A charging station group collaborative scheduling method supporting multi-market participation

Abstract

The application discloses a kind of support multi-market participation's charging station group collaborative scheduling method, it is related to power system dispatching technical field, including: by exerting power disturbance, initial evaluation parameter is constructed;Update evaluation parameter, and generate scheduling unit in combination with power-time decoupling model;When the execution time of scheduling unit overlaps, determine the coupling influence offset between each scheduling unit according to the updated evaluation parameter, and identify the space-time occupation segment of each scheduling unit;According to the space-time occupation segment and coupling influence offset, adjust scheduling unit, and form executable unit output that satisfies power grid bearing constraint.It solves the problem that nonlinear coupling influence cannot be dynamically quantified in existing charging station group scheduling, and charging request cannot be identified space-time occupation and closed-loop adjustment.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching technology, and more specifically, to a method for coordinated dispatching of charging station clusters that supports multi-market participation. Background Technology

[0002] Currently, aggregation platforms are often used to guide charging station clusters to collaboratively participate in grid dispatch and multi-level electricity markets. However, when faced with the need for large-scale cluster response, existing conventional aggregation dispatch solutions have the following technical shortcomings: 1. Inability to dynamically assess the electrical impact on the power grid: The existing dispatching relies on offline static parameters and linear superposition to calculate the power grid carrying capacity, failing to consider dynamic factors such as topology changes and operating point drift during the response of the substation group, and cannot update the assessment parameters online based on measured power disturbances.

[0003] 2. Inability to quantify the offset of coupling effects between overlapping loads: When the execution times of multiple charging requests overlap, due to the electrical coupling between grid nodes and lines, the actual impact of the superposition of the requested power deviates from the linear superposition result. Existing solutions lack a method to quantify this nonlinear coupling offset.

[0004] 3. Unable to identify the spatiotemporal occupancy of charging requests and perform closed-loop adjustments: Existing methods do not jointly model the temporal dimension of charging requests and the spatial dimension of grid nodes-line, and cannot identify the occupancy of each request in time and space, nor can they adjust conflicting requests based on occupancy and coupling offset (such as translation, power reduction, splitting and merging). Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a collaborative scheduling method for charging station groups that supports multi-market participation, thereby solving the technical problems in existing charging station group scheduling that cannot dynamically quantify the effects of nonlinear coupling and cannot identify the spatiotemporal occupancy of charging requests and perform closed-loop adjustments.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A method for coordinated scheduling of charging station clusters supporting multi-market participation includes the following steps: This invention first constructs initial evaluation parameters by acquiring grid electrical quantities and charging gun power and applying actual power perturbations, overcoming the distortion of underlying data caused by unknown local topology or static modeling errors, and providing high-precision dynamic benchmark data for subsequent scheduling. Subsequently, the system updates the aforementioned evaluation parameters in real time based on the online identification results of grid topology and line parameters, giving the system robustness against grid abrupt changes. Furthermore, combined with a power-duration decoupling model, the complex nonlinear battery electrochemical characteristics are reduced in dimensionality and transformed into standardized scheduling units with clear boundaries that can be directly scheduled. Further, when the execution time of each scheduling unit overlaps... During the stacking process, the system accurately calculates the coupling effect offset between each unit based on the updated evaluation parameters, and identifies its spatiotemporal occupancy segments in the time dimension and the grid node-line space. This allows for the advanced prediction and quantification of the nonlinear power flow superposition effect under high-concurrency response scenarios. Finally, based on the spatiotemporal occupancy segments and coupling effect offset, the system performs reverse dynamic adjustments on conflicting scheduling units, generating executable unit outputs that strictly meet grid carrying capacity constraints. This completely eliminates the safety hazard of the response command being disconnected from the underlying physical limits of the grid, and realizes safe, stable, and refined closed-loop scheduling of the distribution network under massive vehicle access.

[0007] The technical effects and advantages of the charging station cluster collaborative scheduling method supporting multi-market participation of the present invention are as follows: This invention applies power perturbation to the charging gun and combines it with online identification of power grid topology and line parameters to dynamically update evaluation parameters. This allows for real-time quantitative characterization of the actual impact of charging load on node voltage and line current carrying capacity, overcoming the evaluation distortion problem caused by static models and realizing dynamic online evaluation of the electrical impact on the power grid.

[0008] In cases of overlapping time among multiple scheduling units, the coupling effect offset is extracted by comparing the actual electrical response with the linear superposition prediction results. The occupied segments of each unit in the time dimension and node-line space are also identified. This solves the problem that multi-load superposition cannot reflect actual electrical coupling changes, enabling the quantification of nonlinear coupling offset of overlapping loads. Furthermore, based on the deviation between the effect occupancy and grid carrying capacity constraints, a standardized dispatchable capacity basis is provided for multi-market participation, enabling closed-loop adjustment of conflicting loads and forming executable units. Attached Figure Description

[0009] Figure 1 A schematic diagram of a collaborative scheduling method for charging station clusters that supports multi-market participation.

[0010] Figure 2 This is a schematic diagram comparing the actual electrical response with the predicted response. Detailed Implementation

[0011] 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 of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0012] Example 1 Please see Figure 1 To address the problems in existing charging station cluster scheduling, such as the inability to dynamically quantify the effects of nonlinear coupling and the inability to identify the spatiotemporal occupancy structure of charging requests and perform closed-loop adjustments, this invention presents a collaborative scheduling method for charging station clusters that supports multi-market participation, comprising the following steps: S1: Obtain the electrical quantities of the power grid and the power of the charging gun, and construct the initial evaluation parameters by applying power perturbation; S2, based on the online identification results of power grid topology and line parameters, updates the evaluation parameters and generates a scheduling unit by combining the power-duration decoupling model; S3, when the execution times of scheduling units overlap, determine the coupling effect offset between scheduling units based on the updated evaluation parameters, and identify the spatiotemporal occupancy segments of each scheduling unit in the time dimension and the grid node-line space. S4. Based on the spatiotemporal occupancy segment and coupling influence offset, the scheduling unit is adjusted to form an executable unit output that satisfies the power grid carrying capacity constraints.

[0013] In one embodiment, S1 includes: S11, obtain the electrical quantities of the power grid and the output power of the charging gun; among which the electrical quantities of the power grid include node voltage (including voltage amplitude and node voltage phase angle) and line current amplitude; The electrical quantities of the power grid are obtained through the feeder terminal units (FTU), substation terminal units (DTU), and the main station SCADA system in the distribution automation system; the output power of the charging gun is obtained in real time by the charging pile controller; the sampling period of the charging gun output power is 100ms to 1s, preferably 200ms; the node voltage amplitude is measured using a voltage transformer with an accuracy of 0.2; the node phase angle is obtained through a synchronous phasor measurement device (PMU), with a sampling frequency ranging from 20Hz to 100Hz, preferably 50Hz, to ensure the observability of phase angle changes; the line current is collected by a smart current sensor with an accuracy of preferably 0.5.

[0014] S12, apply power disturbance and record the change in node voltage phase angle and the change in line current amplitude; Apply a power disturbance to the target charging gun during operation (e.g., issue a step command with a disturbance duration of 5 seconds and an amplitude of 5% of the rated power of the corresponding charging gun). During the execution of this power disturbance, in order to ensure the spatiotemporal consistency of the measurement data, the system calculates the difference between the node voltage phase angle and the line current amplitude at the start time of the applied power disturbance and the end time of the applied power disturbance based on a unified clock synchronization mechanism (such as GPS / BeiDou time synchronization). This difference is the measured change in node voltage phase angle and the change in line current amplitude. Meanwhile, considering the execution deviation of the equipment response, the actual output power of the charging pile at the moment before the disturbance is applied is extracted as the benchmark, and the actual response power of the charging pile during the disturbance is collected. The difference between the two is used as the power change of the disturbance.

[0015] S13, Construct initial evaluation parameters including node voltage phase angle evaluation parameters and line current evaluation parameters; The formula for the initial evaluation parameters is expressed as follows: (1) (2) As shown in the above formula, It serves as a parameter for evaluating the phase angle of node voltage, used to quantify the sensitivity of active power fluctuations to local node voltage phase angle shifts; These are parameters for evaluating line current, used to quantify the impact mapping relationship between active power fluctuations and the current carrying capacity of associated transmission channels; This refers to the change in the phase angle of the node voltage. This refers to the change in the amplitude of the line current. This represents the change in power. The two calculated parameters are packaged and output together to obtain the initial evaluation parameters.

[0016] In one embodiment, S2 includes: S21, based on power grid electrical quantities, identifies the topology and line parameters between nodes online; The electrical quantities of the power grid obtained in step S11 are cached within a set sliding time window to construct a historical time-series data sequence of voltage and current. Subsequently, the historical time-series data sequence is fitted and calculated using recursive least squares (RLS) with a forgetting factor, and the branch admittance correction values ​​between nodes are output in real time. Furthermore, the physical connection status between nodes is determined by judging whether the branch admittance is lower than a set connectivity threshold, thereby completing the online identification of power grid topology and line parameters. The sliding time window is preset to 15 minutes.

[0017] S22, calculate power flow evaluation parameters based on the identified topology and line parameters, and use them to update the evaluation parameters; Based on the real-time identified and updated power grid topology and branch admittance data in step S21, a node admittance matrix for the entire network is constructed. The current node impedance matrix of the power grid is obtained by inverting this admittance matrix (or by using sparse matrix factorization). This node impedance matrix reflects the electrical coupling strength between nodes in the power grid under the current topology.

[0018] Extract the access node number corresponding to each charging gun in the power grid topology. To evaluate the impact of the charging gun's dynamic load on the power grid safety boundary, use the node impedance matrix and access node number to calculate the power flow assessment parameters for each charging gun through power flow sensitivity calculation. The power flow assessment parameters include node power flow assessment parameters and line power flow assessment parameters.

[0019] It should be noted that the power flow assessment parameters in this embodiment are, in essence, power flow sensitivity parameters of the electrical quantities of the grid nodes to the output power of the charging gun (i.e., the active power injected into the grid nodes). These parameters characterize the static power transfer relationship under a purely mathematical topological dimension. The calculation formulas for the power flow assessment parameters are as follows: (3) and (4): (3) (4) In equation (3), Indicates access at node The first Put the charging gun into the node. Node power flow evaluation parameters; (When used as a subscript) is the charging gun access node number; Number the charging gun devices on the node; For the first The node number at one end of the transmission line; Indicates the charging gun connection node The In this embodiment, the change in power output from the charging gun is forcibly constrained in the denominator. This eliminates the division operator, making the power flow assessment parameters directly equivalent to the voltage phase angle change and the line current amplitude change, thus avoiding the division-to-zero value collapse in high-concurrency scenarios. This represents the change in the phase angle of the topological node voltage calculated based on the topology in this step. It is extracted from the complex value of the node voltage change through complex geometric operations. The extraction formula is as follows (5). In equation (4), Indicates the first Connect the charging gun to the connection node With nodes The Power flow assessment parameters for a transmission line; Number the target transmission line in the power grid topology; For the first The node number at the other end of the transmission line (one end is) ); This represents the change in the amplitude of the topological line current calculated based on the topology in this step. It is calculated by the complex difference of the voltage change at both ends of the nodes and the line impedance through the complex Ohm's law. The calculation formula is as follows (7). (When not used as a subscript) represents the imaginary unit, and the superscript * indicates conjugate complex operations.

[0020] (5) (6) (7) (8) (9) In equation (5), The complex value of the node voltage change is calculated using equation (8); The reference voltage before the disturbance is a complex value, which is constructed by synthesizing the real-time voltage amplitude and node phase angle obtained in step S101 using equation (6); In equation (6), The real-time voltage amplitude obtained in step S101 The real-time node phase angle obtained in step S101; In equation (7), and These represent the nodes generated based on topology calculations in this step. With nodes The complex value of the node voltage change is calculated using equations (8) and (9); and The equivalent resistance and reactance of the line are given (from online identification results). In equations (8) and (9), This represents the modulo operation on complex numbers; and Let represent the complex impedance elements in the node impedance matrix calculated above. They represent the slave nodes respectively. To the node and to associated nodes The topological electrical distance and physical transfer impedance.

[0021] It should be noted that, based on formulas (8) and (9), the numerator here directly calls the complex impedance elements in the static nodal impedance matrix, replacing the high-frequency dynamic Jacobian matrix inversion process in the prior art, thus reducing the computational complexity from Directly dropped to Constant level.

[0022] S23, the power flow evaluation parameters are weighted and fused with the initial evaluation parameters to obtain the updated evaluation parameters, wherein the updated evaluation parameters include the updated node evaluation parameters and the updated line evaluation parameters; In a specific embodiment (using weighted fusion processing): When the communication of front-end physical measurement devices (such as PMUs and smart terminals) is detected to be normal, and the power grid topology is in a steady state (i.e., there are no recent topology reconfiguration events such as circuit breaker tripping), the system adopts dynamic weighted fusion processing. This method aims to use the topology physical model to smooth the measurement noise in the physical measurement data, while using the physical measurement data to correct the static parameter errors in the topology model. The specific fusion calculation formula is as follows: (10) (11) In the formula, This represents the updated node evaluation parameters obtained after fusion computation; This represents the updated line evaluation parameters obtained after fusion calculation; and These represent the initial evaluation parameters of the corresponding target node and target line obtained in step S100 above; and These are confidence weight coefficients, and their values ​​range from [0,1].

[0023] In summary, by using a weighted fusion processing method, the updates to the evaluation parameters can simultaneously reflect real-time measurement data and the physical structure characteristics of the power grid. This enables the evaluation parameters to adaptively switch between "data-driven results" and "physical model results," allowing for rapid responses to changes in the power grid's operating state. Furthermore, under stable power grid operating conditions, dynamic updates to the power grid topology ensure the continued effectiveness of the modeling under complex operating conditions, improving the accuracy of the modeling and the stability of subsequent scheduling calculations.

[0024] S24, combining power inflection point, duty cycle and power-duty decoupling model, generates scheduling unit; After obtaining the updated initial evaluation parameters (updated node evaluation parameters and updated line evaluation parameters), to achieve flexible scheduling of charging load, the charging process during the decay phase needs to be abstracted into a structured scheduling unit with a clear power-time elastic boundary. The specific implementation process is as follows: Data preprocessing stage: It should be noted that, firstly, historical electricity load data and new energy output prediction data within the jurisdiction of the target substation are extracted, and conventional algorithms, including but not limited to median filtering, wavelet transform, or spline interpolation, are used to denoise and remove extreme values ​​from the above-mentioned historical electricity load data and new energy output prediction data in order to form a standard prediction curve.

[0025] S241, Construct a power-duration decoupling model; S2411, Determine the power inflection point based on the vehicle's state of charge and charging power attenuation characteristics; To address the nonlinear degradation characteristics of electric vehicles in the later stages of charging, the power inflection point, which characterizes the critical boundary between constant power and the degradation stage, is first determined. Specifically, if a vehicle model or historical charging curve is available, the system directly extracts the starting position where the power begins to continuously decrease using the characteristic line method, and uses the corresponding state of charge (SOC) as the power inflection point. The vehicle SOC is obtained through the charging pile controller or the on-board battery management system. If vehicle model information is missing, the system directly calls an empirical SOC threshold (e.g., 0.65) for substitution mapping.

[0026] Using this power inflection point as a boundary, the system compares the vehicle's current state of charge in real time: when ≥ When the vehicle enters the degradation-dominant phase, the system synchronously extracts the maximum power corresponding to each SOC level in that phase (obtained directly from the vehicle via the charging communication network) as the upper limit of the dynamic response power for subsequent scheduling. in, Assign vehicle number, For vehicles Current state of charge, The state of charge corresponding to the power inflection point; S2412, a flexible interval division based on off-site margin dual thresholds; wherein the flexible interval includes: power reduction interval and intermittent power supply interval; Traditional charging control passively executes only according to the natural decay curve, while this invention introduces off-site margin as a core indicator for flexible space partitioning in the objective function.

[0027] In this embodiment, the departure margin is the maximum allowable time budget during which the charging process can be extended, delayed, or interrupted without affecting timely departure. That is, the adjustable time limit between the current time and the expected departure time of the vehicle, after deducting the shortest charging time required to complete the target energy replenishment; the departure margin is calculated using the following formula (12): (12) (13) (14) In the formula, For vehicles Departure margin For vehicles The expected departure time, For the current moment, The shortest charging time is the shortest continuous time required to replenish the remaining energy demand under the maximum charging power allowed by the current device (the upper limit of the dynamic response power extracted in step S2411), representing the rigid time consumption, and its calculation formula is as shown in equation (13). In equation (13), For vehicles The remaining energy replenishment requirement is calculated using the formula (14). For charging gun The maximum charging power allowed to be output under the current vehicle and equipment conditions is the upper limit of the dynamic response power extracted in step S2411. In equation (14), For vehicles Target state of charge, The rated capacity of the target vehicle's battery.

[0028] The system sets a threshold for determining power reduction (preferably 10 minutes) and a threshold for determining intermittent charging (preferably 20 to 30 minutes) to finely define the charging process entering the degradation-dominant phase: (1) Rigid interval: when At that time, the off-site margin was severely insufficient, the charging process lacked the ability to actively adjust, and the vehicle completed charging according to the natural power curve, without forming a dispatch unit; (2) Reduced power range: when At that time, the off-site capacity margin was sufficient. The system determined that the charging power could be reduced during this period, but the power supply could not be interrupted; (3) Intermittent power supply range: when At that time, the off-site capacity was more than sufficient. The system determined that not only was a power reduction allowed during this period, but a further charging stop period could also be inserted.

[0029] in, The threshold for determining the power reduction range. The threshold for determining the intermittent power supply interval must satisfy the following conditions: In one specific embodiment, preferably Take for 5 to 15 minutes, more preferably 10 minutes; Take 15 to 40 minutes, more preferably 20 to 30 minutes, and .

[0030] S2413, Establish segmented mapping relationships to form a power-duration decoupling model; After identifying the aforementioned flexible zones (the power reduction zone and the intermittent power supply zone), the system constructs a piecewise algebraic mapping model based on the law of conservation of physical energy to establish a quantitative relationship between the charging gun's occupancy time and the power supply intensity. Specifically, for the power reduction zone, the occupancy time is mapped to a binary inverse proportional relationship; for the intermittent power supply zone, the calculation model for the occupancy time is modified to a ternary coupled relationship. Details are as follows: It should be noted that the power supply intensity mentioned in this invention, in its physical essence, refers to the target amplitude of active power output actively sent by the scheduling master station to the underlying charging pile controller and actually executed within a specific discrete charging time window.

[0031] For the power reduction range, the system forces a duty cycle that degenerates to 1, and the occupied time is mapped to a binary inverse proportional relationship: (15) in, Select the power supply intensity for the given interval The corresponding charging gun usage time; For power supply intensity, its adjustable range is calibrated as an interval. ,in The upper limit of dynamic response power extracted in step S2411; The minimum continuous charging power (preferably 10kW) is controlled by the minimum stable output capability of the vehicle charging module and the lower limit of the grid command.

[0032] For intermittently powered sections, in addition to adjustable power, the system introduces a duty cycle to characterize the proportion of effective on-time. In this case, the calculation model for the occupancy time is modified to a ternary coupling relationship: (16) In the formula, For duty cycle, satisfy The lower limit of the duty cycle value is controlled by pre-defined safety engineering parameters such as the minimum continuous charging time of the battery (generally set to 2-5 minutes in engineering) and the minimum charging stop time (generally set to 3-5 minutes in engineering). The formula for calculating the lower limit of the duty cycle value is as follows: (17) in, This indicates the lower limit of the duty cycle that can be adjusted downwards within the intermittent power supply range; This represents the function that takes the maximum value. and These represent the expected departure time and the current time of the target vehicle, respectively. This represents the total available time window at the current distance from the station; Characterized by the minimum duty cycle that must be maintained to meet the vehicle's timely departure and refueling needs; This indicates the pre-calibrated minimum continuous battery power-on time; This indicates the pre-calibrated minimum continuous charging / discharging time for the battery; Characterized by the minimum physical duty cycle that must be maintained to avoid frequent start-stop cycles damaging battery life.

[0033] Preferably, to unify the adjustment boundary on the pure vehicle side with the carrying capacity limit of the power grid layer, the system invokes the conventional Static Security Check mechanism, and combines the updated evaluation parameters to apply a second constraint on the power supply intensity at the power grid topology dimension, thus converging and correcting the final adjustable range of the power supply intensity as follows: (18) In the formula, For the ultimate adjustable range of power supply intensity, The maximum safe power injection boundary that the power grid can support is obtained through a process based on existing linearized sensitivity analysis techniques. The specific calculation formula for the maximum safe power injection boundary is as follows: (19) In the formula, The safe phase angle operating margin of the target node of the power grid is obtained by subtracting the current actual phase angle absolute value from the preset absolute value of the safe phase angle limit of the target node. The safety current-carrying margin of the grid-connected lines is obtained by subtracting the current actual current-carrying capacity from the maximum rated current-carrying capacity allowed by the thermal stability of the connected lines. and These are the updated node evaluation parameters and the updated line evaluation parameters obtained in step S2, respectively.

[0034] At this point, the power-duration decoupling model is complete. This model, in the form of algebraic equations, not only assigns unambiguous time boundaries to different flexibility intervals, but also embeds the spatial physical carrying capacity limit of the grid's underlying structure into the power upper limit of the scheduling unit through updated evaluation parameters, thus achieving the fusion of vehicle-grid two-way physical safety boundaries.

[0035] S242, generates structured scheduling units; After completing the interval partitioning based on power inflection points and constructing the power-duration decoupling model, a structured description of the charging process is needed to transform the charging behavior into an executable scheduling object. Specifically: S2421, Discrete-time slice and candidate cell generation; Based on the aforementioned power-duration decoupling model, the system employs conventional continuous-time discretization methods (such as, but not limited to, fixed-time-step slicing or event-driven slicing) to divide the vehicle's charging and discharging interval into several standard charging time windows. Subsequently, the system directly calls the power adjustable range and duty cycle of each vehicle in different intervals output by the power-duration decoupling model, and structurally encapsulates the power adjustable range, duty cycle adjustable interval, and corresponding charging time window to initially generate candidate scheduling units.

[0036] S2422, introduces space frame evaluation parameters to calculate the impact distribution; Traditional scheduling units typically only contain active load and time sequence boundaries (i.e., only have a time dimension). To address the local network congestion problem caused by massive resource access, this invention introduces the influence distribution characteristics based on the underlying physical network structure into the construction of candidate scheduling units.

[0037] Specifically, the system extracts the target scheduling power from the candidate scheduling units, uses it as a mapping input, and directly performs a linear product mapping with the updated evaluation parameters obtained in step S2 to calculate the impact occupancy of the power scheduling action on each associated node and line in the power grid topology. This impact occupancy characterizes the absolute physical offset caused by the active power execution action of the target scheduling unit on the voltage phase angle of associated nodes and the current amplitude of transmission lines under specific topology constraints. Essentially, it reflects the degree of physical consumption of the voltage stability margin and transmission capacity of local nodes and lines by the scheduling unit. The calculation formula is expressed as follows: (20) (twenty one) In the formula, For the candidate unit to the target node The resulting voltage phase angle affects the occupancy rate; For this candidate unit to the target line The resulting line current affects the occupancy rate; and These are the updated node evaluation parameters and line evaluation parameters, respectively.

[0038] After obtaining the impact occupancy of each node and line, the system uses conventional data serialization or tensor concatenation techniques to combine these impact occupancy values ​​into a multi-dimensional spatial feature vector or feature set, forming the impact distribution of the scheduling unit on the power grid micro-topology. This is represented as: (twenty two) In the formula, To influence the distribution, and These are the occupancy rates affected by voltage phase angle and line current, respectively.

[0039] S2423 obtains the scheduling unit through attribute attachment and encapsulates the output; The system incorporates the influence distribution as a spatial physical attribute into the corresponding candidate scheduling unit data body. The resulting scheduling unit possesses at least an adjustable power range and time window constraints, thereby transforming vehicle scheduling requests into standardized control mechanisms that meet grid capacity constraints. The scheduling unit is represented as follows: (twenty three) In the formula, As a scheduling unit, For scheduling unit index, For vehicle or equipment identification. For physical topology access node identification; To adjust the type label (Boolean or enumerated integer), for example, to take... Represents the "power reduction range" (forced at this time) ),Pick Represents "intermittently powered area"; The charging time window is represented as a set of discretized boundaries for continuous time.

[0040] This step completely shields the complex internal electrochemical dynamics of the battery and transforms them into standardized scheduling units that are transparent to the power grid and have clear boundaries. The scheduling master station does not need to be aware of the state of charge and internal resistance changes of the underlying battery. It only needs to directly read the power range and influence distribution vector within the aforementioned scheduling unit and directly substitute them into a conventional linear programming solver for optimization and clearing. This greatly reduces the curse of dimensionality and computational latency in high-concurrency, massive resource access scenarios.

[0041] In one embodiment, S3 includes: S31, when the execution times of scheduling units overlap, determine the offset of coupling effects between scheduling units based on the updated initial evaluation parameters; Specifically, when multiple scheduling units are detected to have overlapping execution times, the system calculates and extracts the core mathematical formula set for the offset affected by coupling, as follows: (twenty four) (25) (26) (27) (28) In equation (24), This is the original deviation. These are electrical changes (the changes in node voltage phase angle and line current amplitude calculated in step S12). To predict electrical changes, the system calls the updated evaluation parameters obtained in the preceding steps, and combines them with the... The power supply intensity of each concurrent scheduling unit is calculated by the pure theoretical superposition result obtained by the conventional linear product summation algorithm, and its calculation formula is Equation (25). In equation (25), This represents the total number of scheduling units whose current execution time windows overlap. That is, the first step generated by the aforementioned steps. Power supply intensity within each scheduling unit; That is, the first The updated evaluation parameters corresponding to the topology nodes where each scheduling unit is located; In equation (26), For coupling to affect the offset, when the offset is valid (i.e. When the offset occurs, the system not only extracts its instantaneous absolute value, but also synchronously calls the conventional time series database to accurately record the topological location, duration and rate of change of the offset (the calculation formula is (28) through the conventional time series difference algorithm) as the direct data source for the subsequent construction of graph-theoretic spatiotemporal occupancy segments; Dynamic coupling threshold, calculated by taking the historical average of the original deviation over a historical time period. with standard deviation The result is obtained and its calculation is as shown in equation (27).

[0042] In equation (27), Threshold coefficient (preferably within a certain range) ).

[0043] In equation (28), The coupling affects the instantaneous rate of change of the offset. For discrete sampling step size, and They are respectively Time and The coupling at any given moment affects the offset.

[0044] like Figure 2 As shown in the figure, curve A represents the actual electrical change, and curve B represents the predicted change obtained by linear superposition of the initial evaluation parameters. The solid and dashed lines are used to distinguish between the actual and predicted curves, respectively. The vertical difference segment between the two curves represents the deviation range, and the dotted line is used to mark this difference interval (the dotted line is not visible). The reference segment is used to indicate the location of the coupling effect offset. It can be seen that there is a difference between the actual response and the predicted response, which indicates that there is a nonlinear coupling effect under the superposition of multiple scheduling units.

[0045] After dynamically extracting the coupling influence offset and its instantaneous rate of change of each grid entity in step S31, the process proceeds to step S32, where computer graph theory methods are used to transform them into physical control objects with clearly defined load-bearing boundaries. S32, identify the spatiotemporal occupancy segments of the scheduling unit in the time and power grid node space; The spatiotemporal occupancy segment, in essence, is a set of morphologically stable connected response structures within the topological space, used to characterize a stable power grid occupancy pattern formed by the combined action of multiple scheduling units within a certain time range. This is achieved through "electrical response extraction". Joint Space Construction Relational structure mapping The derivation of "spatiotemporal fragment recognition" transforms complex cross-spatiotemporal physical impacts into clear graph-theoretic topological entities. The specific steps are as follows: S321, Based on the electrical responses of nodes and lines during the time-domain overlap of scheduling units, construct a response relationship structure in the joint space of time and nodes; the electrical responses include phase angle changes, power changes, and timing characteristics; 1) Electrical response extraction: The electrical responses (including phase angle changes, power changes, and timing features) of each node during the overlap period are extracted. Specifically, the phase angle change characteristics, power change characteristics, and timing features of each grid entity at each discrete time step are collected in real time using a synchronous phasor measurement unit (PMU) and pile-end sensors, and these are assembled into a node response feature vector. : (29) (30) In equation (29), Represents a node No. Real-time change of voltage phase angle at each discrete time step; Indicates the first A discrete time step access node The algebraic sum of the power supply intensity of all overlapping scheduling units; This represents the real-time rate of change of the power supply intensity; The discrete time step during the overlap period ( , (This is the maximum allowable discrete time step).

[0046] 2) Joint Space Construction: Construct a two-dimensional joint space of time and node-path, expressed by the formula: .in, For network structure entity topology index ( The set of space frame entities This refers to the real-time topology structure identified online via RLS, which represents the active interconnections in the current cross-section. A set of dynamic nodes and The set of dynamic lines is merged to form the total spatial feature dimension, which dynamically converges to... .

[0047] 3) Relational structure mapping: Furthermore, within the joint space of time and nodes, the feature vectors of all nodes throughout the entire overlapping time period are arrayed using spatial topology and temporal dimensions to construct a response relationship matrix: (31) In the formula, For order is The response relationship matrix; Since the underlying elements in the response relation matrix contain different physical dimensions such as phase angle, power and rate of change, the system further calls the preset Min-Max normalization algorithm to perform dimensionless normalization processing on each feature dimension in the response relation matrix to generate a standardized response relation matrix that is convenient for subsequent collaborative optimization scheduling.

[0048] S322, Based on the connectivity of the response relationship structure (the spatiotemporal response topology structure constructed based on the standardized response relationship matrix mapping) and its persistence in the time dimension, extract the response structure that remains stable at different time scales (spatiotemporal segment identification). To filter out transient high-frequency noise, this invention focuses on verifying the time persistence of the topology: The system performs continuous time window connectivity tracking on the spatiotemporal response topology constructed based on the standardized response relation matrix mapping; connectivity tracking only occurs when a connected subgraph does not split in the topological space and its duration is... satisfy ( Only when the preset stable duration threshold is reached will the system establish the stable response topology as a spatiotemporal occupancy segment. This segment is essentially a logical encapsulation of the concentrated consumption of power grid resources in the spatiotemporal dimension.

[0049] It should be noted that the construction of the spatiotemporal response topology and the extraction of connected subgraphs employ conventional spatiotemporal graph processing methods. Specifically, the system maps the identified nodes and lines to vertices and edges in the topology graph, respectively. The time-series electrical characteristics (such as phase angle and power) of the column vectors in the standardized response relation matrix are assigned as weights to the corresponding vertices or edges in real time, thereby forming a spatiotemporal response topology graph that dynamically evolves with discrete time steps. The preferred graph traversal algorithm is either Depth-First Search (DFS) or Breadth-First Search (BFS).

[0050] The system constructs a feature discrimination space based on the duration of connected subgraphs and the average rate of change of coupling influence offsets, and adaptively classifies and labels spatiotemporal occupancy segments: Short-term impact type: when the following conditions are met and At that time, it was determined to be a short-term power surge impact; Continuous accumulation type: when the condition is met and When this occurs, it is judged as a slow accumulation of minor deviations over a long period of time exceeding the limit; Periodic fluctuation type: For segments that do not meet the above two categories but exhibit alternation at a specific frequency, they are judged to be periodic fluctuation type.

[0051] In the formula For the preset time threshold, The rate of change threshold, The duration of the connected subgraph is calculated using the following formula; The average rate of change is calculated using the following formula: (32) (33) In equation (32), The duration of the existence of this specific connected subgraph (i.e., the corresponding spatiotemporal occupancy segment) on the time axis; The initial discrete moment when this particular connected subgraph is first captured and identified in the two-dimensional joint space of time and node-line; The cutoff discrete moment is when the specific connected subgraph in the joint space finally dissipates, undergoes topological splitting, or disintegration. For discrete time step indexes within the lifecycle of this specific connected subgraph ( ); For a moment Below is a real-time topological connectivity subgraph configuration composed of a set of strongly electromagnetically correlated network entities; This is a preset time window indicator function used to determine the continuity of the topological shape. It applies if and only if the current time... Connected subgraph When the space remains topologically strongly connected and morphologically stable without splitting compared to the previous time step, the function output is: If the subgraph splits or falls below the intensity threshold, the function output will be: ; Preferably, it is a constant length, such as configured as 5 seconds or 15 seconds; In equation (33), The average rate of change of this specific spatiotemporal occupancy segment over the entire lifecycle is used to perform fine-grained classification and labeling of the segment in the discrimination space in conjunction with a time threshold; The total number of discrete time steps spanning the lifetime of this spatiotemporal segment (i.e., equal to the total number of sample points) is logically determined by a closed-loop operation of the difference between the start and end discrete times, satisfying the following relationship: ; This is the local temporal traversal index within the segment, from the start time to the end time. ); Corresponding to the space frame entity described in the formula of step S31 above. At any moment The coupling affects the instantaneous rate of change of the offset. During actual active summation traversal, this entity index... The set of active network entities belonging to the specific spatiotemporal segment occupied.

[0052] Based on this classification result, the system marks different types of spatiotemporal occupancy segments to directly guide the selection of adjustment strategies by subsequent scheduling units.

[0053] By identifying spatiotemporal occupancy segments, the coupling effect can be classified and characterized, thus providing a more refined basis for subsequent scheduling optimization. Simultaneously, classification results are obtained based on the aforementioned classification rules, and different types of spatiotemporal occupancy segments are labeled according to these results for use in selecting adjustment strategies for subsequent scheduling units.

[0054] In one embodiment, S4 includes: After identifying and classifying the spatiotemporal occupancy segments, this embodiment further implements a two-layer adjustment mechanism—"preliminary correction based on coupling offset" and "secondary optimization based on segment deviation"—on the scheduling units causing nonlinear congestion, based on the actual physical carrying capacity of the power grid. This abstracts and outputs a combination of underlying physical executable units that meet safe operation constraints. This includes: S41, Based on the distribution range of the spatiotemporal occupied segments and the difference between the initial evaluation parameters and the upper limit of the carrying capacity, determine the degree of deviation of the impact occupancy of the scheduling unit relative to the upper limit of the carrying capacity. The system invokes the spatiotemporal occupancy segments identified and marked in step S32 to pinpoint their distribution range. The distribution range of these spatiotemporal occupancy segments is represented by overlapping discrete moments in the time dimension. In the spatial dimension, it is represented as a target node. Related lines to the target Simultaneously, the system extracts the difference between the initial evaluation parameters and the upper limit of the carrying capacity. This difference manifests as a safe phase angle operating margin at the grid node and a safe current carrying margin on the line. In conventional grid safety verification, only static linear superposition of influence quantities is possible. The formula increment of this invention lies in the additional introduction of a nonlinear coupling influence offset into the numerator, quantitatively calculating the deviation of the influence occupancy of the scheduling unit relative to the upper limit of the carrying capacity. The calculation formula is expressed as follows: (34) (35) In the formula, and They represent the times respectively. Next target node Voltage phase angle deviation index and target associated circuit The deviation index of line current; This represents the set of active scheduling units with overlapping execution times within the current spatiotemporal segment; For the first Candidate units for the target node The resulting voltage phase angle affects the occupancy rate; For the first Each candidate unit corresponds to the target line. The resulting line current affects the occupancy rate; The offset is affected by coupling. and They represent the times respectively. Next, target node Voltage phase angle coupling affects offset and target associated lines The line current coupling effect offset, which are used as the coupling effect offset. Spatiotemporal component mapping on a specific space frame entity; The safe phase angle operating margin of the target node of the power grid is obtained by subtracting the current actual phase angle absolute value from the preset absolute value of the safe phase angle limit of the target node. This refers to the safety current-carrying margin of power grid-connected lines.

[0055] The system will satisfy the calculation results or The spatiotemporal region is determined as the over-limit conflict zone of the underlying physical limit of the distribution network, which is used to quantify the nonlinear power flow superposition effect under high-concurrency response scenarios, and serves as the trigger criterion and control boundary for executing the differentiated reverse adjustment in the subsequent step S42.

[0056] S42, based on the degree of deviation, the execution time of the scheduling unit is processed, including shifting the execution time of the scheduling unit, adjusting the power, or splitting and merging the scheduling units; the specific implementation method is as follows: The system calculates and outputs the deviation index based on step S41. or ), for sets The scheduling unit that triggers a physical limit violation conflict will be dynamically adjusted accordingly: 1) Shift the execution time of the scheduling unit: For the conflict segments marked as "short-term impact type" in step S322, the system selects the scheduling units with the adjustment type label "intermittent power supply interval" and uses a conventional time shift operator to shift the charging time window of the scheduling unit to the low load period of the entire network outside the coverage of the conflict segment. 2) Split and merge scheduling units: For scheduling units that cannot be moved as a whole due to the expected departure time of the vehicle, or conflict segments marked as "periodic fluctuation type", the system performs the following actions: splitting the scheduling unit, discretizing it along the time axis into multiple sub-candidate units including but not limited to equally divided micro-slices or variable step-size discrete blocks, and embedding them in a staggered manner; or performing the following actions: merging the scheduling units, interleaving and recombining multiple scheduling units that start and stop intermittently on the time axis to distribute the peak of the total active power load.

[0057] 3) Adjust the power: For conflict segments marked as "continuous accumulation type," the system adjusts the power. The system uses the excess proportion of the deviation index as the feedback control variable, and adjusts the voltage drop proportionally based on the active power supply intensity or effective duty cycle according to the following formula: (36) (37) In the formula, The adjusted active power supply intensity; This refers to the adjusted effective duty cycle. Minimum continuous charging power limit; This represents the lower limit of the duty cycle for the corresponding vehicle.

[0058] S43: Extract continuous power supply segments, interruptible segments, and constant power segments from the processed scheduling units to generate a combination of executable units that satisfy node voltage and line current carrying constraints. The specific implementation method is as follows: From the scheduling unit processed in step S42, extract the physical control features that the corresponding underlying controller can directly execute, including continuous power supply segments, interruptible segments, and constant power segments: Extracting continuous power supply segments: The system selects scheduling units that have been processed and whose adjustment type label is "power reduction range", maintains the continuous power supply constraint with the duty cycle degraded to 1, extracts the time series curve of the adjusted active power supply intensity, and forms a continuous power supply segment whose power amplitude is adaptively adjusted according to the safety margin of the grid.

[0059] Interruptible Segment Extraction: The system filters out scheduling units with the processed adjustment type label "Intermittent Power Supply Section." Based on the adjusted duty cycle and shifted time window, it extracts pulse timing control signals that alternate between active power output periods and zero-power charging stop periods to form interruptible segments. The pulse timing control signal sequence is generated through pulse width modulation (PWM), among other things. Conventional digital modulation tools such as Sigma-Delta modulation, sinusoidal pulse width modulation (SPWM), or sequential logic state machine allocation algorithms are used to generate binary sequences that are interleaved within the charging time window.

[0060] Extracting constant power segments: The system retrieves scheduling units that belong to rigid intervals or have not triggered over-limit conflicts, eliminating the need for any power and timing adjustments, and directly extracts their inherent original natural charging power curves to form constant power segments.

[0061] Finally, the system seamlessly splices the control segments according to the spatial topology access nodes and device identifiers, thereby generating a combination of executable units that meet the node voltage and line current carrying constraints.

[0062] The system combines the executable units that meet the node voltage constraints and line current carrying constraints into a control sequence, and sends it in a closed loop to the underlying terminal charging pile controllers through conventional data communication tools, including but not limited to distribution automation systems or SCADA systems that comply with the IEC61850 protocol. This drives the physical hardware modules to execute the corresponding power output, thereby completing the refined closed-loop collaborative scheduling of the distribution network.

[0063] Example 2 illustrates a specific implementation method for mapping executable units to different markets in a collaborative scheduling method for charging station clusters that supports multi-market participation. The method includes the following steps: S43, the step of forming an executable unit output that satisfies the grid carrying capacity constraints further includes: converting the optimized scheduling results into standard parameters that can directly participate in electricity market transactions; S431, to exclude invalid units that cannot be effectively utilized by the market due to resource fragmentation, insufficient adjustment margin, or severe over-limit at the underlying level, the system performs a schedulable resource access screening. The system will then select units that, after reverse adjustment in step S42, have output power not exceeding the grid's capacity limit and a duty cycle... The scheduling unit is used as the schedulable unit; preferably, the schedulable unit also needs to meet the preset minimum duration constraint, the value of the minimum duration is in the range of 5 to 30 minutes, and in this embodiment it is preferably 10 minutes.

[0064] Because charging loads exhibit different physical regulation characteristics at different states of charge (e.g., power reduction ranges, intermittent power supply ranges), the system pre-stores the technical specifications of different electricity trading markets. The system maps and compares the adjustable power range of each selected dispatchable unit with the entry thresholds of each market to determine one or more target electricity markets that each dispatchable unit can meet the specifications of. These target electricity markets include, but are not limited to, peak-shaving markets, frequency regulation markets, ancillary service markets, or implicit spot trading markets. The system performs rapid matching using methods including, but not limited to, static rule bases, feature lookup tables, or multi-dimensional matrix lookup tables. Specific matching criteria are as follows: If the response time and power continuity indicators of the schedulable unit meet the technical specifications of the frequency modulation market, then the unit is matched to the corresponding frequency modulation market. If the continuous response time of the schedulable unit meets the technical specifications of the peak shaving market, then the unit is matched to the corresponding peak shaving market.

[0065] Through the above rules, the system has constructed an adaptive market rule base, achieving precise alignment between regulatory flexibility and the demands of a multi-tiered electricity market.

[0066] S432 calculates the allocatable capacity for each market based on power values, according to the technical specifications applicable to multiple markets for the same dispatchable unit. Since a single dispatchable unit may simultaneously meet the technical specifications of multiple target electricity markets, the system quantitatively calculates the allocatable capacity of each target electricity market based on power values, applicable to multiple market technical specifications for the same dispatchable unit. The system calculates its allocatable capacity in each target electricity market based on the overlap between the dispatchable unit's own power adjustable range and the technical specification constraints of each target electricity market (including but not limited to minimum declared capacity, maximum adjustment step size, and response frequency limits). The formula for calculating the allocatable capacity is as follows: (38) (39) in, For allocable capacity, This represents the power adjustability margin of the scheduling unit. The maximum application capacity allowed by market rules. For market numbering.

[0067] Furthermore, when a dispatchable unit is applicable to multiple target electricity markets simultaneously, the system employs conventional allocation strategies, including but not limited to the principle of maximizing benefits, dynamic electricity price guidance rules, or asset matching priority, to allocate capacity across multiple markets. The allocation formula is as follows: (40) in Weights are assigned to the market, with values ​​ranging from 0 to 1; Preferably, in a practical system, the market allocation weight can be dynamically adjusted in conjunction with electricity prices or market revenue.

[0068] This step takes into account that the adjustability of the charging load comes from its power elasticity rather than its absolute power level. Therefore, compared with allocation based on "rated capacity", dynamic allocation based on "adjustable capacity" is adopted, which realizes a fine alignment between scheduling resources and market demand, thereby avoiding the problems of insufficient resource utilization or scheduling failure under the traditional static matching method.

[0069] S433, based on the same market technical specifications applicable to multiple dispatchable units, sums their power values ​​as the dispatchable capacity of that market; The system accumulates the power values ​​of multiple dispatchable units based on the same market technical specifications applicable to the target electricity market, thus determining the total dispatchable capacity of that electricity market. When multiple dispatchable units across the entire network pass the initial screening and are matched to the same electricity market, their power values ​​are algebraically recombined and superimposed within the same control time window. The calculation formula is as follows: (41) in For the market In time Total schedulable capacity, The allocatable capacity of the scheduling unit in this market. To meet the requirements across the entire network A set of dispatchable units for a target electricity market technical specification. The capacity accumulation process is based on the premise that each dispatchable unit has met the physical carrying capacity constraints of the underlying power grid and that its topological influences can be completely superimposed in the microscopic space.

[0070] Preferably, in scenarios involving multiple markets participating in parallel, capacity allocation priority or revenue weight can be introduced to coordinate the allocation of capacity among different markets. For example, a sorting rule can be constructed based on the revenue per unit capacity, prioritizing the allocation of scheduling units to markets with higher revenue, thereby improving overall economic efficiency.

[0071] After the multiple markets have cleared and the dispatchable capacity of each target market has been determined, the dispatch master station decouples this capacity in reverse to the target power supply intensity and duty cycle of the corresponding charging gun within each discrete charging time window, thereby completing a top-down, precise physical-side response and control closed loop. Through this process, the power regulation capacity of all charging piles is effectively allocated, maximizing market participation while ensuring that the grid load does not exceed its carrying capacity.

[0072] This invention overcomes static model prediction bias and achieves high-fidelity dynamic sensing of the power grid by applying power perturbations and obtaining updated evaluation parameters through online identification. Secondly, it constructs a power-duration decoupling model, transforming naturally decaying passive charging into clearly defined schedulable units and precisely delineating flexible adjustment ranges. Next, to address nonlinear disturbances from concurrent responses of multiple units, it extracts spatiotemporal occupancy segments and adjusts unit power or discrete charging time windows to eliminate the risk of node and line overload. Finally, it inversely decouples the cross-market aggregated winning capacity with the hardware stability lower limit to the target power supply intensity and duty cycle of the corresponding charging gun. This invention achieves a complete closed loop from physical sensing and spatiotemporal decoupling to market clearing and precise equipment control.

[0073] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0074] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0075] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0076] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0077] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0078] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for collaborative scheduling of charging station clusters supporting multi-market participation, characterized in that, include: Obtain the electrical quantities of the power grid and the power of the charging gun, and construct initial evaluation parameters by applying power perturbations; Based on the online identification results of power grid topology and line parameters, the evaluation parameters are updated, and a scheduling unit is generated by combining the power-duration decoupling model. The updated evaluation parameters include: identifying the topology and line parameters between nodes online based on node electrical quantities; constructing the node impedance matrix of the power grid based on the identified topology and line parameters; using the impedance matrix and the access node of the charging gun, calculating the power flow evaluation parameters of each charging gun through power flow sensitivity; and weighting and fusing the power flow evaluation parameters with the initial evaluation parameters to obtain the updated evaluation parameters. When the execution times of scheduling units overlap, the coupling effect offset between scheduling units is determined based on the updated evaluation parameters, and the spatiotemporal occupancy segments of each scheduling unit in the time dimension and the power grid node-line space are identified. The identification of the spatiotemporal occupancy segments includes: constructing a response relationship structure in the time and node joint space based on the electrical responses of nodes and lines during the overlapping execution times of scheduling units; extracting a stable response structure at different time scales based on the connectivity of the response relationship structure and its persistence in the time dimension; and determining the stable response structure as the spatiotemporal occupancy segment. Based on the spatiotemporal occupancy segments and coupling effect offsets, the scheduling unit is adjusted to form an executable unit output that satisfies the power grid carrying capacity constraints.

2. The charging station cluster collaborative scheduling method supporting multi-market participation as described in claim 1, characterized in that, The initial evaluation parameters include node voltage phase angle evaluation parameters and line current evaluation parameters; the construction of initial evaluation parameters includes: Record the changes in node voltage phase angle and line current amplitude before and after the disturbance; The ratio of the change in node voltage phase angle to the change in power is used to obtain the node voltage phase angle evaluation parameter. Calculate the ratio of the change in current amplitude to the change in power to obtain the line current evaluation parameters.

3. The charging station cluster collaborative scheduling method supporting multi-market participation as described in claim 1, characterized in that, Before generating the scheduling unit, a power-duration decoupling model is also constructed: The power inflection point is determined based on the vehicle's state of charge and the characteristics of charging power decay. Based on the power inflection point, the power reduction range and the intermittent power supply range are divided, and a segmented mapping relationship between the charging gun occupation time and the power supply intensity is established. Based on the segmented mapping relationship and the updated evaluation parameters, a power-duration decoupling model is constructed.

4. The charging station cluster collaborative scheduling method supporting multi-market participation according to claim 3, characterized in that, The generation scheduling unit includes: Based on the power-duty decoupling model, the adjustable power range and duty cycle of each vehicle in different intervals are determined, and the charging time window is divided. Based on the power adjustable range, duty cycle, and charging time window, candidate scheduling units are generated; Based on the updated evaluation parameters, the impact distribution of candidate scheduling units on nodes and lines is calculated. The influence distribution is added as one of the attributes to the candidate scheduling unit to form a scheduling unit that includes at least the power adjustable range, duty cycle, charging time window and influence distribution.

5. The charging station cluster collaborative scheduling method supporting multi-market participation according to claim 4, characterized in that, The calculation of the influence distribution includes: Based on the updated evaluation parameters and the power supply intensity of the candidate scheduling units, the impact occupancy of each node and line is calculated, and the impact occupancy is combined into a vector or set as the impact distribution.

6. The charging station cluster collaborative scheduling method supporting multi-market participation according to claim 1, characterized in that, The calculation of the coupling effect offset includes: When the execution times of scheduling units overlap, the actual electrical changes of nodes and lines are obtained; Based on the updated evaluation parameters and the power supply intensity of each dispatching unit, the predicted electrical changes of nodes and lines are calculated and linearly superimposed to obtain the superposition result. The difference between the actual electrical change and the superposition result is calculated to obtain the original deviation; The coupling threshold is dynamically calculated based on the current power grid operating status. The coupling effect offset is extracted from the original deviation based on the threshold, and its duration and rate of change are recorded.

7. The charging station cluster collaborative scheduling method supporting multi-market participation according to claim 1, characterized in that, The identification of the spatiotemporal occupancy segment includes: The electrical response includes phase angle changes, power changes, and timing characteristics; Different types of spatiotemporal occupancy segments are distinguished based on their duration and evolution characteristics.

8. The charging station cluster collaborative scheduling method supporting multi-market participation according to claim 1, characterized in that, The executable unit output that satisfies the power grid carrying capacity constraints includes: Based on the distribution range of the spatiotemporal occupancy segments and the difference between the initial evaluation parameters and the capacity limit, the degree of deviation of the scheduling unit's impact occupancy from the capacity limit is determined. Based on the degree of deviation, the execution time of the scheduling unit is processed, including shifting the execution time of the scheduling unit, adjusting the power, or splitting and merging the scheduling unit; Extract continuous power supply segments, interruptible segments, and constant power segments from the processed scheduling units to generate a combination of executable units that meet node voltage and line current carrying constraints.

9. The charging station cluster collaborative scheduling method supporting multi-market participation according to claim 8, characterized in that, The process of generating an executable unit output that satisfies the power grid carrying capacity constraints also includes: Extract the execution time, power value, and duty cycle of each unit from the combination of executable units; Units that meet the preset schedulable conditions are identified as schedulable units; Match the power adjustable range of the dispatchable units with the technical specifications of different markets to determine the target market for each dispatchable unit. For the same schedulable unit, the allocatable capacity of the schedulable unit in each target market is determined based on the overlap between its power adjustable range and the technical specification constraints of the target market. For the same target market, the allocatable capacity of each unit is summed to obtain the schedulable capacity of that market.

Images