and site-side resource status. Then, a mixed-integer programming model is constructed. This model is then solved using branch and bound, cutting plane, sequential linearization, or mixed-integer quadratic programming methods. If all excess relaxation variables in the solution are zero, the reservation scheme satisfies the distribution network safety constraints. If excess relaxation variables exist, the system re-solves the problem by rejecting low-priority requests, adjusting the time window, reducing the issued capacity, calling energy storage support, or triggering a reassignment strategy. The final optimal solution is used not only to determine whether the current reservation request is accepted but also for subsequent steps involving reservation price and congestion price adjustments, as well as rolling adjustments and cross-site reassignment.

[0162] In one operating mode, the solution process is as follows: Taking into account user reservation costs, station-side operating costs, distribution network security costs, and penalties for unfulfilled reservations, the overall optimization objective function is defined as follows: ; in: ; ; ; The cost of power distribution network security is: ; in, , , , These represent the relaxation variables for the feeder group power, branch current, upper voltage limit, and lower voltage limit constraints, respectively.

[0163] The power opportunity constraint of the feeder group is: ; in, This is for random loads that are charged as soon as they arrive. If Approximately follows the mean Standard deviation is If the distribution is such that the chance constraint can be transformed into a deterministic constraint: ; in, Confidence level The quantile coefficients below.

[0164] Branch current constraints are: ; The upper limit constraint on node voltage is: ; The node voltage lower limit constraint is: ; All slack variables satisfy: ; ; ; ; If all slack variables in the solution are zero, it means that the reservation plan and station-side scheduling meet the safety constraints of feeder group power, branch current and node voltage under the set confidence level; if any slack variable is greater than zero, the system determines that there is a risk of exceeding the limit and corrects it by increasing the congestion price, tightening the reservation capacity, calling up station-side energy storage, reassigning low-priority reservations or performing derated charging.

[0165] In one embodiment, after the steps of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, the method may further include the following sub-steps: S21. Extract the feeder congestion coefficient and capacity shadow price from the target scheduling scheme, and construct a reservation price model using the feeder congestion coefficient and the capacity shadow price.

[0166] S22. Calculate the dynamic price using the reservation price model, and update the preset reservation cost function using the dynamic price.

[0167] To transform the safety constraints of the distribution network and the reserved capacity status of the substations into adjustment signals that are perceptible to users and executable by the platform, this invention further establishes a reservation price, a capacity reservation price, and a performance incentive correction mechanism. This mechanism does not manage prices in isolation, but rather maps the feeder group safety capacity, substation reserved capacity, substation occupancy rate, user no-show risk, and time flexibility obtained in the aforementioned steps into economic signals, which are then fed back into subsequent reservation request allocation, rolling correction, and cross-substation reassignment processes.

[0168] Specifically, the sites obtained in step S131 can publish reservation slots. Site reservation power limit Remaining reservation slots at the site and remaining reservation power at the site Used to measure the site In the time window The degree of urgency in booking; the shadow price of the feeder group capacity obtained by solving step S14. and the remaining safety margin of the feeder group Used to measure feeder groups In the time window The scarcity of capacity. These factors together determine the reservation price and the capacity reservation price.

[0169] First, define the site. In the time window The reservation occupancy rate is: ; in, Indicates site In the time window Number of reservations accepted Indicates site In the time window The planned power has been reserved. It is a very small positive number. This formula considers both the occupancy rate of reservation slots and the occupancy rate of reservation power. When The closer the value is to 1, the more limited the reservation resources are for that time window.

[0170] Define feeder group In the time window The capacity congestion coefficient is: ; in, This represents the remaining safety power margin of the feeder group obtained after solving step S14. Indicates feeder group In the time window The available capacity for reservations can be released. To avoid outliers, it can be limited to a range. Inside: ; when The closer the value is to 1, the closer the feeder group is to the safe capacity boundary within that time window, and the greater the impact of new reservations on distribution network security.

[0171] If step S14 is solved using a mixed-integer optimization model, then the shadow price or congestion marginal price corresponding to the feeder group reservation capacity constraint can be expressed as: ; Its meaning is feeder group In the time window The marginal change in the system objective function caused by changes in the reservable capacity of internal units. If the dual variable is not directly output in the optimization model, it can be approximated using the finite difference method: ; in, This represents the optimal objective function value of the optimization model in step S14 under the corresponding feeder group reserved capacity. This indicates a small capacity disturbance. This indicates the length of the time window. The larger this value, the scarcer the feeder capacity is within that time window, and the higher the marginal cost to the system caused by users reserving that capacity.

[0172] Define request The risk indicator for no-shows is: ; in, This represents information such as the user's historical fulfillment records, advance booking time, booking time period, historical cancellation rate, and vehicle arrival trajectory. If a probabilistic model is not used, it can also be represented using historical statistics: ; in, , , These represent the number of times a user has canceled, been late, or missed an appointment in their history. This indicates the total number of reservations made by the user throughout their history.

[0173] Define request For time windows The time elasticity index is: ; in, Indicates a request On the site Acceptable set of time windows This represents the total number of booking time windows on the platform. The larger this indicator is, the more time windows users can accept, and the stronger the platform's ability to adjust to off-peak periods.

[0174] Based on the above indicators, define the site. In the time window For the request The reservation price is: ; in, Indicates site In the time window Basic charging reservation price; This indicates the minimum allowed booking price; Indicates site Belongs to the feeder group; to This is the price adjustment factor. The formula indicates that when the feeder group capacity shadow price is high, the site reservation occupancy rate is high, the feeder congestion level is high, or the risk of users defaulting on reservations is high, the reservation price increases; when users have strong time flexibility and are willing to accept off-peak time windows, the reservation price decreases.

[0175] If request Assigned to a site Time window The reservation fee can be expressed as: ; in, This indicates the planned replenishment power determined in step S14.

[0176] Further define the capacity reservation price. The capacity reservation price measures the marginal system cost incurred by a reservation request occupying the reserveable capacity of the feeder group within a specific time window. If the request... On the site Time window The planned charging power within is The corresponding capacity reservation fee is: ; in, The length of the reservation time window is defined by this formula. This shows that for the same charging demand, if scheduled within a time window with sufficient feeder capacity, the capacity reservation cost is lower; conversely, if scheduled within a time window near the feeder's capacity limit, the capacity reservation cost is higher. Therefore, the safety margin on the distribution network side is determined by... This translates into economic signals that are perceptible to the user.

[0177] Considering the impact of user fulfillment behavior on the effective utilization rate of reservation capacity, the user reservation deposit is defined as: ; in, Based on the basic deposit ratio, and These are the deposit adjustment coefficients corresponding to the risk of no-shows and the degree of feeder congestion, respectively. This formula indicates that when a user has a high historical risk of no-shows, or when their reservation occupies peak capacity in a congested feeder network, the deposit ratio should be increased to reduce false reservations and wasted peak capacity.

[0178] If the user arrives at the designated location as agreed and completes charging within the allowable deviation range, the deposit will be fully refunded: ; If a user cancels, is late, overstays their reservation time, or fails to arrive at their destination, a portion of their deposit will be deducted. The deducted amount can be expressed as follows: ; in, , , , These are indicator variables representing user cancellation, lateness, failure to arrive, and timeout. to The above coefficients are the corresponding deduction factors. These coefficients can be further adjusted based on the amount of time allowed for cancellation, the duration of lateness, and the degree of capacity shortage.

[0179] To encourage users to accept low-congestion sites or low-congestion time windows, the time elasticity incentive is defined as follows: ; in, This is the incentive factor. If the user accepts the system's recommended low-congestion time window or alternative site, this incentive can be deducted from the reservation fee. Therefore, the user's final reservation payment amount can be expressed as: ; And it satisfies: ; The aforementioned reservation price, capacity reservation price, and incentive adjustment results will be fed back into the generalized reservation cost function in the above steps. Specifically, the allocation cost of the site charging window time information in step S132 can be updated as follows: ; in, Calculated in this step, It can be determined by the congestion coefficient of the feeder group Or capacity shadow price The calculation yields the result. Therefore, the reservation price is no longer an external parameter independent of the optimization model, but is determined by feeder capacity, site occupancy, user risk, and time flexibility, and directly participates in subsequent reservation allocation.

[0180] During the rolling optimization process, the above steps S131-133 provide , , and This is used to calculate the site reservation occupancy rate; step S14 provides... and This is used to calculate capacity reservation price and feeder congestion level; the above results can be transformed into... , , and These price and incentive signals are then fed back into the reservation request allocation model of the above steps, and used to recalculate the combined cost of candidate alternative sites and alternative time windows when feeder margin changes or cross-site reassignment occur in subsequent steps.

[0181] Thus, the reservation price, capacity reservation price, and incentive adjustment mechanism realize the transformation from "distribution network physical security constraints" to "user-side reservation behavior guidance". This allows feeder group capacity shortages, site resource occupation, and user performance risks to enter the reservation decision-making process in the form of prices, deposits, and incentives, thereby improving the utilization rate of reserved capacity and reducing the risks of false reservations and local feeder over-limits during peak hours.

[0182] The purpose of this step, involving reservation pricing and capacity reservation pricing, is to transform the feeder group safety margin, remaining reserved capacity at the site, feeder congestion level, and user fulfillment risk obtained in the previous steps into economic signals that users can perceive. Its aim is not to simply add a charging step, but rather to guide users to avoid periods of feeder capacity shortages, limited site reservation slots, and peak congestion times through pricing and deposit mechanisms, thereby aligning user-side reservation behavior with distribution network safety constraints.

[0183] Specifically, step S131 generates the feeder group's publishable reservation capacity, the site's reservation power limit, and the site's publishable reservation quota; step S132's generalized reservation cost function needs to call the site-time window reservation price to compare different candidate sites and different candidate time windows; after solving the reservation linkage overall objective function in step S14, the feeder group's capacity shadow price, the site's remaining reservation power, the site's remaining reservation quota, and the feeder safety margin can be obtained. These results are further used to correct the reservation price, capacity reservation price, and deposit coefficient.

[0184] In one embodiment, after the steps of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, the method may further include the following sub-steps: S31. Obtain real-time arrival data and adjust the charging power parameters based on the real-time arrival data.

[0185] During operation, demand forecasts can be revised based on actual arrival data. ; in, The charging power parameters before correction. The corrected charging power parameters. To measure the actual arrival demand, This is for updating coefficients.

[0186] When the safety capacity of the feeder group changes abruptly, update: ; in, This could be caused by factors such as increased base load, localized faults, sudden drops in photovoltaic power, or unavailability of on-site energy storage. If the following conditions are met: ; This triggers a re-optimization or emergency reassignment. At this point, priority is calculated for reservation requests that have not yet started charging: ; in, Indicates whether it is a high-priority vehicle. Weighted. System priority protection. Higher requests, for For lower-value, flexible rescheduling requests, cross-site reassignment or postponement is performed. Cross-site reassignment can be represented as: ; in, The remaining reservation capacity at the station. If no feasible alternative station is available, the station will enter a reduced or partially charged mode to at least meet the minimum SOC requirement for future travel.

[0187] Compared to existing prediction methods that primarily focus on historical sequences of single sites, this invention first introduces "cross-site knowledge transfer" into the charging station reservation release process. Recent preprints have shown that for new stations and stations with few samples, few-shot prediction and transfer learning significantly outperform the approach of treating each station in isolation; however, most of these works stop at "more accurate predictions." This invention further transforms the prediction results directly into publishable reservation capacity, reservation price, and joint optimization inputs for station clusters and feeder clusters, thereby truly transforming "predictive capability" into "reservation control capability."

[0188] Compared to existing solutions that primarily focus on charging station site selection or grid impact assessment, this invention does not determine whether a station should be built at the planning level, but rather answers the question at the operational level: "How to issue reservations, allocate users, and coordinate energy storage and photovoltaics under multiple feeder constraints for an already built multi-station cluster." Grid-ECO and related high-fidelity simulations have shown that voltage and current sensitivity, as well as distribution network physical constraints, are crucial for site planning and impact analysis. This invention further extends this "grid-side sensitivity" from planning analysis to the reservation and operation stage, making the feeder cluster safety margin a direct determinant of reservation slots and reserved power.

[0189] Compared with existing in-station EMS or real-time charging control schemes, the innovation of this invention lies in extending the control boundary from a single station to a two-layer collaboration of "station group - feeder group". Existing work uses hierarchical EMS, distributed optimization and incentive design to manage multiple charging points, energy storage and photovoltaic resources within a single large charging station; this invention adds upper-level reservation linkage and cross-station reassignment mechanisms on this basis, so that in-station optimization is no longer an isolated occurrence, but serves the global reservation fulfillment under the safety boundary of the entire feeder group.

[0190] Compared to mechanisms like reserve-based tariffs that address dynamic congestion, this invention is not limited to the idea of ​​"pre-reserving flexibility contracts." Instead, it couples capacity reservation with user bookings, vehicle travel constraints, station-side energy storage, and station cluster reassignment simultaneously. Existing reserve-based mechanisms emphasize obtaining flexibility for short-term congestion through pricing and reservations; this invention further refines "reservation" into reserved capacity for specific stations, specific time windows, and specific user requests, and utilizes mobility-aware constraints to support partial charging, deferral, and cross-station substitution, thus making it more suitable for the engineering operation scenarios of public charging networks.

[0191] In this embodiment, the present invention provides a joint scheduling method for feeders and charging clusters. Its advantages are as follows: The present invention can construct a transfer learning prediction model based on the feature information of charging stations, and use the transfer learning prediction model to determine demand prediction parameters; determine the safe power envelope of the feeder cluster throughout the entire time period using the demand prediction parameters as constraints; determine the reservation allocation constraint information for charging based on the safe power envelope; determine the station charging window time information using charging reservation requests and reservation allocation constraint information; determine the station-side photovoltaic-storage scheduling constraint information based on the station charging window time information; construct an objective function based on the reservation allocation constraint information, station charging window time information, and station-side photovoltaic-storage scheduling constraint information, and solve to obtain the target scheduling scheme; and perform scheduling processing on the stations and feeders of the charging cluster based on the target scheduling scheme. The present invention determines the demand prediction parameters through the feature information of charging stations, and then combines the demand prediction parameters and charging reservation requests to construct an objective function to obtain the scheduling scheme, thereby conforming to the actual needs of charging stations and charging vehicles and improving scheduling accuracy.

[0192] This invention also provides a joint scheduling device for feeders and charging clusters, see [link to relevant documentation]. Figure 2 The diagram shows a structural schematic of a joint scheduling device for feeders and charging clusters provided in an embodiment of the present invention.

[0193] As an example, the joint scheduling device for the feeder and charging cluster may include: The prediction determination module 201 is used to construct a transfer learning prediction model based on the feature information of the charging station, and to determine the demand prediction parameters using the transfer learning prediction model. The demand prediction parameters include: charging power parameters and feeder group load parameters. The power determination module 202 is used to determine the safe power envelope of the feeder group for the entire time period based on the demand forecast parameters as constraints, wherein the safe power envelope is the upper limit of the maximum total charging power allowed by the distribution network. Information determination module 203 is used to determine the charging reservation allocation constraint information based on the safe power envelope, determine the site charging window time information using the charging reservation request and the reservation allocation constraint information, and determine the site-side photovoltaic-storage scheduling constraint information according to the site charging window time information. The scheduling module 204 is used to construct an objective function and solve it to obtain a target scheduling scheme based on the reservation allocation constraint information, the site charging window time information and the site-side optical storage scheduling constraint information, and to perform scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme.

[0194] Optionally, the step of constructing a transfer learning prediction model based on the feature information of the charging station, and using the transfer learning prediction model to determine the demand prediction parameters, includes: Obtain feature information of the charging station, and construct a transfer learning prediction model based on the feature information; The transfer learning prediction model is invoked to predict charging power parameters, wherein the charging power parameters include: reserved power demand parameters and immediate power demand parameters; The charging power parameters are equivalently converted based on the preset time window length to obtain the feeder group load parameters.

[0195] Optionally, determining the safe power envelope of the feeder group throughout the entire time period using the demand forecast parameters as constraints includes: The transformer margin, current margin, and voltage margin of the feeder group are determined respectively using the aforementioned demand forecast parameters as constraints. The safe power envelope to be released by the feeder group at a future scheduled time is determined using the transformer margin, the current margin, and the voltage margin.

[0196] Optionally, the step of determining the charging reservation allocation constraint information based on the safe power envelope, determining the site charging window time information using the charging reservation request and the reservation allocation constraint information, and determining the site-side photovoltaic-storage scheduling constraint information based on the site charging window time information includes: The available capacity for reservation is calculated using the safe power envelope and the preset reservation release coefficient. The upper limit of the reservation power is determined based on the available capacity for reservation and the preset station-level weight to obtain reservation allocation constraint information. The preset station-level weight is determined based on the charging power parameters. Obtain charging reservation requests and candidate charging stations respectively. Construct a reservation time window set based on the charging reservation requests and candidate charging stations. Divide the reservation time window set according to a preset reservation cost function with the reservation allocation constraint information as the constraint condition to obtain the station charging window time information. Based on the charging reservation request, the range of charging power and the number of reservations within the charging window time information of the site are determined, and the site-side photovoltaic-storage scheduling constraint information is obtained.

[0197] Optionally, the step of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side photovoltaic-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, includes: The site charging window time information and the site-side photovoltaic-storage scheduling constraint information are integrated into an optimization decision variable set, and the objective function is constructed using the optimization decision variable set. Using the reservation allocation constraint information as constraints, a preset solution algorithm is used to solve the objective function to obtain the target scheduling scheme. The preset solution algorithm includes branch and bound, cutting plane, sequential linearization, or mixed integer quadratic programming methods. The charging cluster's sites and feeders are scheduled according to the target scheduling scheme.

[0198] Optionally, the device further includes: The extraction module is used to extract the feeder congestion coefficient and capacity shadow price from the target scheduling scheme, and to construct a reservation price model using the feeder congestion coefficient and the capacity shadow price; The update module is used to call the reservation price model to calculate the dynamic price and use the dynamic price to update the preset reservation cost function.

[0199] Optionally, the device further includes: An adjustment module is used to acquire real-time arrival data and adjust the charging power parameters based on the real-time arrival data.

[0200] Those skilled in the art will understand that, for ease of description and brevity, the specific working process of the device described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0201] Furthermore, this application also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the joint scheduling method of feeder and charging cluster as described in the above embodiments.

[0202] Furthermore, embodiments of this application also provide a computer-readable storage medium storing a computer-executable program for causing a computer to execute the feeder and charging cluster joint scheduling method as described in the above embodiments.

[0203] In the description of the embodiments of the present invention, it should be noted that the terms "above," "below," etc., indicating the orientation or positional relationship are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the embodiments of the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. When an element such as a layer, region, or substrate is referred to as being "above" or "on top of" another element, it may be directly on the other element, or there may be an intermediate element. Conversely, when an element is referred to as being "directly on" or "above" another element, there is no intermediate element. It should also be understood that when an element is referred to as being "below" or "under" another element, it may be directly below or under the other element, or there may be an intermediate element. Conversely, when an element is referred to as being "directly below" or "under" another element, there is no intermediate element. Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections or electrical connections; they can refer to direct connections or indirect connections through an intermediate medium; and they can refer to the internal connection between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0204] Those skilled in the art will understand that embodiments of this application may also include computer program products. Therefore, this application may take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0205] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), devices, and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0206] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0207] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0208] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A joint scheduling method for feeders and charging clusters, characterized in that, The method includes: A transfer learning prediction model is constructed based on the feature information of the charging station, and the demand prediction parameters are determined using the transfer learning prediction model. The demand prediction parameters include: charging power parameters and feeder group load parameters. The safe power envelope of the feeder group is determined with the demand forecast parameters as constraints, wherein the safe power envelope is the upper limit of the maximum total charging power allowed by the distribution network. Based on the safety power envelope, the charging reservation allocation constraint information is determined. The charging reservation request and the reservation allocation constraint information are used to determine the site charging window time information. The site-side photovoltaic-storage scheduling constraint information is determined based on the site charging window time information. Based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information, an objective function is constructed and solved to obtain the target scheduling scheme. Based on the target scheduling scheme, the sites and feeders of the charging cluster are scheduled.

2. The joint scheduling method for feeders and charging clusters according to claim 1, characterized in that, The step of constructing a transfer learning prediction model based on the feature information of charging stations, and using the transfer learning prediction model to determine demand prediction parameters, includes: Obtain feature information of the charging station, and construct a transfer learning prediction model based on the feature information; The transfer learning prediction model is invoked to predict charging power parameters, wherein the charging power parameters include: reserved power demand parameters and immediate power demand parameters; The charging power parameters are equivalently converted based on the preset time window length to obtain the feeder group load parameters.

3. The joint scheduling method for feeders and charging clusters according to claim 2, characterized in that, The process of determining the safe power envelope of the feeder group throughout the entire time period using the demand forecast parameters as constraints includes: The transformer margin, current margin, and voltage margin of the feeder group are determined respectively using the aforementioned demand forecast parameters as constraints. The safe power envelope to be released by the feeder group at a future scheduled time is determined using the transformer margin, the current margin, and the voltage margin.

4. The joint scheduling method for feeders and charging clusters according to claim 2, characterized in that, The process of determining charging reservation allocation constraint information based on the safe power envelope, determining site charging window time information using the charging reservation request and the reservation allocation constraint information, and determining site-side photovoltaic-storage scheduling constraint information based on the site charging window time information includes: The available capacity for reservation is calculated using the safe power envelope and the preset reservation release coefficient. The upper limit of the reservation power is determined based on the available capacity for reservation and the preset station-level weight to obtain reservation allocation constraint information. The preset station-level weight is determined based on the charging power parameters. Obtain charging reservation requests and candidate charging stations respectively. Construct a reservation time window set based on the charging reservation requests and candidate charging stations. Divide the reservation time window set according to a preset reservation cost function with the reservation allocation constraint information as the constraint condition to obtain the station charging window time information. Based on the charging reservation request, the range of charging power and the number of reservations within the charging window time information of the site are determined, and the site-side photovoltaic-storage scheduling constraint information is obtained.

5. The joint scheduling method for feeders and charging clusters according to claim 4, characterized in that, The step of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side photovoltaic-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, includes: The site charging window time information and the site-side photovoltaic-storage scheduling constraint information are integrated into an optimization decision variable set, and the objective function is constructed using the optimization decision variable set. Using the reservation allocation constraint information as constraints, a preset solution algorithm is used to solve the objective function to obtain the target scheduling scheme. The preset solution algorithm includes branch and bound, cutting plane, sequential linearization, or mixed integer quadratic programming methods. The charging cluster's sites and feeders are scheduled according to the target scheduling scheme.

6. The joint scheduling method for feeders and charging clusters according to any one of claims 1-5, characterized in that, After the steps of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, the method further includes: Extract the feeder congestion coefficient and capacity shadow price from the target scheduling scheme, and construct the reservation price model using the feeder congestion coefficient and the capacity shadow price; The reservation price model is invoked to calculate the dynamic price, and the preset reservation cost function is updated using the dynamic price.

7. The joint scheduling method for feeders and charging clusters according to any one of claims 1-5, characterized in that, After the steps of constructing and solving an objective function based on the reservation allocation constraint information, the site charging window time information, and the site-side optical-storage scheduling constraint information to obtain a target scheduling scheme, and then performing scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme, the method further includes: Obtain real-time arrival data and adjust the charging power parameters based on the real-time arrival data.

8. A joint scheduling device for feeders and charging clusters, characterized in that, The device includes: The prediction determination module is used to construct a transfer learning prediction model based on the feature information of the charging station, and to determine the demand prediction parameters using the transfer learning prediction model. The demand prediction parameters include: charging power parameters and feeder group load parameters. The power determination module is used to determine the safe power envelope of the feeder group for the entire time period based on the demand forecast parameters as constraints, wherein the safe power envelope is the upper limit of the maximum total charging power allowed by the distribution network. The information determination module is used to determine the charging reservation allocation constraint information based on the safe power envelope, determine the site charging window time information using the charging reservation request and the reservation allocation constraint information, and determine the site-side photovoltaic-storage scheduling constraint information based on the site charging window time information. The scheduling module is used to construct an objective function and solve it to obtain a target scheduling scheme based on the reservation allocation constraint information, the site charging window time information and the site-side optical-storage scheduling constraint information, and to perform scheduling processing on the sites and feeders of the charging cluster based on the target scheduling scheme.

9. An electronic device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the program, it implements the joint scheduling method for feeders and charging clusters as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to perform the joint scheduling method for feeders and charging clusters as described in any one of claims 1-7.

Images