Method for optimizing energy storage capacity of electric vehicle to grid microgrid based on dynamic programming

Abstract

The application discloses a dynamic programming-based electric vehicle V2G microgrid energy storage capacity optimization method and particularly relates to the technical field of microgrid energy storage optimization, and is used for solving the problem of the existing electric vehicle mobility-induced spatial dynamic characteristics leading to energy storage capacity optimization defects, collecting power grid topology parameters and electric vehicle moving track data, performing geographical grid division on node positions, generating a space-time matrix in combination with time stamps and identifying high-frequency access nodes; constructing a state transition equation based on an impedance matrix between nodes, setting a node-level safety boundary with a line capacity change rate as a dynamic constraint; adopting a phased reverse dynamic programming algorithm, preferentially performing spatial dimension optimization on the high-frequency access nodes, generating charging and discharging constraints that fuse voltage deviation and capacity overrun penalty items; performing global optimization with the minimum total operation cost as the target after integrating the constraints, outputting energy storage capacity configuration strategies in each period; and finally generating an energy storage deployment scheme.

Description

Technical Field

[0001] This invention relates to the field of microgrid energy storage optimization technology, and more specifically, to a method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming. Background Technology

[0002] With the deep integration of electric vehicle V2G technology and microgrids, dynamic programming methods are widely used in energy storage capacity optimization to achieve economic dispatch and supply-demand balance. Existing technologies typically divide optimization stages based on the time dimension, generating capacity configuration and charging / discharging strategies for energy storage devices by predicting the charging and discharging demand of electric vehicles and the output of renewable energy. Such methods assume that the physical location of electric vehicles connected to the microgrid is fixed during the optimization period and assume that the grid topology parameters (such as node voltage and line capacity) are uniformly distributed in the spatial dimension. This makes it difficult to adapt to the transmission constraints of 750 kV and above AC transmission systems on the stability of microgrid access. Therefore, complex microgrids are simplified into single-node or homogeneous multi-node models for solution.

[0003] In practice, the lack of consideration for the spatial dynamic characteristics caused by the mobility of electric vehicles leads to shortcomings when participating in large-scale grid coordinated dispatch: the dynamic programming strategy ignores the differential impact of changes in the access location of electric vehicles on the voltage and line capacity of grid nodes, resulting in spatial mismatch problems in actual multi-node microgrid scenarios. This triggers a chain reaction of safety risks to the backbone network. Specifically, the energy storage capacity configuration and dispatch strategy cause voltage overruns and line overloads in some grid nodes, while other nodes have redundant capacity. This reduces the safety margin of the intelligent dispatch system, resulting in a double deterioration of grid safety risks and investment costs. Summary of the Invention

[0004] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a dynamic programming-based method for optimizing the energy storage capacity of electric vehicle V2G microgrids to address the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] The method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming includes the following steps:

[0007] S1. Collect the grid topology parameters of the target microgrid and the electric vehicle movement trajectory data. The grid topology parameters include the inter-node impedance matrix, node voltage limits and line capacity thresholds.

[0008] S2. Geographically divide the node locations in the electric vehicle movement trajectory data into grids, generate a spatiotemporal matrix by combining timestamps, and identify high-frequency access nodes.

[0009] S3. Construct state transition equations based on the inter-node impedance matrix, take the rate of change of line capacity caused by the movement of electric vehicles as a constraint, and set node-level safety boundaries according to node voltage limits and line capacity thresholds.

[0010] S4. A phased reverse dynamic programming algorithm is adopted, with node voltage deviation and line capacity overrun as penalty terms. Spatial dimension optimization is performed first for high-frequency access nodes to generate node charging and discharging constraints.

[0011] S5. Integrate node charging and discharging constraints, perform global optimization with the goal of minimizing total operating costs, and output a global strategy for energy storage capacity configuration in each time period;

[0012] S6. Generate an energy storage deployment plan based on the global strategy and the distribution of high-frequency access nodes.

[0013] In a preferred embodiment, S1 includes:

[0014] S1a. The power grid topology parameters are obtained in real time through the microgrid monitoring system. The power grid topology parameters include the inter-node impedance matrix, node voltage limits and line capacity thresholds. The node voltage limits and line capacity thresholds are extracted from the preset safe operation database of the microgrid management platform.

[0015] S1b: Collect electric vehicle movement trajectory data through the vehicle positioning device. The movement trajectory data includes the node location and corresponding timestamp of the electric vehicle when it connects to the microgrid during a continuous period.

[0016] S1c. Based on the inter-node connection relationship in the power grid topology parameters, calculate the numerical update result of the inter-node impedance matrix in real time, and perform spatial correlation mapping between the numerical update result of the inter-node impedance matrix and the node position in the electric vehicle movement trajectory data.

[0017] S1d performs data verification on the numerical update results of node voltage limits, line capacity thresholds, and inter-node impedance matrices, and removes abnormal parameters that exceed the preset deviation range.

[0018] In a preferred embodiment, S2 includes:

[0019] S2a. Dynamically divide the node locations in the electric vehicle movement trajectory data into grids according to the preset geographic grid density to generate geographic grid regions corresponding to the distribution of power grid nodes.

[0020] S2b: Divide the electric vehicle access events in each geographic grid area into consecutive time periods according to timestamps, count the number of access events in each time period, and generate a spatiotemporal matrix;

[0021] S2c: Based on preset clustering analysis rules, cluster the number of access events in the spatiotemporal matrix, identify geographical grid areas where the number of access events exceeds a preset high-frequency threshold, and mark them as high-frequency access nodes;

[0022] S2d: Perform spatiotemporal correlation verification on high-frequency access nodes. If the same geographic grid area is marked as a high-frequency access node in several consecutive time periods, it is confirmed as a valid high-frequency access node.

[0023] In a preferred embodiment, S3 includes:

[0024] S3a. Based on the dynamic change characteristics of the inter-node impedance matrix, a state transition equation is constructed. The numerical update result of the inter-node impedance matrix is ​​dynamically associated with the node position in the electric vehicle trajectory data through a weighting factor to generate a spatiotemporally coupled state transition weight.

[0025] S3b. Based on the rate of change of line capacity caused by the movement of electric vehicles, and combined with historical load data and line capacity thresholds, establish a linear constraint relationship between the rate of change of line capacity and the charging and discharging power of nodes, and generate dynamic constraint conditions.

[0026] S3c. Based on the node voltage limit and the line capacity threshold, the node-level safety boundary is set. The node-level safety boundary includes the allowable range of node voltage deviation and the tolerance of line capacity exceeding the limit. The node-level safety boundary is embedded in the state transition equation as a hard constraint.

[0027] In a preferred embodiment, S4 includes:

[0028] S4a. The optimization phase is divided into spatial dimension optimization and time dimension recursion. The priority order of spatial dimension optimization is determined based on the distribution of high-frequency access nodes.

[0029] S4b. In the spatial dimension optimization stage, the initial charging and discharging power constraints of the high-frequency access node are generated by using the node voltage deviation and line capacity over-limit as penalty terms.

[0030] S4c: Dynamically adjust the weight of the penalty term based on the inter-node impedance matrix in the power grid topology parameters, iteratively correct the initial charging and discharging power constraints, and generate optimized constraints that satisfy the node-level safety boundary.

[0031] In the time-dimensional recursive phase, S4d integrates the optimization constraints of all high-frequency access nodes, eliminates local constraints that conflict with the global strategy, and generates the final node charging and discharging constraints.

[0032] In a preferred embodiment, during the spatial dimension optimization stage, a reverse dynamic programming algorithm is used to recursively calculate from the lowest priority node, using node voltage deviation and line capacity overrun as penalty terms to construct an objective function. The specific form of the objective function includes:

[0033] The voltage deviation penalty and the line capacity over-limit penalty are weighted and summed proportionally. The voltage deviation penalty is the percentage of the absolute difference between the actual voltage and the rated voltage to the rated voltage, and the line capacity over-limit penalty is the square of the ratio of the actual line current to the capacity threshold. The weighting ratio is dynamically adjusted according to the power grid topology parameters.

[0034] In a preferred embodiment, S5 includes:

[0035] S5a. Summarize the charging and discharging constraints of all high-frequency access nodes to generate a node-level charging and discharging feasible region set, which includes the upper and lower limits of power and the time validity range of each node.

[0036] S5b. Construct a charging and discharging cost coefficient matrix based on the inter-node impedance matrix in the power grid topology parameters. The charging and discharging cost coefficient matrix includes the weights of the combined impact of node charging and discharging power on line loss, equipment depreciation and peak-valley electricity price.

[0037] S5c. Taking the minimization of total operating cost as the objective function, the charging and discharging cost coefficient matrix and the node-level charging and discharging feasible domain set are input into the linear programming model to solve the global charging and discharging power allocation scheme that satisfies the time-period coupling constraint.

[0038] S5d performs N-1 safety checks on the global charging and discharging power allocation scheme, eliminates schemes that cause node voltage to exceed limits or capacity to exceed limits when a single line fails, and outputs the global strategy for energy storage capacity configuration in each time period.

[0039] In a preferred embodiment, S6 includes:

[0040] S6a. Integrate the energy storage capacity configuration schemes for each time period and the distribution data of high-frequency access nodes in the global strategy to generate a node energy storage demand matrix. The node energy storage demand matrix includes the capacity demand, charging and discharging power demand of each node and the time distribution characteristics that match the high-frequency time period.

[0041] S6b. Based on the node energy storage demand matrix and the inter-node impedance matrix, calculate the energy storage deployment priority of each node. The energy storage deployment priority is dynamically adjusted according to the access frequency during high-frequency periods and the electrical centrality of the nodes.

[0042] S6c uses node-level security boundaries as constraints, combined with charging and discharging power limits and adaptive priority rules, to generate deployment schemes for the capacity, power limits, and geographical locations of energy storage devices using a multi-objective optimization algorithm;

[0043] S6d: Perform node-level voltage stability verification on the energy storage deployment scheme, eliminate schemes that exceed the charging and discharging power limit or voltage deviation tolerance, and output the final energy storage deployment scheme.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] 1. This invention dynamically maps the movement trajectory of electric vehicles to the physical topology of the power grid through geographic grid division and spatiotemporal matrix construction, realizing accurate identification and spatiotemporal distribution quantification of high-frequency access nodes. Combined with dynamically updated node impedance matrix to adjust charging and discharging constraints, it solves the spatial mismatch defects of traditional single-node models, ensures the spatial consistency between energy storage capacity allocation and load demand, and effectively suppresses the risks of local voltage overruns and line overloads.

[0046] 2. By employing a phased inverse dynamic programming algorithm, high-frequency node spatial optimization is deeply coupled with global security objectives, solving the problem of difficulty in coordinating local optimization and global constraints. Based on priority optimization in the spatial dimension, a node-level security boundary and an N-1 fault pre-verification mechanism are embedded to construct a hierarchical and progressive global strategy generation framework. Through a multi-objective optimization algorithm, multi-objective coordination of economy, security, and operability is achieved, overcoming the strategy oscillations and security risks caused by the lack of closed-loop verification in existing technologies. This ensures the robustness of energy storage deployment schemes in complex power grid scenarios and provides a systematic solution that balances efficiency and reliability for microgrid energy storage capacity optimization. Attached Figure Description

[0047] Figure 1 This is a flowchart of the method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming, as described in this invention. Detailed Implementation

[0048] 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.

[0049] Example: Figure 1 This invention presents a dynamic programming-based method for optimizing the energy storage capacity of V2G microgrids for electric vehicles, which includes the following steps:

[0050] S1. Collect the grid topology parameters of the target microgrid and the electric vehicle movement trajectory data. The grid topology parameters include the inter-node impedance matrix, node voltage limits and line capacity thresholds.

[0051] S2. Geographically divide the node locations in the electric vehicle movement trajectory data into grids, generate a spatiotemporal matrix by combining timestamps, and identify high-frequency access nodes.

[0052] S3. Construct state transition equations based on the inter-node impedance matrix, take the rate of change of line capacity caused by the movement of electric vehicles as a constraint, and set node-level safety boundaries according to node voltage limits and line capacity thresholds.

[0053] S4. A phased reverse dynamic programming algorithm is adopted, with node voltage deviation and line capacity overrun as penalty terms. Spatial dimension optimization is performed first for high-frequency access nodes to generate node charging and discharging constraints.

[0054] S5. Integrate node charging and discharging constraints, perform global optimization with the goal of minimizing total operating costs, and output a global strategy for energy storage capacity configuration in each time period;

[0055] S6. Generate an energy storage deployment plan based on the global strategy and the distribution of high-frequency access nodes.

[0056] S1. Collect the grid topology parameters of the target microgrid and the electric vehicle movement trajectory data. The grid topology parameters include the inter-node impedance matrix, node voltage limits, and line capacity thresholds, specifically:

[0057] The microgrid monitoring system acquires the grid topology parameters of the target microgrid in real time. These parameters include the inter-node impedance matrix, node voltage limits, and line capacity thresholds. The node voltage limits and line capacity thresholds are extracted from the microgrid management platform's preset safe operation database. This database is a standardized database pre-set during the microgrid planning and design phase based on grid equipment parameters and operating procedures. For example, the node voltage limit is ±10% of the rated voltage, and the line capacity threshold is 85% of the line's maximum designed transmission power.

[0058] The vehicle-mounted positioning device collects the movement trajectory data of electric vehicles. The movement trajectory data includes the node location of the electric vehicle connected to the microgrid during continuous time periods and the corresponding timestamp. The vehicle-mounted positioning device adopts the Global Positioning System or the Beidou Navigation System, and the timestamp accuracy is at the second level, ensuring the accurate spatial correspondence between the electric vehicle location and the microgrid node.

[0059] Based on the inter-node connectivity in the power grid topology parameters, the numerical update of the inter-node impedance matrix is ​​calculated in real time. Specifically, the calculation of the inter-node impedance matrix is ​​based on the inter-node connectivity in the power grid topology parameters, using the node admittance calculation method in circuit topology, and updating the matrix values ​​according to the electrical distance between adjacent nodes and the line impedance parameters. For example, when the impedance value of a node changes due to line modification, the values ​​of the corresponding rows and columns in the inter-node impedance matrix will be updated synchronously.

[0060] The numerical update results of the impedance matrix between nodes are spatially correlated and mapped with the node positions in the electric vehicle movement trajectory data. Spatial correlation mapping refers to matching the location information of the electric vehicle with the geographical coordinates of the power grid node. According to the geographical grid division rules, the grid area where the electric vehicle is located is mapped to the corresponding power grid node, thereby establishing a dynamic correlation between the electric vehicle movement trajectory and the impedance matrix of the power grid node.

[0061] The updated values ​​of node voltage limits, line capacity thresholds, and inter-node impedance matrices are validated, and abnormal parameters exceeding the preset deviation range are removed. The preset deviation range is set based on historical operating data and equipment error ranges. For example, the allowable deviation for node voltage limits is ±2%, the allowable deviation for line capacity thresholds is ±3%, and a deviation of ±5% in the inter-node impedance matrix is ​​considered an abnormal parameter.

[0062] During data verification, if a node voltage limit is found to exceed the preset deviation range, the voltage limit for that node is retrieved again from the preset safe operation database. If the updated values ​​of the line capacity threshold or the inter-node impedance matrix exceed the preset deviation range, an alarm signal is sent to the microgrid monitoring system and a manual review is requested. Through these steps, the accuracy and reliability of the input power grid topology parameters and electric vehicle trajectory data are ensured, providing a high-quality data foundation for the subsequent construction of the dynamic programming model.

[0063] During implementation, the following points should be noted: First, when the microgrid monitoring system acquires grid topology parameters in real time, it must maintain synchronization with the data interface of the microgrid management platform to avoid parameter expiration due to communication delays. Second, the node location data collected by the vehicle-mounted positioning device needs to be calibrated with the microgrid's geographic information system to ensure the spatial mapping accuracy between the electric vehicle's location and the grid nodes. Third, the real-time calculation of the impedance matrix between nodes needs to consider the dynamic changes in the grid's operating status, such as impedance parameter changes caused by distributed power source switching or load fluctuations. Fourth, the preset deviation range of the data verification process needs to be dynamically adjusted according to the actual operating environment. For example, the deviation range of the line capacity threshold should be appropriately widened in high-temperature or low-temperature environments to avoid erroneously removing normal parameters. Fifth, the handling of abnormal parameters needs to be recorded in the operation log, including the abnormal value, the time of occurrence, and the handling measures, to facilitate subsequent traceability and analysis.

[0064] S2. Geographically divide the node locations in the electric vehicle trajectory data into grids, generate a spatiotemporal matrix by combining timestamps, and identify high-frequency access nodes. Specifically:

[0065] The node locations in the electric vehicle movement trajectory data are dynamically divided into grids based on a preset geographic grid density, generating geographic grid regions corresponding to the distribution of power grid nodes.

[0066] The preset geographic grid density is set according to the historical access frequency distribution of electric vehicles in the target microgrid coverage area. The specific rules are as follows: a high-density grid of 10×10 grids per square kilometer is used in urban commercial areas, and a low-density grid of 5×5 grids per square kilometer is used in residential areas.

[0067] During dynamic grid partitioning, if the number of power grid nodes in a certain geographic grid area exceeds 3, the grid density of that area is automatically increased to 15×15 per square kilometer, ensuring that each geographic grid area is associated with only one power grid node. For example, if a commercial area is initially partitioned into 10×10 grids, after detecting 4 power grid nodes in the area, the grid density is automatically adjusted to 15×15.

[0068] Electric vehicle access events within each geographic grid area are divided into consecutive time periods based on timestamps. The number of access events within each time period is counted to generate a spatiotemporal matrix.

[0069] The rule for dividing continuous time periods is a fixed time interval, which can be set to 15 minutes per period. The timestamp accuracy is consistent with the collected electric vehicle movement trajectory data.

[0070] The rows of the spatiotemporal matrix represent geographic grid regions, and the columns represent consecutive time periods. The matrix element values ​​are the number of electric vehicle access events for the corresponding geographic grid region within the corresponding time period. For example, the element value N in the i-th row and j-th column of the spatiotemporal matrix... ij This represents the number of access events for the i-th geographic grid region in the j-th time period, where i = 1, 2, ..., M (M is the total number of geographic grid regions) and j = 1, 2, ..., T (T is the total number of time periods).

[0071] Clustering of access event counts in the spatiotemporal matrix using statistical analysis methods identifies high-frequency access nodes. Specifically, for each time period j, the average access event count μ across all geographic grid regions during that time period is calculated. j and standard deviation σ j Set the preset high-frequency threshold to μ j +2σ j If the number of access events N in geographic grid region i during time period j... ij If the value exceeds the preset high-frequency threshold, the area is marked as a high-frequency access node for time period j.

[0072] Spatiotemporal correlation verification is performed on high-frequency access nodes. The verification rule is as follows: if the same geographic grid area is marked as a high-frequency access node in three or more consecutive time periods, it is confirmed as a valid high-frequency access node. The criterion for consecutive time periods is an adjacent and uninterrupted sequence of time periods. For example, if geographic grid area A is marked as a high-frequency access node in time periods t1, t2, and t3 respectively, it is confirmed as a valid high-frequency access node; if it is marked only in time periods t1 and t3, the continuity condition is not met. During the verification process, if the proportion of high-frequency markings in a certain area within a consecutive time period exceeds 80%, the length of the consecutive time period is automatically extended to four time periods to accommodate traffic flow fluctuations.

[0073] The following details should be noted during implementation: the preset geographic grid density should be adjusted in real time according to changes in the microgrid topology. For example, when a new grid node is added, the associated geographic grid area should be automatically re-divided; the setting of the high-frequency threshold can be dynamically adjusted in conjunction with the microgrid load capacity. For example, the threshold can be reduced by 10% during peak electricity consumption periods to identify more potential high-frequency nodes; the length of the continuous time period in the spatiotemporal correlation verification can be adjusted according to actual needs. For example, it can be shortened to two consecutive time periods during peak traffic periods and extended to four consecutive time periods during off-peak periods.

[0074] S3. Based on the inter-node impedance matrix, a state transition equation is constructed, using the rate of change in line capacity caused by the movement of electric vehicles as a constraint condition. Node-level safety boundaries are set according to node voltage limits and line capacity thresholds, specifically:

[0075] A state transition equation is constructed based on the dynamic change characteristics of the inter-node impedance matrix. The numerical update results of the inter-node impedance matrix are dynamically correlated with the node positions in the electric vehicle trajectory data through weighting factors.

[0076] The weighting factor is calculated by updating the numerical values ​​of the impedance matrix between power grid nodes. (Represents the real-time impedance value from node x to node y, in ohms) and the electric vehicle access frequency within the corresponding geographic grid area. (Represents the number of electric vehicle moves from node x to node y, in times / hour) Perform linear weighting ( and Dimensionless processing was performed before linear weighting to generate spatiotemporally coupled state transition weights. The calculation formula is: Wherein, α and β are preset balance coefficients that satisfy α+β=1, for example, α is 0.6 and β is 0.4, which are used to adjust the contribution ratio of impedance value and moving frequency to weight; Obtained through real-time calculation in step S1c. The results are obtained through statistical analysis of the high-frequency access node identification results in step S2c.

[0077] By calculating the state transition weights of spatiotemporal coupling To achieve spatiotemporal coupling between the physical characteristics of the power grid and the mobile behavior of electric vehicles.

[0078] Based on the rate of change in line capacity caused by the movement of electric vehicles, and combining historical load data with line capacity thresholds, a linear constraint relationship is established between the rate of change in line capacity and the charging and discharging power of nodes. Specifically, for each line k, its rate of change in capacity is... With node charging and discharging power The relationship was fitted with historical data to obtain a linear equation: in, and The coefficient represents the fitting coefficient, which is obtained by regression analysis of the charging and discharging power and capacity change rate in historical operating data using the least squares method; k represents the line number.

[0079] For example, historical data analysis of routes in a certain commercial area yielded... Then when the node charging and discharging power hour,

[0080] Based on this linear relationship, dynamic constraints are generated: in, The capacity threshold (unit: kW) for line k is obtained through the grid topology parameters in step S1 to ensure that the adjustment of charging and discharging power will not cause the line capacity to exceed the limit.

[0081] Node-level safety boundaries are set based on node voltage limits and line capacity thresholds, including the allowable range of node voltage deviation and the tolerance for line capacity over-limit.

[0082] The allowable voltage deviation range is set to ±5% of the node voltage limit. For example, when the node's rated voltage is 380V, the allowable deviation is ±19V. The line capacity over-limit tolerance is set to 10% of the line capacity threshold. For example, when the capacity threshold is 100kW, temporary over-limit to 110kW is allowed. Safety boundaries are embedded in the state transition equation as hard constraints. Specifically, a penalty term is introduced into the state transition equation. When the node voltage deviation or line capacity over-limit exceeds the safety boundary, the penalty term becomes infinitely large, forcing the optimization algorithm to avoid such state transition paths. For example, if the voltage deviation of node x reaches 20V (exceeding the allowable 19V), the weight of the corresponding path in the state transition equation... It is set to infinity, thus disallowing the selection of this path.

[0083] It is worth noting that the balance coefficients α and β can be dynamically adjusted according to the microgrid's operating scenario. For example, β can be increased to 0.5 during peak hours to enhance the impact of electric vehicle movement frequency on the weighting; α can be increased to 0.7 during off-peak hours to prioritize grid impedance characteristics; and the balance coefficients can be refitted monthly based on the latest historical data. and The update cycle is synchronized with the microgrid operation and maintenance plan to ensure that the linear relationship is consistent with the actual operating status; under extreme weather conditions (such as high or low temperature), the tolerance for line capacity overload can be temporarily reduced to 5% to prevent equipment overload; in areas with frequent voltage fluctuations, the allowable range of voltage deviation is tightened to ±3%; when the voltage deviation or capacity overload exceeds the safety boundary, the current optimization calculation is immediately interrupted and the state transition equation is reinitialized to avoid invalid iteration. During reinitialization, the path weights that have not triggered penalty terms are retained first to improve computational efficiency.

[0084] By deeply integrating the dynamic changes of the impedance matrix between grid nodes with the spatiotemporal characteristics of electric vehicle trajectories, a spatiotemporally coupled state transition equation is constructed, solving the energy storage optimization mismatch problem caused by neglecting spatial dimension parameter differences in traditional dynamic programming methods. Specifically, the weighted factor dynamically correlates the impedance value with the electric vehicle's movement frequency, reflecting the impact of grid topology changes on charging and discharging strategies in real time; the linear constraint relationship based on historical data regression transforms the complex line capacity change rate into a calculable charging and discharging power limit, significantly reducing model complexity; the hard constraint mechanism of node-level safety boundaries, through dual control of the allowable voltage deviation range and capacity over-limit tolerance, ensures that the optimization path strictly conforms to the grid safety operation standards.

[0085] S4. A phased inverse dynamic programming algorithm is adopted, with node voltage deviation and line capacity exceeding limits as penalty terms. Spatial dimension optimization is performed first on high-frequency access nodes to generate node charging and discharging constraints, specifically:

[0086] The optimization process is divided into two stages: spatial dimension optimization and temporal dimension recursion. Based on the high-frequency access node distribution identified in step S2, the priority order of spatial dimension optimization is determined. The priority ranking rule is: based on the electrical position weight of the node in the power grid topology (calculated from the inter-node impedance matrix) and the comprehensive score of the access frequency.

[0087] For example, for high-frequency nodes located at the end of the main line (electrical weight ≥ 0.8) and with an access frequency ≥ 10 times / hour, the priority is raised to the highest level; for branch line nodes (electrical weight < 0.5) and with an access frequency < 5 times / hour, the priority is set to the lowest level.

[0088] After the priority list is generated, the space optimization problem of each node is processed in descending order of priority.

[0089] In the spatial dimension optimization phase, a reverse dynamic programming algorithm is used to recursively calculate from the lowest priority node. Node voltage deviation and line capacity exceeding limits are used as penalty terms to construct the objective function. The specific form of the objective function includes, but is not limited to, the following two:

[0090] Linear combination form: The voltage deviation penalty and the line capacity over-limit penalty are weighted and summed proportionally. The voltage deviation penalty is the percentage of the absolute difference between the actual voltage and the rated voltage to the rated voltage, and the line capacity over-limit penalty is the square of the ratio of the actual line current to the capacity threshold. The weighting ratio is dynamically adjusted according to the power grid topology parameters (such as the inter-node impedance matrix).

[0091] Piecewise function form: When the line capacity exceeds the limit by less than 10% of the threshold, a linear penalty coefficient is used; when the exceedance exceeds 10%, the penalty coefficient increases exponentially to strengthen the suppression of severe over-limits. The voltage deviation penalty term is set with asymmetrical weights according to the direction of deviation (too high or too low), for example, the weight for excessively high voltage is 1.2 times that for excessively low voltage.

[0092] For each high-frequency access node, its feasible charging and discharging power range (e.g., -1.5MW to +1.5MW) is traversed, and power ranges with penalty values ​​lower than a preset threshold are selected as initial charging and discharging constraints. For example, if the combined penalty value of all power points of a node within the power range [-1.2MW, +1.0MW] is lower than the preset threshold, then initial constraints are generated.

[0093] The penalty term weights are dynamically adjusted based on the inter-node impedance matrix in the power grid topology parameters. For nodes with tight electrical coupling (e.g., inter-node impedance less than 0.1 ohms), the voltage deviation penalty weight is increased; for lines with low capacity margins (e.g., line load rate exceeding 90%), the capacity over-limit penalty weight is increased. After adjustment, an iterative method is used to correct the constraint boundaries, for example, by gradually narrowing the feasible power range using a bisection method until the node-level safety boundaries (voltage deviation ≤ 5%, line capacity over-limit ≤ 10%) are met.

[0094] In the time-dimensional recursive phase, optimization constraints from all high-frequency access nodes are integrated. A global conflict detection rule is established: if the charging / discharging power constraint of a node causes the voltage of adjacent nodes to exceed limits or the line capacity to exceed limits, then the constraint is removed. For example, if the constraint of node A causes the voltage deviation of adjacent node B to exceed 5%, the power limit of node A is reduced to below the conflict threshold. Finally, a set of charging / discharging constraints consistent with the global optimization objective is generated.

[0095] It is worth noting that in the objective function parameter settings, the initial value of the voltage deviation penalty weight is set according to the electrical centrality of the node in the power grid. The higher the centrality (such as being close to the main transformer), the greater the weight. The capacity over-limit penalty weight is dynamically adjusted according to the historical load rate of the line. When the load rate exceeds 80%, the weight is increased by 20%.

[0096] In the dynamic rules for penalties, when the node voltage deviation exceeds 3% for three consecutive time periods, the voltage deviation penalty weight is increased by 20%; when the line capacity over-limit probability exceeds 15%, the capacity over-limit penalty weight is increased by 30%.

[0097] In the conflict resolution mechanism, a relaxation factor is introduced into the conflict constraints, allowing fine-tuning of the constraint boundaries within a preset tolerance range (such as ±5%), ensuring that local optimization is compatible with the global objective.

[0098] The Backward Dynamic Programming algorithm recursively calculates the current charging and discharging strategy based on the optimal solution of the subsequent stage, the node voltage deviation, and the line capacity over-limit penalty term in each stage, ensuring that the local decision is consistent with the global safety boundary, and finally generating the optimization result that satisfies the node-level constraints.

[0099] It is worth noting that node voltage limits are the preset safe operating range of grid node voltages (e.g., rated voltage ±10%), set by grid design specifications or management regulations, and used to define the allowable voltage fluctuation boundaries (e.g., 380V ± 38V). Node voltage deviation is the real-time difference between the actual operating voltage and the rated voltage (e.g., +25V or -20V), reflecting the degree to which the node voltage deviates from the rated value, and must be maintained within the limit range through monitoring and control measures.

[0100] It is worth noting that the line capacity threshold is the maximum power or current that the line is allowed to carry (e.g., 1000A), determined by line design parameters or safety regulations, and is used to ensure that equipment is not overloaded. Line capacity exceeding the limit refers to a state where the actual operating power or current exceeds the threshold (e.g., 1050A), which is an abnormal operating condition and needs to be restored to within the threshold by adjusting the charging and discharging strategy or disconnecting the load.

[0101] S5. Integrate node charging and discharging constraints, perform global optimization with the goal of minimizing total operating cost, and output a global strategy for energy storage capacity configuration in each time period, specifically:

[0102] Summarize the charging and discharging constraints of all high-frequency access nodes output in step S4 to generate a set of node-level charging and discharging feasible regions.

[0103] Each node constraint in the feasible region set includes upper and lower power limits and a time validity range. The upper and lower power limits are directly obtained from the spatial optimization results in step S4 (e.g., the upper power limit of node A is +1.2MW and the lower limit is -0.8MW). The time validity range is determined according to the time period marking of the high-frequency access node in step S2 (e.g., the charging and discharging operation of node B is only valid from 9:00 to 12:00 and from 18:00 to 21:00 every day).

[0104] The generation of the feasible region set requires verification of the compatibility between the constraints of each node. If the upper and lower power limits of adjacent nodes overlap (such as [-0.5MW, +1.0MW] for node C and [-0.8MW, +0.9MW] for node D), the overlapping parts are automatically merged to form a joint feasible region.

[0105] Based on the inter-node impedance matrix in the power grid topology parameters, a charging and discharging cost coefficient matrix is ​​constructed. Each element of the cost coefficient matrix represents the comprehensive impact weight of node charging and discharging power on power grid operating costs, including line loss costs, equipment depreciation costs, and peak-valley electricity price difference costs.

[0106] The line loss cost weight is calculated from the mutual impedance elements of the inter-node impedance matrix. For example, the greater the mutual impedance from node M to node N, the higher the line loss cost weight. The equipment depreciation cost weight is set according to the rated power and historical usage frequency of the node's charging and discharging equipment. For example, if the rated power of a node's charging and discharging equipment is 2MW and the average number of times it is used per day is ≥10, its depreciation cost weight is increased by 20%. The peak-valley electricity price difference weight is dynamically adjusted according to the real-time electricity price data collected in step S1. For example, the weight is reduced to 0.7 during the off-peak period (0:00-8:00) and increased to 1.3 during the peak period (18:00-22:00).

[0107] Using the minimization of total operating cost as the objective function, the charging and discharging cost coefficient matrix and the node-level feasible region set are input into a linear programming model to solve for a global charging and discharging power allocation scheme that satisfies time-period coupling constraints. Specifically, the objective function is to minimize the weighted sum of line loss cost, equipment depreciation cost, and peak-valley electricity price cost. Time-period coupling constraints include limits on the rate of power change between adjacent time periods (e.g., charging and discharging power change between adjacent time periods does not exceed ±20%), constraints on node charging and discharging power within the feasible region, and time-effectiveness constraints on high-frequency access nodes.

[0108] During the solution process, the simplex method is used to iteratively calculate the linear programming model until all constraints are met and the solution converges to the optimal solution. For example, the global allocation scheme for a certain time period requires that the charging and discharging power of node A be +0.9MW from 9:00 to 10:00, and that of node B be -0.3MW, with the power change rate of both nodes not exceeding 20%.

[0109] An N-1 safety check is performed on the global charging and discharging power allocation scheme. This involves simulating the power grid operation under any single line fault scenario to detect whether node voltage exceeds limits or line capacity is exceeded. The check rules include:

[0110] Node voltage verification: If the voltage deviation of a node exceeds its safety boundary (e.g., rated voltage ±5%) under fault conditions, the solution is deemed unqualified.

[0111] Line capacity verification: If the load rate of a line exceeds 10% of the capacity threshold under fault conditions (e.g., when the threshold is 100kW, the load is >110kW), it is deemed unqualified.

[0112] For unqualified schemes, the charging and discharging power allocation of the relevant nodes is eliminated or adjusted. For example, when a fault in line L1 causes the voltage deviation at node E to reach 6%, the power limit of node E is adjusted from +1.0MW to +0.8MW, and the global strategy is recalculated.

[0113] The final output is a global strategy for configuring energy storage capacity for each time period, which has passed verification. This includes the charging and discharging power plan, timing, and cost budget for each node.

[0114] It is worth noting that the interval intersection algorithm can be used to detect the overlap of power constraints between adjacent nodes. If the overlap interval ratio is less than 30%, independent constraints are maintained; if it is ≥30%, joint constraints are generated. In scenarios with frequent electricity price fluctuations, the peak-valley electricity price weight is updated every hour. During the equipment maintenance cycle (e.g., every quarter), the equipment depreciation cost weight is updated. For the charging and discharging power change rate limit, if the power change of a node is less than 5% in three consecutive time periods, the limit is relaxed to ±30% to improve scheduling flexibility. Priority is given to verifying fault scenarios of trunk lines and high-frequency access node associated lines. For example, full verification is performed on the associated lines of nodes with a daily access frequency of ≥20 times, and other lines are sampled and verified at a rate of 10%.

[0115] By integrating node-level charging and discharging constraints with multi-dimensional cost modeling, the global optimization strategy achieves synergistic improvements in safety, economy, and operability. The charging and discharging cost coefficient matrix integrates multiple factors such as line loss, equipment depreciation, and peak-valley electricity prices, solving the problem of strategy one-sidedness caused by single cost optimization in traditional methods. Time-coupling constraints (such as the power change rate limit between adjacent time periods) are embedded through a linear programming model to suppress drastic fluctuations in charging and discharging power and avoid losses from frequent equipment start-ups and shutdowns. The N-1 safety verification mechanism is deeply coupled with the global optimization process, eliminating high-risk strategies in fault scenarios in real time and significantly improving the robustness of the power grid.

[0116] S6. Generate an energy storage deployment plan based on the global strategy and the distribution of high-frequency access nodes, specifically:

[0117] By integrating the global energy storage capacity configuration strategy for each time period output in step S5 with the high-frequency access node distribution data generated in step S2, a node energy storage demand matrix is ​​generated. Each element of this matrix corresponds to a grid node and includes the following characteristics:

[0118] Capacity requirement: Calculated by summing the charging and discharging power requirements of the node during each time period in the global strategy. For example, if a node needs to provide a cumulative discharge capacity of 10MWh during the daily peak period (18:00-22:00), then the capacity requirement is 10MWh; during the off-peak period (0:00-8:00), 8MWh of charging capacity needs to be reserved.

[0119] Charging and discharging power requirements: Directly refer to the node charging and discharging power limits generated in step S4. For example, the upper limit of node A is +1.2MW (discharging) and the lower limit is -0.8MW (charging), and it is only valid from 9:00 to 12:00 and 18:00 to 21:00 daily;

[0120] Temporal distribution characteristics: The time validity range that matches high-frequency periods. For example, the high-frequency periods for node B are the morning and evening peak hours on weekdays (7:00-9:00 and 17:00-19:00), and its energy storage operations are only allowed to be carried out during these periods.

[0121] Based on the node energy storage demand matrix and the inter-node impedance matrix from step S1, the energy storage deployment priority of each node is calculated. The priority scoring rules are as follows:

[0122] The higher the access frequency during high-frequency periods, the greater the weight. For example, if node C accesses the network 15 times per hour during peak hours, the weight coefficient is set to 1.2; if node D accesses the network 5 times per hour, the weight coefficient is set to 0.6.

[0123] Based on the electrical position of the nodes in the power grid topology (calculated by the inter-node impedance matrix), the nodes closer to the main transformer (electrical centrality ≥ 0.8) are assigned a weight of 1.5, and the end nodes (electrical centrality ≤ 0.3) are assigned a weight of 0.5.

[0124] When a node's capacity demand exceeds 20% of its historical average, its priority score is increased by 10%; when the impedance between nodes changes due to grid upgrades, the electrical centrality weight is recalculated. For example, if a node's self-impedance decreases by 30% due to line upgrades, its electrical centrality weight increases from 0.7 to 0.9.

[0125] Using the node-level safety boundaries defined in step S3 as hard constraints (e.g., voltage deviation ≤ 5%, line capacity overrun ≤ 10%), and combining the charging / discharging power limits and adaptive priority rules generated in step S4, a multi-objective optimization algorithm is employed to generate an energy storage deployment scheme. The specific optimization logic includes:

[0126] Minimize the investment and operation and maintenance costs of energy storage equipment. For example, the unit capacity cost of lithium-ion batteries is 1000 yuan / kWh, while that of lead-acid batteries is 800 yuan / kWh. Prioritize the type with the best cost-effectiveness ratio.

[0127] To maximize charge and discharge efficiency, priority should be given to energy storage device types with an efficiency of ≥95%. For example, if the alternatives for a certain node include a lithium-ion battery with an efficiency of 97% and a flow battery with an efficiency of 92%, the former should be preferred.

[0128] For nodes with a priority score ≥ 0.8, at least one energy storage device must be deployed; for nodes with a score < 0.5, capacity is allocated flexibly according to demand. For example, node E has a priority score of 0.85 and needs to deploy an energy storage device with a capacity ≥ 5MWh; node F has a score of 0.4 and capacity is allocated according to actual demand.

[0129] Prioritize deploying energy storage devices in areas with low impedance and close to high-frequency access nodes. For example, if node G has an electrical centrality of 0.9 and is adjacent to three high-frequency access nodes, its energy storage device should be deployed at the grid junction with the lowest impedance.

[0130] The voltage stability of the energy storage deployment scheme is verified to ensure that the voltage deviation of all nodes is within the node-level safety boundary defined in step S3 (e.g., ±5%). Verification rules include:

[0131] Simulate full-power charging and discharging scenarios for energy storage devices to detect whether the node voltage exceeds the limit. For example, when node H is discharging at full power, the voltage drops to 363V (rated 380V), with a deviation of -4.47%, which meets the safety boundary; if the deviation reaches -5.3%, its power limit needs to be adjusted.

[0132] Simulating a scenario where a single energy storage device fails and goes out of service, the system checks whether the remaining devices can still meet voltage stability requirements. For example, if the primary energy storage device at node I fails, the backup device automatically switches over, and the node voltage deviation remains at 4.8%.

[0133] If the voltage deviation of a node exceeds the tolerance limit, the capacity or location of its energy storage device will be adjusted. For example, if the deviation of node J reaches 5.2%, its energy storage capacity will be increased from 3MWh to 3.5MWh, and it will be recalibrated until it meets the standard.

[0134] It is worth noting that during off-peak electricity pricing periods (0:00-8:00), the access frequency weight in the priority scoring can be reduced to 0.8 to encourage energy storage charging during off-peak hours; and during extreme weather warnings, the electrical centrality weight can be increased to 1.1 to strengthen coverage of key nodes.

[0135] When selecting energy storage equipment types, a life cycle cost model is introduced, prioritizing equipment with a lifespan of ≥10 years; when deploying geographically, GIS coordinates and power grid topology maps are combined to ensure the physical accessibility of the equipment.

[0136] For solutions that fail voltage stability verification, the deployment parameter adjustment process is automatically triggered, such as increasing capacity by 5% or tightening power limits by 10%. The adjusted solution needs to re-execute the full process verification until all nodes meet the safety boundaries.

[0137] It is worth noting that this embodiment effectively suppresses the risk of microgrid voltage fluctuations being transmitted to AC transmission networks above 750 kV by using node-level security boundaries and penalty constraints, providing spatial collaborative optimization support for large-scale power grid defense systems and enhancing the intelligent dispatch system's boundary control capabilities over distributed resource clusters.

[0138] The above formulas are all dimensionless calculations. The formulas are derived from software simulations using a large amount of collected data, and are the closest to the real situation. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0139] 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, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0140] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0141] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0142] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0143] 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.

[0144] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0145] 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.

[0146] 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 optimizing the energy storage capacity of V2G microgrids for electric vehicles based on dynamic programming, characterized in that, Includes the following steps: S1. Collect the grid topology parameters of the target microgrid and the electric vehicle movement trajectory data. The grid topology parameters include the inter-node impedance matrix, node voltage limits and line capacity thresholds. S2. Geographically divide the node locations in the electric vehicle movement trajectory data into grids, generate a spatiotemporal matrix by combining timestamps, and identify high-frequency access nodes. S3. Construct state transition equations based on the inter-node impedance matrix, take the rate of change of line capacity caused by the movement of electric vehicles as a constraint, and set node-level safety boundaries according to node voltage limits and line capacity thresholds. S4. A phased reverse dynamic programming algorithm is adopted, with node voltage deviation and line capacity overrun as penalty terms. Spatial dimension optimization is performed first for high-frequency access nodes to generate node charging and discharging constraints. S4 includes: S4a. The optimization phase is divided into spatial dimension optimization and time dimension recursion. The priority order of spatial dimension optimization is determined based on the distribution of high-frequency access nodes. S4b. In the spatial dimension optimization stage, the initial charging and discharging power constraints of the high-frequency access node are generated by using the node voltage deviation and line capacity over-limit as penalty terms. S4c: Dynamically adjust the weight of the penalty term based on the inter-node impedance matrix in the power grid topology parameters, iteratively correct the initial charging and discharging power constraints, and generate optimized constraints that satisfy the node-level safety boundary. S4d, in the time-dimensional recursive stage, integrates the optimization constraints of all high-frequency access nodes, eliminates local constraints that conflict with the global strategy, and generates the final node charging and discharging constraints. In the spatial dimension optimization stage, a reverse dynamic programming algorithm is used to recursively calculate from the lowest priority node, using node voltage deviation and line capacity overrun as penalty terms to construct an objective function. The specific form of the objective function includes: The voltage deviation penalty and the line capacity over-limit penalty are weighted and summed proportionally. The voltage deviation penalty is the percentage of the absolute difference between the actual voltage and the rated voltage to the rated voltage, and the line capacity over-limit penalty is the square of the ratio of the actual line current to the capacity threshold. The weighting ratio is dynamically adjusted according to the power grid topology parameters. S5. Integrate node charging and discharging constraints, perform global optimization with the goal of minimizing total operating costs, and output a global strategy for energy storage capacity configuration in each time period; S6. Generate an energy storage deployment plan based on the global strategy and the distribution of high-frequency access nodes.

2. The method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming according to claim 1, characterized in that, S1 includes: S1a. The power grid topology parameters are obtained in real time through the microgrid monitoring system. The power grid topology parameters include the inter-node impedance matrix, node voltage limits and line capacity thresholds. The node voltage limits and line capacity thresholds are extracted from the preset safe operation database of the microgrid management platform. S1b: Collect electric vehicle movement trajectory data through the vehicle positioning device. The movement trajectory data includes the node location and corresponding timestamp of the electric vehicle when it connects to the microgrid during a continuous period. S1c. Based on the inter-node connection relationship in the power grid topology parameters, calculate the numerical update result of the inter-node impedance matrix in real time, and perform spatial correlation mapping between the numerical update result of the inter-node impedance matrix and the node position in the electric vehicle movement trajectory data. S1d performs data verification on the numerical update results of node voltage limits, line capacity thresholds, and inter-node impedance matrices, and removes abnormal parameters that exceed the preset deviation range.

3. The method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming according to claim 1, characterized in that, S2 include: S2a. Dynamically divide the node locations in the electric vehicle movement trajectory data into grids according to the preset geographic grid density to generate geographic grid regions corresponding to the distribution of power grid nodes. S2b: Divide the electric vehicle access events in each geographic grid area into consecutive time periods according to timestamps, count the number of access events in each time period, and generate a spatiotemporal matrix; S2c: Based on preset clustering analysis rules, cluster the number of access events in the spatiotemporal matrix, identify geographical grid areas where the number of access events exceeds a preset high-frequency threshold, and mark them as high-frequency access nodes; S2d: Perform spatiotemporal correlation verification on high-frequency access nodes. If the same geographic grid area is marked as a high-frequency access node in several consecutive time periods, it is confirmed as a valid high-frequency access node.

4. The method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming according to claim 1, characterized in that, S3 include: S3a. Based on the dynamic change characteristics of the inter-node impedance matrix, a state transition equation is constructed. The numerical update result of the inter-node impedance matrix is ​​dynamically associated with the node position in the electric vehicle trajectory data through a weighting factor to generate a spatiotemporally coupled state transition weight. S3b. Based on the rate of change of line capacity caused by the movement of electric vehicles, and combined with historical load data and line capacity thresholds, establish a linear constraint relationship between the rate of change of line capacity and the charging and discharging power of nodes, and generate dynamic constraint conditions. S3c. Based on the node voltage limit and the line capacity threshold, the node-level safety boundary is set. The node-level safety boundary includes the allowable range of node voltage deviation and the tolerance of line capacity exceeding the limit. The node-level safety boundary is embedded in the state transition equation as a hard constraint.

5. The method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming according to claim 1, characterized in that, S5 include: S5a. Summarize the charging and discharging constraints of all high-frequency access nodes to generate a node-level charging and discharging feasible region set, which includes the upper and lower limits of power and the time validity range of each node. S5b. Construct a charging and discharging cost coefficient matrix based on the inter-node impedance matrix in the power grid topology parameters. The charging and discharging cost coefficient matrix includes the weights of the combined impact of node charging and discharging power on line loss, equipment depreciation and peak-valley electricity price. S5c. Taking the minimization of total operating cost as the objective function, the charging and discharging cost coefficient matrix and the node-level charging and discharging feasible domain set are input into the linear programming model to solve the global charging and discharging power allocation scheme that satisfies the time-period coupling constraint. S5d performs N-1 safety checks on the global charging and discharging power allocation scheme, eliminates schemes that cause node voltage to exceed limits or capacity to exceed limits when a single line fails, and outputs the global strategy for energy storage capacity configuration in each time period.

6. The method for optimizing the energy storage capacity of electric vehicle V2G microgrids based on dynamic programming according to claim 1, characterized in that, S6 include: S6a. Integrate the energy storage capacity configuration schemes for each time period and the distribution data of high-frequency access nodes in the global strategy to generate a node energy storage demand matrix. The node energy storage demand matrix includes the capacity demand, charging and discharging power demand of each node and the time distribution characteristics that match the high-frequency time period. S6b. Based on the node energy storage demand matrix and the inter-node impedance matrix, calculate the energy storage deployment priority of each node. The energy storage deployment priority is dynamically adjusted according to the access frequency during high-frequency periods and the electrical centrality of the nodes. S6c uses node-level security boundaries as constraints, combined with charging and discharging power limits and adaptive priority rules, to generate deployment schemes for the capacity, power limits, and geographical locations of energy storage devices using a multi-objective optimization algorithm; S6d: Perform node-level voltage stability verification on the energy storage deployment scheme, eliminate schemes that exceed the charging and discharging power limit or voltage deviation tolerance, and output the final energy storage deployment scheme.

Images