New energy charging optimization method and system based on multi-source data and storage medium

Abstract

The present application relates to the technical field of wisdom energy, specifically to a new energy charging optimization method and system based on multi-source data and a storage medium, the method comprising: cross-source mapping of real-time position, historical trajectory of new energy vehicles and real-time load data of power grid to build a multi-dimensional feature dataset; deducing the spatio-temporal distribution of future charging demand of a target area according to the multi-dimensional feature dataset to build a prediction atlas; dynamically adjusting the electricity price of charging stations in combination with real-time load to generate initial charging scheduling instructions; responding to closed-loop correction instructions according to the charging willingness of vehicles to obtain a target optimization scheme; monitoring the execution result of the target charging optimization scheme, and updating the multi-dimensional feature dataset according to the execution result and iteratively optimizing the charging demand prediction atlas and the electricity price system of charging stations when it is detected that the execution result deviates from the preset optimization target; the present application can improve the efficiency of new energy charging.

Description

Technical Field

[0001] This invention relates to the field of smart energy technology, and in particular to a new energy charging optimization method, system, and storage medium based on multi-source data. Background Technology

[0002] Current new energy charging management relies solely on single-dimensional data for scheduling, failing to conduct cross-source collaborative analysis of vehicle driving trajectories, location information, and grid load data. This makes it impossible to accurately depict the spatiotemporal evolution of charging demand, resulting in significant discrepancies between charging demand forecasts and actual conditions, and making it difficult to adapt to the real-time operating status of the power grid.

[0003] Traditional charging services use a fixed electricity price mechanism and lack dynamic adjustment methods based on demand and load. Charging dispatch instructions are not optimized in a closed loop according to the actual charging intentions of vehicles, resulting in insufficient flexibility in dispatch schemes. This can easily lead to problems such as grid load fluctuations and imbalances in charging resource allocation. Therefore, how to improve the optimization efficiency of new energy charging has become an urgent problem to be solved. Summary of the Invention

[0004] This disclosure provides a new energy charging optimization method, system, and storage medium based on multi-source data.

[0005] Firstly, this disclosure provides a new energy charging optimization method based on multi-source data, including: A1. Mapping the real-time location information and historical driving trajectory data of new energy vehicles with the real-time load data of the power grid across sources to obtain a multi-dimensional feature dataset between the new energy vehicles and the power grid; A2. Based on the multi-dimensional feature dataset, performing spatiotemporal distribution extrapolation of the charging demand distribution in the target area over future periods to construct a charging demand prediction map of the target area; A3. Based on the charging demand prediction map and the real-time load data, dynamically guiding the pricing system of charging stations in the target area to generate an initial charging scheduling instruction for the new energy vehicles; A4. Sending the initial charging scheduling instruction to the new energy vehicles, and performing closed-loop correction of the scheduling strategy based on the charging willingness response data fed back by the new energy vehicles to obtain a target charging optimization scheme for the new energy vehicles; A5. Monitoring the execution result of the target charging optimization scheme, and when the execution result is detected to deviate from the preset optimization target, updating the multi-dimensional feature dataset according to the execution result, and iteratively optimizing the charging demand prediction map and the pricing system of the charging stations.

[0006] In a preferred embodiment, the step of cross-source data mapping of the real-time location information and historical driving trajectory data of the new energy vehicle with the real-time load data of the power grid to obtain a multi-dimensional feature dataset between the new energy vehicle and the power grid includes: extracting the geographic grid code from the real-time location information of the new energy vehicle and obtaining the node load status corresponding to the geographic grid code in the real-time load data of the power grid to establish a spatial association mapping relationship between the new energy vehicle and the power grid; based on the spatial association mapping relationship, performing feature cross-coding on the travel pattern features in the historical driving trajectory data of the new energy vehicle and the time-period load fluctuation features in the real-time load data to obtain a spatiotemporal coupling feature vector between the new energy vehicle and the power grid; normalizing the spatiotemporal coupling feature vector and combining it with the remaining battery power features of the new energy vehicle and the load capacity features of the power grid nodes to construct a multi-dimensional feature dataset between the new energy vehicle and the power grid.

[0007] In a preferred embodiment, the step of extrapolating the spatiotemporal distribution of charging demand in the target area over future periods based on the multidimensional feature dataset and constructing a charging demand prediction map of the target area includes: extracting the vehicle aggregation evolution trend and the load change trend of power grid nodes within the target area based on the spatiotemporal coupled feature vector to obtain the basic driving factors of the target area; predicting the future charging probability of each geographical grid unit within the target area based on the basic driving factors to obtain a potential charging demand density distribution map of the target area; and overlaying the potential charging demand density distribution map with the load capacity features to construct the charging demand prediction map of the target area.

[0008] In a preferred embodiment, the basic driving factor is calculated using the following formula: In the formula, Geographic grid unit In the time window The fundamental driving factors For the geographic grid unit In the time window The evolution trend of vehicle aggregation. For the geographic grid unit and its adjacent grid cell set In the time window The sum of the trends in vehicle aggregation evolution, The preset load fluctuation impact factor, For the geographic grid unit In the time window The trend of load change For the geographic grid unit The load capacity characteristics of the corresponding power grid nodes.

[0009] In a preferred embodiment, the step of dynamically guiding the pricing of the electricity system of charging stations in the target area based on the charging demand forecast map and the real-time load data to generate the initial charging dispatch instructions for the new energy vehicles includes: classifying and evaluating the load carrying capacity of charging stations in the target area according to the spatiotemporal distribution characteristics in the charging demand forecast map to obtain the basic pricing range of the charging stations; mapping the peak and valley time information of the power grid in the real-time load data to the electricity price fluctuation adjustment factor of the charging stations; weighting and superimposing the basic pricing range and the electricity price fluctuation adjustment factor to obtain the dynamic time-of-use electricity price system of the charging stations; and providing economic incentives to guide the charging behavior of the new energy vehicles based on the dynamic time-of-use electricity price system to obtain the initial charging dispatch instructions for the new energy vehicles.

[0010] In a preferred embodiment, the step of providing economic incentives to guide the charging behavior of the new energy vehicle based on the dynamic time-of-use pricing system to obtain the initial charging scheduling instruction for the new energy vehicle includes: acquiring the electricity price period division nodes and the charging unit price value corresponding to each electricity price period in the dynamic time-of-use pricing system; extracting the remaining battery power value and expected parking duration from the vehicle status data of the new energy vehicle; performing a matching analysis between the charging unit price value and the expected parking duration to determine the target electricity price period set covered by the new energy vehicle within the expected parking duration; generating a quantitative value of the charging cost difference of the new energy vehicle in each electricity price period based on the remaining battery power value and the charging unit price value of each electricity price period in the target electricity price period set; comparing the quantitative value of the charging cost difference with a preset incentive threshold, and selecting the electricity price period in the target electricity price period set where the quantitative value of the charging cost difference is lower than the incentive threshold as the economic incentive period; and encapsulating the economic incentive period and the corresponding charging unit price value into the initial charging scheduling instruction for the new energy vehicle.

[0011] In a preferred embodiment, the step of sending the initial charging scheduling instruction to the new energy vehicle and, based on the charging intention response data fed back by the new energy vehicle, performing closed-loop correction of the scheduling strategy for the initial charging scheduling instruction to obtain the target charging optimization scheme for the new energy vehicle includes: issuing the initial charging scheduling instruction to the new energy vehicle and receiving the charging intention response data fed back by the new energy vehicle; comparing the desired charging period in the charging intention response data with the suggested charging period in the initial charging scheduling instruction to obtain the time period deviation result between the desired charging period and the suggested charging period; performing power difference analysis between the desired charging power value in the charging intention response data and the suggested charging power value in the initial charging scheduling instruction to obtain the power deviation result between the desired charging power value and the suggested charging power value; and reconstructing the scheduling strategy for the suggested charging period and suggested charging power value in the initial charging scheduling instruction based on the time period deviation result and the power deviation result to obtain the target charging optimization scheme for the new energy vehicle.

[0012] In a preferred embodiment, monitoring the execution result of the target charging optimization scheme, and updating the multidimensional feature dataset based on the execution result when the execution result deviates from the preset optimization target, and iteratively optimizing the charging demand prediction map and the electricity price system of the charging station, includes: acquiring actual charging execution data generated during the execution of the target charging optimization scheme, the actual charging execution data including actual charging time period, actual charging power value, and actual charging completion time; comparing and analyzing the actual charging execution data with the expected charging time period, expected charging power value, and expected charging completion time in the target charging optimization scheme to obtain the execution deviation data of the target charging optimization scheme; when the execution deviation data exceeds the tolerance threshold in the preset optimization target, adding the execution deviation data as a feedback feature to the multidimensional feature dataset to generate an updated multidimensional feature dataset; and iteratively correcting the charging demand prediction map and the electricity price system of the charging station based on the updated multidimensional feature dataset.

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] 1. This invention constructs a multi-dimensional feature dataset of new energy vehicles and the power grid through cross-source mapping of multi-source data. It accurately realizes the spatiotemporal coupling correlation between vehicle location, driving trajectory, and power grid load. Based on this dataset, it completes the spatiotemporal distribution projection of charging demand in the target area, constructing an accurate charging demand prediction map. This provides scientific data support for charging scheduling, ensuring that the charging demand prediction aligns with actual operating scenarios. Based on the prediction map and real-time power grid load, it dynamically optimizes the charging station electricity price system. Dynamic pricing forms economic incentives, generating initial charging scheduling instructions adapted to the power grid load. This effectively balances the regional charging load distribution, optimizes the peak-valley load fluctuation state of the power grid, achieves efficient matching between charging resources and power grid carrying capacity, and improves the stability of power grid operation and the rationality of charging resource allocation.

[0015] 2. This invention utilizes vehicle charging willingness response data to complete the closed-loop correction of dispatch instructions, generating a target charging optimization scheme that aligns with the actual needs of vehicles, significantly improving the adaptability and execution efficiency of the dispatch scheme. Simultaneously, by monitoring the scheme execution results in real time, it updates the multi-dimensional feature dataset promptly when deviations occur, and iteratively optimizes the charging demand prediction map and electricity pricing system, forming a continuously iterative intelligent optimization mechanism to ensure long-term stable achievement of charging dispatching results. The entire technology achieves intelligent management and control of the entire new energy charging dispatching process, improving quality and efficiency across all stages from data fusion, demand forecasting, pricing guidance to strategy correction and iterative optimization, comprehensively enhancing the overall operational efficiency of new energy charging, strengthening the collaborative operation capabilities of the power grid and charging services, and providing stable and reliable technical support for the efficient development of the new energy charging industry. Attached Figure Description

[0016] The present disclosure will be described in more detail below based on embodiments and with reference to the accompanying drawings:

[0017] Figure 1 The flowchart of the new energy charging optimization method based on multi-source data according to Embodiment 1 of the present invention is shown.

[0018] Figure 2 The diagram shows the functional modules of the new energy charging optimization system based on multi-source data according to Embodiment 2 of the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the technical solutions of this disclosure, and to fully understand and implement the process of how this disclosure applies technical means to solve technical problems and achieve corresponding technical effects, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, not all embodiments. The embodiments of this disclosure and the various features within them can be combined with each other without conflict, and the resulting technical solutions are all within the protection scope of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort should fall within the protection scope of this disclosure.

[0020] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0021] Example 1

[0022] Figure 1 This is a flowchart illustrating the new energy charging optimization method based on multi-source data provided in an embodiment of this disclosure. Figure 1 As shown, a new energy charging optimization method based on multi-source data includes: A1. Mapping the real-time location information and historical driving trajectory data of new energy vehicles with the real-time load data of the power grid across sources to obtain a multi-dimensional feature dataset between the new energy vehicles and the power grid; In this embodiment of the invention, the step of mapping the real-time location information and historical driving trajectory data of new energy vehicles with the real-time load data of the power grid to obtain a multi-dimensional feature dataset between the new energy vehicles and the power grid includes: extracting the geographic grid code from the real-time location information of the new energy vehicles and obtaining the node load status corresponding to the geographic grid code in the real-time load data of the power grid to establish a spatial association mapping relationship between the new energy vehicles and the power grid; based on the spatial association mapping relationship, performing feature cross-coding on the travel pattern features in the historical driving trajectory data of the new energy vehicles and the time-period load fluctuation features in the real-time load data to obtain a spatiotemporal coupling feature vector between the new energy vehicles and the power grid; normalizing the spatiotemporal coupling feature vector and combining it with the remaining battery power features of the new energy vehicles and the load capacity features of the power grid nodes to construct a multi-dimensional feature dataset between the new energy vehicles and the power grid.

[0023] The system continuously collects real-time location information from the on-board positioning terminal of the new energy vehicle. It accurately extracts the longitude and latitude values ​​of the vehicle's current location from this location information. According to the pre-set geographic grid division rules, the entire urban power supply area is divided into regular rectangular geographic grids with uniform size and clear boundaries. The extracted longitude and latitude values ​​are compared with the boundary range of each geographic grid in turn. The unique geographic grid code corresponding to the current location of the vehicle is determined by matching the coordinates with the grid number one by one.

[0024] Real-time load data of the power grid covering the entire power supply service area is retrieved from the power grid dispatch and operation management system. This data includes the number of all power grid nodes and the load information at the corresponding time. According to the pre-established correspondence table between power grid nodes and geographic grids, power grid nodes that are completely consistent with the geographic grid code of the new energy vehicle are selected. The actual load value of the power grid node at the current collection time is read, and the value is directly determined as the node load status of the corresponding geographic grid code. By binding and recording the vehicle geographic grid code with the power grid node load status, the spatial association mapping relationship between the new energy vehicle and the power grid is established.

[0025] Based on the established spatial correlation mapping relationship, historical driving trajectory data of new energy vehicles within a preset 30-day time period are extracted from the vehicle data management platform. The continuous trajectory data is split into segments at fixed time intervals of 15 minutes. The number of trips of vehicles in each time period of the day and the duration of stay in the corresponding geographic grid are counted. The statistical values ​​are combined to form the characteristics of vehicle travel patterns.

[0026] Extract time-period load data that is completely consistent with the above-mentioned 30-day time cycle from the real-time load data of the power grid. Divide the load data according to the same fixed time interval of 15 minutes, calculate the load value difference between adjacent time nodes, and statistically analyze the load change amplitude in each time period. Combine the statistically obtained values ​​to form the time-period load fluctuation characteristics.

[0027] The travel pattern characteristics and the corresponding time period load fluctuation characteristics are matched and combined one by one according to the same time node and the same geographical grid. The two types of feature content are arranged in order of time sequence and spatial correspondence. Feature cross-coding is completed by feature splicing to generate a spatiotemporal coupling feature vector that can simultaneously reflect the temporal and spatial correlation between new energy vehicles and the power grid.

[0028] The various feature values ​​in the spatiotemporal coupled feature vector are scaled according to a unified numerical range. The scaling benchmark is the maximum and minimum values ​​of the feature within a preset 30-day time period. The feature value is subtracted from the minimum value of the corresponding category and then divided by the difference between the maximum and minimum values ​​of the category, so that all feature values ​​fall within a fixed range of zero to one, thus completing the normalization process.

[0029] The system collects the actual value of the current remaining battery power uploaded by the battery management system of new energy vehicles, calculates the ratio between the actual value and the total battery capacity value specified by the vehicle at the factory, and uses the ratio result as the characteristic of the remaining battery power.

[0030] The system collects the pre-set rated load capacity value of the corresponding node in the power grid dispatching system, calculates the difference between the rated load capacity value and the actual load value of the node at the current time, and uses the difference result as the load capacity characteristic of the power grid node.

[0031] The normalized spatiotemporal coupling feature vector, the calculated remaining battery power feature, and the load capacity feature of the grid node are integrated and arranged in a fixed order. All feature content is stored in a structured data format to construct a multi-dimensional feature dataset between new energy vehicles and the power grid.

[0032] This step accurately establishes a spatial mapping relationship between new energy vehicles and grid nodes, achieving precise matching between vehicle location and grid load, avoiding spatial correspondence deviations. By combining travel patterns with spatiotemporal characteristics of load fluctuations through cross-coding, it can fully reflect the coupling state between vehicles and the grid in both time and space dimensions, providing a comprehensive feature foundation for vehicle-grid interaction analysis. Combined with the normalization and integration of multi-dimensional features, it eliminates the impact of differences in the magnitude of different feature values, forming a structurally complete multi-dimensional feature dataset that can be directly used for subsequent analysis. This effectively improves the accuracy and reliability of new energy vehicle-grid interaction analysis, providing stable and reliable data support for orderly vehicle charging and discharging scheduling and grid peak shaving and valley filling control. Simultaneously, it ensures that the entire data processing process is clear, reproducible, and free of ambiguous judgment steps, avoiding data processing deviations and improving the control precision of vehicle-grid collaborative operation.

[0033] A2. Based on the multidimensional feature dataset, the spatiotemporal distribution of charging demand in the target area during future periods is extrapolated to construct a charging demand prediction map of the target area. In this embodiment of the invention, the step of extrapolating the spatiotemporal distribution of charging demand in the target area during future periods based on the multidimensional feature dataset to construct a charging demand prediction map of the target area includes: extracting the vehicle aggregation evolution trend and the load change trend of the power grid nodes within the target area based on the spatiotemporal coupled feature vector to obtain the basic driving factors of the target area; predicting the future charging probability of each geographical grid unit within the target area based on the basic driving factors to obtain a potential charging demand density distribution map of the target area; and overlaying the potential charging demand density distribution map with the load capacity features to construct a charging demand prediction map of the target area.

[0034] The formula for calculating the basic driving factor is as follows: In the formula, Geographic grid unit In the time window The fundamental driving factors For the geographic grid unit In the time window The evolution trend of vehicle aggregation. For the geographic grid unit and its adjacent grid cell set In the time window The sum of the trends in vehicle aggregation evolution, The preset load fluctuation impact factor, For the geographic grid unit In the time window The trend of load change For the geographic grid unit The load capacity characteristics of the corresponding power grid nodes.

[0035] Based on the generated spatiotemporal coupling feature vector, the specific boundary range of the target area is first defined. This range is delineated based on the power grid power supply zone, and the boundary coordinates are determined according to the preset power supply zone boundary in the power grid dispatch system, covering all geographic grid units and corresponding power grid nodes within the target area.

[0036] The location and time association information of all new energy vehicles in the target area are extracted from the spatiotemporal coupling feature vector. The location association information includes the geographic grid code to which the vehicle belongs, and the time association information includes the time of collection of the vehicle positioning data. At a fixed interval of 15 minutes, all geographic grid units are traversed and the number of vehicles in each grid in each time slice is counted. The total number of vehicles corresponding to each time slice is arranged in chronological order to form a continuous and traceable trend of vehicle clustering evolution in the target area.

[0037] Simultaneously, load-related features corresponding to all power grid nodes within the target area are extracted from the spatiotemporal coupling feature vector. The load-related features include the real-time load values ​​of each power grid node and its corresponding geographic grid code. Similarly, at a fixed interval of 15 minutes per time slice, the actual load values ​​of each power grid node within each time slice are read. The load values ​​of all power grid nodes within the target area are added together and divided by the total number of power grid nodes to obtain the average load of that time slice. The average load values ​​of each time slice are arranged sequentially according to time to form a continuous and traceable trend of power grid node load changes.

[0038] The obtained vehicle aggregation evolution trend and the load change trend of the power grid node are integrated one by one according to the same time slice. The data set arranged in time order after integration is the basic driving factor of the target area. The vehicle aggregation evolution trend corresponds to the vehicle number statistics of the geographic grid unit in the corresponding time window. This result is obtained by summarizing the vehicle location and time correlation information of the corresponding grid unit vehicle by vehicle.

[0039] The sum of the vehicle aggregation trends of a geographic grid cell and its adjacent grid cells is obtained by first determining all geographic grid cells that are directly adjacent to the central grid in both horizontal and vertical directions to form a set of adjacent grid cells, then extracting the vehicle aggregation trend values ​​of the central grid and its adjacent grid cells within the same time window, and finally summing all the extracted values.

[0040] The load fluctuation impact factor is determined by retrieving complete power grid load fluctuation data and vehicle aggregation data for the past 30 days in the target area, statistically analyzing the correlation between load change magnitude and vehicle charging behavior in each time period, and averaging the results of multiple statistical analyses to determine a fixed value. This value is specifically used to quantify the impact weight of load change trends on the basic driving factor, ensuring that the contribution of load fluctuations to the driving factor is within a reasonable range.

[0041] The load change trend is obtained by extracting the load-related features of the corresponding power grid node of the geographic grid unit from the spatiotemporal coupling feature vector. The actual load value of the node is read in a 15-minute time window. The trend is statistically analyzed by combining the load values ​​of the previous time window and the next time window to obtain the load change trend of the grid unit in the corresponding time window. The result is completely consistent with the load change trend of the previously generated power grid node.

[0042] The load capacity characteristics of power grid nodes are obtained by collecting the rated load capacity values ​​of the corresponding nodes from the power grid dispatch system, reading the actual load value of the node at the current moment, and subtracting the current actual load value directly from the rated load capacity value. The difference obtained is the corresponding feature. This result is completely consistent with the load capacity characteristics of power grid nodes obtained when constructing the multidimensional feature dataset.

[0043] The calculation formula normalizes the vehicle aggregation trend by dividing the evolution trend of a single grid by the sum of its own and neighboring grid aggregation, thereby eliminating the calculation bias caused by the difference in the number of vehicles between different geographical grid units. At the same time, the ratio of load change trend to load capacity characteristics is combined with load fluctuation impact factors for weighted calculation, quantifying the actual impact of grid load fluctuations on the basic driving factors. This ensures that the final calculated basic driving factors can accurately reflect the vehicle aggregation status and grid load operation status of geographical grid units in the corresponding time window, providing accurate and comprehensive core driving data for the prediction of charging probability distribution of each geographical grid unit in the target area in the future. This forms a complete logical closed loop with the previously constructed spatial correlation mapping relationship and spatiotemporal coupling feature vector, ensuring that the entire data processing flow is coherent, unified, and closely aligned with the core characteristics of vehicle-grid coupling.

[0044] Based on the obtained fundamental driving factors, these factors are used as key input features and input into a pre-trained time-series prediction model to perform AI-based inference on the charging demand distribution of the target area in future periods. The time-series prediction model can employ a Long Short-Term Memory (LSTM) network, a Gated Recurrent Unit (GRU) network, or a Random Forest Regression model. Its training samples are derived from a multi-dimensional feature dataset of historical records and the corresponding actual charging locations and times. Specifically, the future prediction period is defined as the next 24 hours, and the time is divided into independent time slices of 15 minutes. Each geographical grid unit within the target area is then used as a separate prediction unit for calculation.

[0045] By combining the vehicle aggregation evolution trend in the basic driving factors, the vehicle number records of each geographic grid unit in the same time slice of the same historical period are retrieved, and the vehicle number change pattern in the continuous period is analyzed to determine the expected vehicle number of the geographic grid unit in each corresponding time slice in the future.

[0046] By combining the travel pattern characteristics extracted from the historical driving trajectory data of new energy vehicles, the total number of times vehicles in the same geographical grid unit engaged in charging behavior within the same historical time slice is counted. The charging probability of a single vehicle within the time slice is obtained by dividing the total number of charging times by the total number of vehicles appearing in the same historical time slice.

[0047] The overall charging probability of a geographic grid cell in the future time slice is obtained by multiplying the calculated charging probability of a single vehicle by the expected number of vehicles in the corresponding future time slice.

[0048] The charging probabilities calculated for each geographic grid cell in each future time slice are matched one-to-one with the spatial coordinates of each grid cell. A fixed gradient color is used to indicate the magnitude of the charging probability values: light gray for charging probabilities in the range of 0 to 0.2, light blue for the range of 0.2 to 0.4, blue for the range of 0.4 to 0.6, dark blue for the range of 0.6 to 0.8, and black for the range of 0.8 to 1.0. This forms a visualized map of the potential charging demand density distribution in the target area.

[0049] Using the generated potential charging demand density distribution map as the base map, the load capacity characteristics of the corresponding power grid node for each geographical grid unit in the target area are extracted one by one. This characteristic is the difference between the rated load capacity of the corresponding node and the current actual load value.

[0050] The load capacity characteristic value corresponding to each geographic grid unit is clearly marked at the center of the corresponding grid unit in the potential charging demand density distribution map. The marked numbers are uniformly retained with one integer place and one decimal place to ensure that the numerical display is clear and easy to read.

[0051] At the same time, the original gradient color markings in the potential charging demand density distribution map are fully preserved, so that each geographical grid unit can intuitively reflect the potential charging demand density through color, and the load capacity of the corresponding grid node can be directly identified through the labeled numbers.

[0052] The superimposed graphics are standardized and uniformly processed, and the font size and position of the charging probability range corresponding to the color labels and the load capacity characteristic values ​​are standardized to ensure that the charging demand information and load capacity information of each geographic grid unit are displayed independently, clearly distinguishable and without mutual overlap and obstruction, and finally a complete charging demand prediction map of the target area is constructed.

[0053] In a specific implementation scenario of the present invention, the charging behavior within the target area includes at least the following four typical scenarios: Home scenario: Vehicles are parked for extended periods at night, typically >6 hours, with highly predictable charging demand and a time window granularity of 1 hour; Office scenario: Vehicles are parked continuously during the day, typically 4-8 hours, with the charging window coinciding with working hours and a time window granularity of 30 minutes; Public place scenario: Vehicles are parked irregularly, typically 1-3 hours, with charging probability related to the business district and parking lot saturation; Driving scenario: Vehicles are not parked, requiring dynamic identification of their charging urgency, with a time window granularity of 5 minutes based on remaining battery power and distance to the destination; For different scenarios, the time window granularity for future periods is dynamically adjusted, and corresponding granularities are used to predict the charging probability distribution.

[0054] This step extracts and accurately calculates the fundamental driving factors of the target area based on spatiotemporal coupled feature vectors, comprehensively and realistically capturing the temporal variation patterns of vehicle aggregation and grid load. This provides stable and reliable data support for subsequent charging probability prediction. By extrapolating historical patterns and performing numerical calculations, the future charging probability of each geographical grid unit is obtained. The resulting potential charging demand density distribution map can intuitively and clearly present the spatial distribution and temporal changes of charging demand in the target area. This map is then visualized and overlaid with the load capacity characteristics of grid nodes. The constructed charging demand prediction map can simultaneously reflect the distribution status of potential charging demand and the upper limit of grid carrying capacity, intuitively reflecting the matching relationship between charging demand and grid load in each grid unit. This provides accurate and implementable decision-making basis for the rational layout of charging facilities, precise peak shaving and valley filling of grid load, and orderly charging and discharging scheduling of new energy vehicles. The entire process is clearly defined, the numerical calculations are reproducible, and there are no ambiguous judgment steps, effectively improving the accuracy and practicality of charging demand prediction and ensuring the stability and efficiency of vehicle-grid collaborative operation.

[0055] A3. Based on the charging demand forecast map and the real-time load data, dynamically guide the pricing of the electricity price system of charging stations in the target area to generate the initial charging dispatch instruction for the new energy vehicle; In this embodiment of the invention, the dynamic pricing guidance of the electricity price system of charging stations in the target area based on the charging demand forecast map and the real-time load data to generate the initial charging dispatch instruction for the new energy vehicle includes: classifying and evaluating the load carrying capacity of charging stations in the target area according to the spatiotemporal distribution characteristics in the charging demand forecast map to obtain the basic pricing range of the charging station; mapping the peak and valley time information of the power grid in the real-time load data to the electricity price fluctuation adjustment factor of the charging station; weighting and superimposing the basic pricing range and the electricity price fluctuation adjustment factor to obtain the dynamic time-of-use electricity price system of the charging station; and providing economic incentive guidance for the charging behavior of the new energy vehicle based on the dynamic time-of-use electricity price system to obtain the initial charging dispatch instruction for the new energy vehicle.

[0056] The method of providing economic incentives to guide the charging behavior of new energy vehicles based on the dynamic time-of-use pricing system to obtain initial charging scheduling instructions for the new energy vehicles includes: acquiring the electricity price period division nodes and the charging unit price value corresponding to each electricity price period in the dynamic time-of-use pricing system; extracting the remaining battery power value and expected parking duration from the vehicle status data of the new energy vehicles; performing matching analysis between the charging unit price value and the expected parking duration to determine the target electricity price period set covered by the new energy vehicles within the expected parking duration; generating a quantitative value of the charging cost difference of the new energy vehicles in each electricity price period based on the remaining battery power value and the charging unit price value of each electricity price period in the target electricity price period set; comparing the quantitative value of the charging cost difference with a preset incentive threshold, and selecting the electricity price period in the target electricity price period set where the quantitative value of the charging cost difference is lower than the incentive threshold as the economic incentive period; and encapsulating the economic incentive period and the corresponding charging unit price value into an initial charging scheduling instruction for the new energy vehicles.

[0057] Based on the constructed charging demand forecast map, the spatiotemporal distribution characteristics of each charging station's coverage geographic grid unit are extracted. These characteristics include the charging demand density values ​​for each time period and the corresponding grid node load capacity characteristics. Hardware parameters such as the rated charging power and the number of charging piles for each charging station are also collected. The charging station load capacity is divided into three levels: Level 1 corresponds to a charging demand density of 0.6 to 1.0 and a grid load capacity greater than 100kW; Level 2 corresponds to a charging demand density of 0.3 to 0.6 and a grid load capacity of 50kW to 100kW; and Level 3 corresponds to a charging demand density of 0 to 0.3 and a grid load capacity less than 50kW. A corresponding basic pricing range is set according to different load capacity levels: Level 1 is 1.0 yuan / kWh to 1.2 yuan / kWh, Level 2 is 0.8 yuan / kWh to 1.0 yuan / kWh, and Level 3 is 0.6 yuan / kWh to 0.8 yuan / kWh. This basic pricing range is the basic pricing range for the charging stations.

[0058] Real-time load data from the power grid dispatching system is retrieved to extract peak-valley time period division information. The division criteria are defined as follows: peak periods are 7:00-10:00 and 18:00-21:00; flat periods are 10:00-18:00 and 21:00-23:00; and valley periods are 23:00-7:00 the next day. This division standard is determined based on historical power grid load data. During peak periods, the actual power grid load reaches more than 80% of the rated load; during flat periods, it reaches 50% to 80%; and during valley periods, it is below 50%. Each time period is mapped to a corresponding electricity price adjustment factor. The adjustment factor is set to 1.2 for peak periods, 1.0 for flat periods, and 0.8 for valley periods. The value of the adjustment factor is determined based on the load pressure of the power grid during each time period; the higher the load pressure, the higher the adjustment factor, thereby achieving linkage between electricity prices and power grid load.

[0059] The basic pricing range for charging stations is weighted and superimposed with the corresponding time-period electricity price fluctuation adjustment factor. The superposition method involves multiplying the upper and lower limits of the basic pricing range by the corresponding time-period electricity price fluctuation adjustment factor to obtain the charging unit price range for each time period. All time-period charging unit price ranges are then integrated in chronological order to clarify the unit price range and corresponding time nodes for each time period, forming a complete dynamic time-of-use electricity price system for charging stations. This system is directly related to the peak and off-peak periods of the power grid and the load-bearing capacity of charging stations, ensuring that electricity price adjustments closely reflect the actual vehicle-grid operating status.

[0060] From the generated dynamic time-of-use electricity pricing system, the division nodes and corresponding charging unit prices for each electricity price period are extracted one by one. The division nodes are marked with the hour and half-hour to clearly define the start and end times of each electricity price period. For example, the division nodes for the peak period of 7:00-10:00 are 7:00 and 10:00. The corresponding charging unit price is the midpoint of the unit price range for that period. The unit price for a Level 1 charging station during peak hours is 1.1 yuan / kWh, Level 2 is 0.9 yuan / kWh, and Level 3 is 0.7 yuan / kWh. All the division nodes and unit price values ​​for all periods are organized into an ordered set to ensure that the information for each period is complete and searchable.

[0061] Vehicle status data is collected from the on-board battery management system of new energy vehicles. The current remaining battery power value is extracted from it. This value is the ratio of the current actual battery power to the total battery capacity, rounded to one decimal place, for example, 50.0%. At the same time, by analyzing the dwell time in the current geographic grid cell, the current time and destination planning information in the vehicle's historical driving trajectory data, the estimated dwell time of the vehicle is determined. The estimated dwell time is accurate to the minute, for example, 120 minutes, to ensure that the data is accurate and traceable.

[0062] The extracted charging unit price values ​​for each electricity price period are matched and analyzed with the expected parking time of the vehicle. First, the start and end times of the expected parking time of the vehicle are determined. Then, the electricity price period division nodes in the dynamic time-of-use electricity price system are compared to determine all the electricity price periods covered by the parking period. These covered electricity price periods are integrated into a set, which is the target electricity price period set covered by the new energy vehicle within the expected parking time. This ensures that each covered period can be accurately matched without omission or redundancy.

[0063] Based on the vehicle's current remaining battery charge, determine the required charging amount. The required charging amount is the total battery capacity minus the current actual charge. Then, combine the charging unit price values ​​for each electricity price period in the target electricity price period set, calculate the total charging cost for the vehicle to complete the required charging amount in each electricity price period. Calculate the difference between the total charging cost for each period and the average charging cost for all periods. The resulting difference is the quantified value of the charging cost difference in each electricity price period. The quantified value is rounded to two decimal places. A positive number indicates that the cost for that period is higher than the average level, and a negative number indicates that it is lower than the average level.

[0064] The preset charging cost difference incentive threshold is 2.0 yuan. This threshold is determined based on statistical analysis of historical charging cost-sensitive data of new energy vehicle users, ensuring that cost differences below this threshold can effectively incentivize users to adjust their charging times. The quantified charging cost difference value for each electricity price period is compared one by one with the 2.0 yuan incentive threshold. Electricity price periods with a quantified charging cost difference value below 2.0 yuan are selected and integrated as the economic incentive periods for new energy vehicles, ensuring that the selection criteria for incentive periods are clear and reproducible.

[0065] The selected economic incentive periods are integrated with the corresponding charging unit price values ​​to clarify the start and end times and corresponding charging unit price of each economic incentive period. At the same time, the required charging amount of the vehicle is supplemented with information. This information is organized and packaged in a fixed format to form the initial charging scheduling instruction for new energy vehicles. This instruction can be directly sent to the vehicle's on-board terminal to guide users to charge during the economic incentive period, ensuring that the instruction information is complete and executable.

[0066] In addition, the initial charging scheduling command also includes an optional V2G bidirectional scheduling mode: when it is predicted that the vehicle will successively experience "peak electricity price period" and "off-peak electricity price period" during the current parking period, and the vehicle's current battery remaining charge is higher than a preset high threshold of 80%, the dynamic pricing module further generates a combined scheduling suggestion, including: sending power back to the grid during peak hours (discharging) to obtain electricity sales revenue; and then charging to full charge at a low price during off-peak hours, thereby obtaining net revenue through the price difference. Enabling the V2G mode requires prior authorization from the vehicle user and must comply with the grid's scheduling constraints on distributed power source access.

[0067] This process combines charging demand forecasting maps to assess the load-bearing capacity of charging stations and determine basic pricing ranges. By integrating peak-valley time-of-use pricing factors with grid peak-valley time-of-use periods, a dynamic time-of-use pricing system is constructed. This system enables precise linkage between charging station electricity prices and vehicle-grid operation status. By matching vehicle parking duration with electricity price periods and quantifying charging cost differences, economic incentive periods are selected and initial charging dispatch instructions are encapsulated. This not only guides new energy vehicle users to charge during off-peak grid periods, reducing user charging costs, but also optimizes grid load allocation, alleviating grid pressure during peak hours. Furthermore, the entire process is clear, standardized, and reproducible, with no ambiguous decision-making steps, effectively improving the accuracy and efficiency of vehicle-grid coordinated dispatch and ensuring a dual improvement in grid operation stability and the economic benefits of new energy vehicle charging.

[0068] A4. Sending the initial charging scheduling instruction to the new energy vehicle, and performing closed-loop correction of the scheduling strategy based on the charging willingness response data fed back by the new energy vehicle to obtain the target charging optimization scheme for the new energy vehicle; In this embodiment of the invention, sending the initial charging scheduling instruction to the new energy vehicle, and performing closed-loop correction of the scheduling strategy based on the charging willingness response data fed back by the new energy vehicle to obtain the target charging optimization scheme for the new energy vehicle includes: issuing the initial charging scheduling instruction to the new energy vehicle and receiving the charging willingness response data fed back by the new energy vehicle; The desired charging time period in the charging intention response data is compared with the suggested charging time period in the initial charging scheduling instruction to obtain the time period deviation result between the desired charging time period and the suggested charging time period; the desired charging power value in the charging intention response data is analyzed with the suggested charging power value in the initial charging scheduling instruction to obtain the power deviation result between the desired charging power value and the suggested charging power value; based on the time period deviation result and the power deviation result, the scheduling strategy is reconstructed for the suggested charging time period and suggested charging power value in the initial charging scheduling instruction to obtain the target charging optimization scheme for the new energy vehicle.

[0069] The system sends a pre-packaged initial charging scheduling command to the vehicle terminal of the new energy vehicle via wireless communication. After sending the command, the system monitors the feedback signal from the vehicle terminal in real time and receives charging intention response data submitted by the new energy vehicle user through the vehicle terminal interface. This data includes the user's self-set desired charging period, desired charging power value, and charging demand priority. The desired charging period is accurate to the minute and has a clear start and end time. The desired charging power value is set in units of 0.5kW. The charging demand priority is divided into three categories: urgent, regular, and non-urgent. This ensures that the received data is complete, accurate, and traceable.

[0070] The desired charging time period in the charging intention response data is compared with the suggested charging time period in the initial charging scheduling instruction. First, the start and end times of the two time periods are extracted, and the overlap time of the two time periods is calculated. The overlap time is divided by the total duration of the desired charging time period to obtain the matching degree. The matching degree is calculated as a percentage and kept to an integer. A matching degree of 100% is set as no deviation, 70% to 99% is a slight deviation, and below 70% is a serious deviation. The time period deviation result between the desired charging time period and the suggested charging time period is determined according to the calculated matching degree. At the same time, the specific start and end times of the deviation time period are recorded to ensure that the comparison process is reproducible and the results are clear.

[0071] A power difference analysis is performed between the expected charging power value in the charging intention response data and the suggested charging power value in the initial charging scheduling instruction. The suggested charging power value is set according to the load carrying capacity of the charging station: 7kW for Level 1, 5kW for Level 2, and 3kW for Level 3. First, the absolute difference between the expected charging power value and the suggested charging power value is calculated. A difference of 0kW is set as no deviation, 0.5kW to 1kW as a slight deviation, and greater than 1kW as a serious deviation. At the same time, a positive difference indicates that the expected power is higher than the suggested power, and a negative difference indicates that the expected power is lower than the suggested power. Based on this, the power deviation result between the expected charging power value and the suggested charging power value is obtained, and the magnitude and direction of the deviation are determined.

[0072] Based on the obtained time period deviation and power deviation results, the scheduling strategy is reconstructed for the suggested charging time period and suggested charging power value in the initial charging scheduling instruction. If both deviation results are without deviation, the initial instruction content remains unchanged. If the time period has a slight deviation and the power has no deviation, the suggested charging time period is adjusted so that the overlap between the adjusted time period and the expected charging time period accounts for more than 90% of the total expected time period, and the adjusted time period still belongs to the economic incentive period. If the power has a slight deviation and the time period has no deviation, the suggested charging power value is adjusted so that the absolute difference between the adjusted power and the expected power does not exceed 0.5kW, and does not exceed the maximum power limit corresponding to the current load carrying capacity of the charging station. If either deviation is a serious deviation, the suggested charging time period and suggested charging power value are re-determined by combining the spatiotemporal distribution characteristics in the charging demand forecast map, real-time grid load data, and user expectations, ensuring that the adjusted instruction not only meets the user's wishes but also complies with the operating constraints of the grid and charging station. After reconstruction, the target charging optimization scheme for new energy vehicles is obtained.

[0073] This step enables bidirectional interaction between dispatch instructions and actual user needs by issuing initial charging dispatch instructions to new energy vehicles and receiving user charging intention response data. By accurately comparing time period and power deviations, the dispatch strategy is reconstructed in a targeted manner. The resulting optimized charging solution can not only meet users' expected charging needs and improve their charging experience, but also take into account grid load regulation and charging station capacity constraints. This avoids grid load fluctuations or charging station overload problems caused by the disconnect between user charging needs and dispatch instructions. The entire process is clear, the judgment criteria are well-defined, and it is reproducible, effectively improving the flexibility and feasibility of charging dispatch and further enhancing the stability and efficiency of vehicle-grid collaborative operation.

[0074] A5. Monitor the execution results of the target charging optimization scheme. When the execution results are detected to deviate from the preset optimization target, update the multidimensional feature dataset according to the execution results, and iteratively optimize the charging demand prediction map and the electricity price system of the charging station.

[0075] In this embodiment of the invention, monitoring the execution result of the target charging optimization scheme, and updating the multidimensional feature dataset based on the execution result when the execution result deviates from the preset optimization target, and iteratively optimizing the charging demand prediction map and the electricity price system of the charging station, includes: acquiring actual charging execution data generated during the execution of the target charging optimization scheme, the actual charging execution data including actual charging time period, actual charging power value, and actual charging completion time; comparing and analyzing the actual charging execution data with the expected charging time period, expected charging power value, and expected charging completion time in the target charging optimization scheme to obtain the execution deviation data of the target charging optimization scheme; when the execution deviation data exceeds the tolerance threshold in the preset optimization target, adding the execution deviation data as a feedback feature to the multidimensional feature dataset to generate an updated multidimensional feature dataset; and iteratively correcting the charging demand prediction map and the electricity price system of the charging station based on the updated multidimensional feature dataset.

[0076] The actual charging execution data generated during the execution of the target charging optimization scheme is obtained. Various data during the charging process are collected in real time through the charging station's charging management terminal. At the same time, charging data uploaded by the new energy vehicle's on-board battery management system is received simultaneously. The actual charging period is the complete time from the start to the stop of power supply of the charging equipment, accurate to the minute and with clear start and end times. The actual charging power value is the real-time power value recorded every 5 minutes during the charging process. The average of all recorded values ​​is taken as the final actual charging power value. The actual charging completion time is the system time when the charging equipment stops supplying power, accurate to the minute. The above three data are integrated to form complete actual charging execution data, ensuring that the data collection is comprehensive, accurate and traceable.

[0077] The actual charging execution data is compared and analyzed one by one with the expected charging period, expected charging power value, and expected charging completion time in the target charging optimization scheme. First, the actual charging period is compared with the expected charging period, and the difference between the start and end times of the two is calculated to determine the duration of the period deviation. At the same time, the specific direction of the deviation is recorded. Then, the actual charging power value is compared with the expected charging power value, and the absolute difference between the two is calculated to determine the magnitude of the power deviation. Finally, the actual charging completion time is compared with the expected charging completion time, and the time difference between the two is calculated to determine the duration of the completion time deviation. The duration of the period deviation, the magnitude of the power deviation, and the duration of the completion time deviation are integrated and organized to form the execution deviation data of the target charging optimization scheme. Each deviation item is marked with a specific value and the direction of the deviation to ensure that the comparison results are clear and verifiable.

[0078] The preset optimization targets include tolerance thresholds: a time period deviation tolerance threshold of 15 minutes, a power deviation tolerance threshold of 1kW, and a completion time deviation tolerance threshold of 10 minutes. These tolerance thresholds are determined based on the grid load control accuracy requirements and charging station operation specifications to ensure that deviations within acceptable ranges do not affect vehicle-grid collaborative operation. Each deviation in the execution deviation data is compared with its corresponding tolerance threshold. If any deviation exceeds the corresponding tolerance threshold, it is considered to have exceeded the tolerance threshold. In this case, the execution deviation data is used as a feedback feature. Following the original format of the multidimensional feature dataset, it is associated with the geographical grid unit and time window to which the target new energy vehicle belongs and added to the multidimensional feature dataset. This updated multidimensional feature dataset is then integrated and arranged with the original feature data to generate an updated multidimensional feature dataset. This updated dataset ensures that it includes historical features and the latest feedback features, and that the data structure remains consistent with the original dataset.

[0079] Based on the updated multidimensional feature dataset, the charging demand prediction map and the charging station electricity pricing system are collaboratively and iteratively corrected. When correcting the charging demand prediction map, feedback features from the updated multidimensional feature dataset and the original spatiotemporal coupling features are extracted. The charging probability of each geographical grid unit within the target area is recalculated, and the charging probability range corresponding to the gradient color markers in the map is adjusted to better align the color markers with the charging probabilities in the actual charging execution data. Simultaneously, the load capacity feature values ​​labeled in each grid unit are updated to ensure the map accurately reflects actual charging demand and grid load status. When correcting the charging station electricity pricing system, the execution deviation data in the updated multidimensional feature dataset is used to adjust the basic pricing range corresponding to each load carrying capacity level. If the actual charging power deviation of a certain level of charging station is generally large, the span of the basic pricing range for that level is appropriately reduced. At the same time, the electricity price fluctuation adjustment factor corresponding to the peak and valley periods of the grid is adjusted. If the actual charging volume during the valley period is lower than expected, the valley period adjustment factor is appropriately reduced, making the electricity pricing system more closely reflect actual charging execution and grid load changes. This ensures that the charging demand prediction map and the electricity pricing system are collaboratively matched and meet the requirements of vehicle-grid collaborative operation.

[0080] Meanwhile, the iterative correction also includes online updates to the time-series prediction model: the updated multidimensional feature dataset is used as new training samples to incrementally learn or fine-tune the parameters of the original model, so that it continuously fits the latest vehicle-to-grid interaction patterns, thereby continuously improving the accuracy of charging demand prediction and the adaptability of electricity price dynamic adjustment.

[0081] The beneficial effects are as follows: by acquiring actual charging execution data during the execution of the target charging optimization scheme, the actual and expected data are accurately compared to obtain the execution deviation. When the deviation exceeds the tolerance threshold, the multidimensional feature dataset is updated. Based on the updated dataset, the charging demand prediction map and the charging station electricity price system are collaboratively and iteratively corrected. This can continuously optimize the prediction accuracy and the rationality of the electricity price, make up for the deviation between the initial scheme and the actual execution, make the charging demand prediction more in line with reality, and make the electricity price system more suitable for the vehicle-grid operation status. This further improves the accuracy and feasibility of charging scheduling, ensures the stability and efficiency of vehicle-grid collaborative operation, and forms a closed-loop mechanism of "execution-feedback-update-correction" to ensure that the entire technical solution can be continuously optimized and adapted to the vehicle-grid operation needs under different scenarios.

[0082] Example 2

[0083] like Figure 2 As shown in the figure, this embodiment also provides a functional module diagram of a new energy charging optimization system based on multi-source data.

[0084] The new energy charging optimization system 100 based on multi-source data described in this embodiment can be installed in a terminal. Depending on the functions implemented, the new energy charging optimization system 100 based on multi-source data may include a cross-source mapping module 101, an angular spatiotemporal extrapolation module 102, a dynamic pricing module 103, a closed-loop correction module 104, and an iterative optimization module 105. The modules described in this invention can also be called units, referring to a series of computer program segments that can be executed by the terminal processor and perform fixed functions, stored in the terminal's memory.

[0085] In this embodiment, the functions of each module / unit are as follows: The cross-source mapping module 101 is used to perform cross-source data mapping between the real-time location information and historical driving trajectory data of new energy vehicles and the real-time load data of the power grid, to obtain a multi-dimensional feature dataset between the new energy vehicles and the power grid; the spatiotemporal projection module 102 is used to perform spatiotemporal distribution projection of the charging demand distribution of the target area in the future time period based on the multi-dimensional feature dataset, and to construct a charging demand prediction map of the target area; the dynamic pricing module 103 is used to dynamically adjust the electricity price system of charging stations in the target area based on the charging demand prediction map and the real-time load data. The system uses a state-based pricing mechanism to generate an initial charging scheduling instruction for the new energy vehicle. A closed-loop correction module 104 sends the initial charging scheduling instruction to the new energy vehicle and, based on the charging willingness response data fed back by the new energy vehicle, performs a closed-loop correction of the scheduling strategy for the initial charging scheduling instruction to obtain a target charging optimization scheme for the new energy vehicle. An iterative optimization module 105 monitors the execution result of the target charging optimization scheme. When the execution result deviates from the preset optimization target, it updates the multi-dimensional feature dataset based on the execution result and iteratively optimizes the charging demand prediction map and the electricity price system of the charging station.

[0086] In detail, each module in the new energy charging optimization system 100 based on multi-source data described in the embodiments of the present invention adopts the same technical means as the new energy charging optimization method based on multi-source data described in Embodiment 1 and Embodiment 2, and can produce the same technical effect, which will not be repeated here.

[0087] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0088] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A new energy charging optimization method based on multi-source data, characterized in that, The method includes: A1. Cross-source data mapping is performed between the real-time location information and historical driving trajectory data of new energy vehicles and the real-time load data of the power grid to obtain a multi-dimensional feature dataset between the new energy vehicles and the power grid. A2. Based on the multidimensional feature dataset, perform spatiotemporal distribution extrapolation of the charging demand distribution in the target area in future time periods, and construct a charging demand prediction map of the target area. A3. Based on the charging demand forecast map and the real-time load data, dynamically guide the pricing of the electricity price system of the charging stations in the target area to generate the initial charging dispatch instructions for the new energy vehicles. A4. Send the initial charging scheduling instruction to the new energy vehicle, and based on the charging intention response data fed back by the new energy vehicle, perform closed-loop correction of the scheduling strategy on the initial charging scheduling instruction to obtain the target charging optimization scheme for the new energy vehicle. A5. Monitor the execution results of the target charging optimization scheme. When the execution results are detected to deviate from the preset optimization target, update the multidimensional feature dataset according to the execution results, and iteratively optimize the charging demand prediction map and the electricity price system of the charging station.

2. The new energy charging optimization method based on multi-source data as described in claim 1, characterized in that, The method involves cross-source data mapping of the real-time location information and historical driving trajectory data of new energy vehicles with the real-time load data of the power grid to obtain a multi-dimensional feature dataset between the new energy vehicles and the power grid, including: Geographic grid codes are extracted from the real-time location information of new energy vehicles, and the node load status corresponding to the geographic grid codes in the real-time load data of the power grid is obtained to establish a spatial association mapping relationship between the new energy vehicles and the power grid. Based on the spatial correlation mapping relationship, the travel pattern features in the historical driving trajectory data of the new energy vehicles and the time-period load fluctuation features in the real-time load data are cross-coded to obtain the spatiotemporal coupling feature vector between the new energy vehicles and the power grid. The spatiotemporal coupling feature vector is normalized, and combined with the remaining battery power of the new energy vehicle and the load capacity of the power grid node, a multidimensional feature dataset between the new energy vehicle and the power grid is constructed.

3. The new energy charging optimization method based on multi-source data as described in claim 2, characterized in that, The step of extrapolating the spatiotemporal distribution of charging demand in the target area over future periods based on the multidimensional feature dataset, and constructing a charging demand prediction map for the target area, includes: Based on the spatiotemporal coupling feature vector, the evolution trend of vehicle aggregation and the load change trend of power grid nodes in the target area are extracted to obtain the basic driving factors of the target area. Based on the fundamental driving factors, the future charging probability distribution of each geographic grid unit in the target area is predicted to obtain the potential charging demand density distribution map of the target area. The potential charging demand density distribution map is overlaid with the load capacity characteristics to construct a charging demand prediction map for the target area.

4. The new energy charging optimization method based on multi-source data as described in claim 3, characterized in that, The formula for calculating the basic driving factor is as follows: ； In the formula, Geographic grid unit In the time window The fundamental driving factors For the geographic grid unit In the time window The evolution trend of vehicle aggregation. For the geographic grid unit and its adjacent grid cell set In the time window The sum of the trends in vehicle aggregation evolution, The preset load fluctuation impact factor, For the geographic grid unit In the time window The trend of load change For the geographic grid unit The load capacity characteristics of the corresponding power grid nodes.

5. The new energy charging optimization method based on multi-source data as described in claim 1, characterized in that, The step of dynamically guiding the pricing system of charging stations within the target area based on the charging demand forecast map and the real-time load data to generate initial charging dispatch instructions for the new energy vehicles includes: Based on the spatiotemporal distribution characteristics in the charging demand forecast map, the load carrying capacity of charging stations in the target area is assessed in a graded manner to obtain the basic pricing range of the charging stations. Map the grid peak and valley time information in the real-time load data to the electricity price floating adjustment factor of the charging station; The basic pricing range is weighted and superimposed with the electricity price fluctuation adjustment factor to obtain the dynamic time-of-use electricity price system of the charging station; Based on the dynamic time-of-use electricity pricing system, economic incentives are used to guide the charging behavior of the new energy vehicles in order to obtain the initial charging scheduling instructions for the new energy vehicles.

6. The new energy charging optimization method based on multi-source data as described in claim 5, characterized in that, The step of providing economic incentives to guide the charging behavior of the new energy vehicles based on the dynamic time-of-use electricity pricing system to obtain the initial charging dispatch instructions for the new energy vehicles includes: Obtain the electricity price period division nodes and the charging unit price value corresponding to each electricity price period in the dynamic time-of-use electricity price system; Extract the remaining battery charge and estimated parking time from the vehicle status data of the new energy vehicle; The charging unit price is matched and analyzed with the expected parking duration to determine the target electricity price time period set covered by the new energy vehicle within the expected parking duration; Based on the remaining battery capacity and the charging unit price for each electricity price period in the target electricity price period set, a quantitative value of the charging cost difference of the new energy vehicle in each electricity price period is generated. The quantified value of the charging cost difference is compared with a preset incentive threshold, and the electricity price period in which the quantified value of the concentrated charging cost difference is lower than the incentive threshold is selected as the economic incentive period. The economic incentive period and the corresponding charging unit price are encapsulated into the initial charging scheduling instruction for the new energy vehicle.

7. The new energy charging optimization method based on multi-source data as described in claim 1, characterized in that, The step of sending the initial charging scheduling command to the new energy vehicle and, based on the charging intention response data fed back by the new energy vehicle, performing a closed-loop correction of the scheduling strategy for the initial charging scheduling command to obtain the target charging optimization scheme for the new energy vehicle includes: The system issues the initial charging scheduling command to the new energy vehicle and receives the charging intention response data fed back by the new energy vehicle. The desired charging time period in the charging intention response data is compared with the suggested charging time period in the initial charging scheduling instruction to obtain the time period deviation result between the desired charging time period and the suggested charging time period. A power difference analysis is performed between the expected charging power value in the charging intention response data and the suggested charging power value in the initial charging scheduling instruction to obtain the power deviation result between the expected charging power value and the suggested charging power value. Based on the time period deviation results and the power deviation results, the suggested charging time period and suggested charging power value in the initial charging scheduling instruction are reconstructed to obtain the target charging optimization scheme for the new energy vehicle.

8. The new energy charging optimization method based on multi-source data as described in claim 1, characterized in that, The monitoring of the execution results of the target charging optimization scheme, when the execution result deviates from the preset optimization target, updates the multidimensional feature dataset based on the execution result, and iteratively optimizes the charging demand prediction map and the electricity price system of the charging station, including: Obtain the actual charging execution data generated during the execution of the target charging optimization scheme. The actual charging execution data includes the actual charging time period, the actual charging power value, and the actual charging completion time. The actual charging execution data is compared and analyzed with the expected charging period, expected charging power value and expected charging completion time in the target charging optimization scheme to obtain the execution deviation data of the target charging optimization scheme. When the execution deviation data exceeds the tolerance threshold in the preset optimization target, the execution deviation data is added to the multidimensional feature dataset as a feedback feature to generate the updated multidimensional feature dataset. Based on the updated multidimensional feature dataset, the charging demand prediction map and the electricity price system of the charging station are collaboratively and iteratively corrected.

9. A new energy charging optimization system based on multi-source data, characterized in that, The system for implementing the new energy charging optimization method based on multi-source data as described in claim 1 includes: The cross-source mapping module is used to perform cross-source data mapping between the real-time location information and historical driving trajectory data of new energy vehicles and the real-time load data of the power grid, so as to obtain a multi-dimensional feature dataset between the new energy vehicles and the power grid. The spatiotemporal extrapolation module is used to extrapolate the spatiotemporal distribution of charging demand in the target area in future time periods based on the multidimensional feature dataset, and to construct a charging demand prediction map of the target area. The dynamic pricing module is used to dynamically guide the pricing of the electricity price system of charging stations in the target area based on the charging demand forecast map and the real-time load data, so as to generate the initial charging dispatch instructions for the new energy vehicles. The closed-loop correction module is used to send the initial charging scheduling instruction to the new energy vehicle, and based on the charging intention response data fed back by the new energy vehicle, to perform closed-loop correction of the scheduling strategy of the initial charging scheduling instruction to obtain the target charging optimization scheme of the new energy vehicle. The iterative optimization module is used to monitor the execution result of the target charging optimization scheme. When the execution result is detected to deviate from the preset optimization target, the multidimensional feature dataset is updated according to the execution result, and the charging demand prediction map and the electricity price system of the charging station are iteratively optimized.

10. A storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the method according to any one of claims 1 to 8.

Images