Mobile charging pile charging demand prediction method and system based on big data analysis

Abstract

The application discloses a mobile charging pile charging demand prediction method and system based on big data analysis, relates to the related technical field of charging piles, and comprises the following steps: acquiring the position distribution and the residual capacity distribution of all mobile charging piles in a preset area in real time; performing first-type charging demand prediction from a wave crest to a wave trough, and performing charging task distribution density partition identification; performing second-type charging demand prediction of the mobile charging pile based on a charging pile distribution map; performing second-type charging scheme decision, and generating a second charging decision; and controlling the charging of the mobile charging pile by the second charging decision. The application solves the technical problems that the prior art cannot match the dynamic trajectory of a user, cannot adapt to dynamically changing charging demand, and cannot reasonably allocate charging resources, thereby causing inaccurate charging demand prediction, poor charging service efficiency and poor user experience. The application achieves the technical effects of optimizing charging resource allocation, improving charging demand prediction accuracy, and improving charging service efficiency and user experience.

Description

Technical Field

[0001] This application relates to the technical field of charging piles, specifically to a method and system for predicting the charging demand of mobile charging piles based on big data analysis. Background Technology

[0002] As the application scenarios of mobile charging piles continue to expand, problems such as uneven distribution of charging resources and prominent supply-demand contradictions are gradually being exposed. On the one hand, the traditional static layout mode of charging piles is difficult to adapt to the dynamic changes in the travel trajectories of new energy vehicle users, resulting in an idle rate of up to 40% for charging piles in some areas, while popular business districts and transportation hubs have long faced the dilemma of not being able to find a charging pile. On the other hand, existing charging demand forecasts rely heavily on statistical analysis of historical data, which cannot accurately capture fluctuations in charging demand under special scenarios such as holidays and extreme weather, making the scheduling of charging resources lack timeliness and scientific rigor. In addition, new energy vehicle users are constantly increasing their demands for the immediacy and convenience of charging services, and the existing charging service model is unable to meet market demand, making it impossible to achieve dynamic optimization of mobile charging pile resources. Consequently, it is difficult to accurately predict charging demand and rationally plan the scheduling of charging resources, affecting the efficiency of charging services and the control of operating costs.

[0003] Therefore, current technologies suffer from several technical problems, including inability to match users' dynamic trajectories, difficulty in adapting to dynamically changing charging demands, and unreasonable allocation of charging resources, leading to inaccurate charging demand prediction, poor charging service efficiency, and a poor user experience. Summary of the Invention

[0004] This application provides a method and system for predicting charging demand for mobile charging piles based on big data analysis. It solves the technical problems in the prior art, such as the inability to match the dynamic trajectory of users, the difficulty in adapting to dynamically changing charging demand, and the unreasonable allocation of charging resources, which lead to inaccurate charging demand prediction, poor charging service efficiency, and poor user experience. It achieves the technical effects of optimizing charging resource allocation, improving the accuracy of charging demand prediction, and enhancing charging service efficiency and user experience.

[0005] This application provides a method for predicting the charging demand of mobile charging piles based on big data analysis. The method includes: acquiring the location distribution and remaining power distribution of all mobile charging piles within a preset area in real time, and constructing a charging pile distribution map; starting from the current time, performing a first type of charging demand prediction from peak to trough within the preset area through big data analysis, and identifying the charging task distribution density partition to generate a first charging demand time series; using the first charging demand time series as the task objective, performing a second type of charging demand prediction of mobile charging piles based on the charging pile distribution map to generate second charging demand information; performing a second type of charging scheme decision based on the second charging demand information to generate a second charging decision; and controlling the charging of the mobile charging piles with the second charging decision.

[0006] In a possible implementation, the mobile charging pile charging demand prediction method based on big data analysis further performs the following processing: the first type of charging demand is the charging demand of the mobile charging pile for charging electric vehicle users in the preset area; the second type of charging demand is the charging demand of the mobile charging pile itself.

[0007] In a possible implementation, the mobile charging pile charging demand prediction method based on big data analysis further performs the following processing: collecting historical charging datasets of mobile charging piles within the preset area, including arbitrary historical charging data including charging location and charging amount at any historical time; performing charging demand analysis at each time of day based on the historical charging dataset to generate a set of historical daily charging volume curves; performing partitioning based on charging task distribution density based on charging location within the historical charging dataset to generate a set of historical density partitioning identifiers; using the set of historical daily charging volume curves to perform peak-to-valley charging volume prediction starting from the current time to generate the first charging demand time series; and performing density partitioning prediction for peak-to-valley charging volume using the set of historical density partitioning identifiers to perform charging task distribution density partitioning.

[0008] In a possible implementation, the mobile charging pile charging demand prediction method based on big data analysis also performs the following processing: collecting charging task cancellation records at any historical moment; collecting historical average charging data records for the requesting users corresponding to the charging task cancellation records; and performing corresponding moment charging volume correction on the historical daily charging volume curve set using the historical average charging data records.

[0009] In a possible implementation, the mobile charging pile charging demand prediction method based on big data analysis further performs the following processing: marking each curve in the historical daily charging volume curve set with time factor features, including marking weekdays, holidays, weather, and seasons, to generate time factor identifiers; clustering the curves according to the time factor identifiers to generate curve sets corresponding to each time factor identifier; using the current moment as the prediction starting point, predicting the peak to trough period of charging volume based on time-based and year-on-year comparisons of each curve set, to generate the first charging demand time series.

[0010] In a possible implementation, the mobile charging demand prediction method based on big data analysis further performs the following processing: Based on the charging task distribution density partitioning identifier of the first charging demand time series, and combined with the location distribution in the charging pile distribution map, a density adaptation calculation is performed to generate a density adaptation index; if the density adaptation index is greater than or equal to a preset adaptation threshold, based on the charging pile distribution map, with the first charging demand time series as the task target, the optimal fitting of charging pile task allocation is performed to determine the identified charging piles and identified interruption tasks that cause charging task interruption; the minimum charging demand constraint and charging time constraint of the identified charging pile are calculated based on the identified interruption task to generate the second charging demand information.

[0011] In a possible implementation, the mobile charging pile charging demand prediction method based on big data analysis further performs the following processing: determining the charging task at each moment according to the first charging demand time sequence, predicting the charging duration by combining the charging pile charging voltage information, and generating a first charging duration time sequence; combining the charging pile distribution map, the first charging demand time sequence, and the first charging duration time sequence, performing charging pile and task matching under distance priority and time matching, and determining unmatched tasks and matched results; for the unmatched tasks, performing task reallocation under the charging pile's own power decision under distance priority in the matched results, and generating the identified charging pile and the identified interrupted task.

[0012] In a possible implementation, the mobile charging pile charging demand prediction method based on big data analysis further performs the following processing: obtaining the distribution information of mobile charging pile operation centers; combining the second charging demand information and the operation center distribution information to perform charging time matching, and generating the second charging decision.

[0013] In a possible implementation, the mobile charging demand prediction method based on big data analysis also performs the following processing: determining whether there are backup mobile charging piles in each mobile charging pile operation center; if so, using the backup mobile charging pile to execute the charging decision for the unmatched task, and combining the matched result to execute the charging pile's own charging decision according to the preset off-peak charging pile allocation strategy after the charging task is completed.

[0014] This application also provides a mobile charging pile charging demand prediction system based on big data analysis. The system includes: a charging pile distribution map construction module, used to acquire the location distribution and remaining power distribution of all mobile charging piles within a preset area in real time, and construct a charging pile distribution map; a first charging demand generation module, used to perform a first type of charging demand prediction from peak to trough within the preset area based on big data analysis, starting from the current time, and to perform charging task distribution density partitioning and identification, generating a first charging demand time series; a second charging demand generation module, used to perform a second type of charging demand prediction of mobile charging piles based on the charging pile distribution map, with the first charging demand time series as the task target, and to generate second charging demand information; a second charging decision generation module, used to perform a second type of charging scheme decision based on the second charging demand information, and generate a second charging decision; and a mobile charging pile control module, used to control the charging of the mobile charging piles based on the second charging decision.

[0015] This application proposes a mobile charging demand prediction method and system based on big data analysis. The method acquires the real-time location distribution and remaining battery power distribution of all mobile charging piles within a preset area; performs a first-type charging demand prediction from peak to trough, and identifies the charging task distribution density; performs a second-type charging demand prediction based on a charging pile distribution map; performs a second-type charging scheme decision, generating a second charging decision; and controls the charging of the mobile charging piles using the second charging decision. This solves the technical problems in existing technologies, such as the inability to match user dynamic trajectories, difficulty in adapting to dynamically changing charging demands, and unreasonable allocation of charging resources, leading to inaccurate charging demand prediction, poor charging service efficiency, and poor user experience. It achieves the technical effects of optimizing charging resource allocation, improving the accuracy of charging demand prediction, and enhancing charging service efficiency and user experience. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a schematic diagram of the mobile charging pile charging demand prediction method based on big data analysis provided in the embodiments of this application.

[0018] Figure 2 This is a schematic diagram of the structure of a mobile charging pile charging demand prediction system based on big data analysis provided in an embodiment of this application.

[0019] Explanation of reference numerals in the attached diagram: 10 for charging pile distribution map construction module, 20 for first charging demand generation module, 30 for second charging demand generation module, 40 for second charging decision generation module, and 50 for mobile charging pile control module. Detailed Implementation

[0020] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below.

[0021] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0023] This application provides a method for predicting the charging demand of mobile charging stations based on big data analysis, such as... Figure 1 As shown, the method includes: Step S100: Obtain the location distribution and remaining power distribution of all mobile charging piles within the preset area in real time, and construct a charging pile distribution map.

[0024] Preferably, the location distribution and remaining power distribution of all mobile charging piles within a preset area are collected in real time through IoT sensors, positioning technology, or the communication module built into the charging pile, and visualized as a charging pile distribution map. The preset area may include a certain administrative district, business district, highway service area, etc. Specifically, the GPS positioning module deployed on the mobile charging pile, combined with IoT communication technology (such as 4G / 5G, NB-IoT), transmits latitude and longitude coordinates to the cloud server at a frequency of minutes (such as once every 5 minutes) and records its current location status, such as whether it is fixed at a certain point or whether it is in transit. The charging pile BMS (Battery Management System) monitors the real-time power information of each charging pile, such as the remaining battery capacity, battery pack SOC (State of Charge), supported charging time or number of charging times, etc., and uploads the data to the main control unit through the internal bus, and transmits it synchronously with the location information. The location and remaining battery power of all collected charging stations are mapped using a geographic information system (GIS). For example, different colored icons represent the location of charging stations (e.g., blue icons represent fixed stations, green icons represent mobile stations), and progress bars, numerical labels, or color intensities (e.g., red indicates low battery, green indicates high battery) visually display the remaining battery power of each charging station. Furthermore, the charging station distribution map is updated in real time (e.g., once per second or minute) to ensure users can always check the latest status of charging stations and avoid resource misjudgments due to information lag. When users search for charging stations through a mobile app, the charging station distribution map can prioritize recommending the nearest charging station with sufficient remaining battery power based on the user's current location, reducing waiting time and improving charging efficiency.

[0025] Step S200: Starting from the current moment, perform peak-to-trough first-type charging demand prediction within the preset area through big data analysis, and perform charging task distribution density partitioning identification to generate the first charging demand time sequence.

[0026] Step S200 further includes step S210, collecting historical charging datasets of mobile charging piles within the preset area, including arbitrary historical charging data including charging location and charging amount at any historical time; step S220, performing charging demand analysis at each time of day based on the historical charging dataset to generate a set of historical daily charging volume curves; step S230, performing partition identification based on charging task distribution density based on the charging location within the historical charging dataset to generate a set of historical density partition identifiers; step S240, using the set of historical daily charging volume curves to perform charging volume peak to trough time prediction starting from the current time to generate the first charging demand time series; step S250, performing density partition prediction for peak to trough time using the set of historical density partition identifiers to perform charging task distribution density partition identification.

[0027] Preferably, starting from the current moment, the system uses big data analysis to predict the first type of charging demand from peak to trough within a preset area. Specifically, it collects historical charging datasets of all mobile charging stations within the preset area, i.e., charging records over a past period (e.g., the past 30 days or six months). Each record includes the charging location (the specific geographical coordinates of the charging station when performing the charging task, such as a shopping mall parking lot or a highway service area) and the charging amount (the charging time or electricity consumed in that charging task, such as how many kilowatt-hours the user charged the vehicle). Then, based on the historical charging dataset, it performs charging demand analysis for each moment of the day. The analysis involves grouping historical charging data by date and time (e.g., by hour or 15-minute intervals), and calculating the total charging volume in a preset area within each time interval. For example, it calculates the total charging volume from 8:00 AM to 9:00 AM and the total charging volume from 5:00 PM to 6:00 PM each day over the past 30 days. It also generates a curve showing the change in charging volume over time for each day, i.e., generating multiple historical daily charging volume curves. The horizontal axis represents time, and the vertical axis represents charging volume. These curves are then combined to form a set of historical daily charging volume curves, which visually demonstrate the fluctuation patterns of charging volume on different days. For example, charging volume is low during the morning peak and high during the evening peak on weekdays, while it may fluctuate less throughout the day on weekends.

[0028] Preferably, the preset area is divided into several small geographical units (e.g., divided by grid, with each grid having a side length of 1 kilometer), or divided by actual functional areas (e.g., commercial districts, residential areas, industrial areas). The number of charging tasks or the total amount of charging for each geographical unit in historical charging data are counted to determine the charging density of the area. Then, the area is zoned and labeled according to the density, which may include high-density areas with frequent charging tasks, such as downtown commercial districts with more than 50 charging times per day; medium-density areas with moderate charging tasks, such as ordinary residential areas with 20-50 charging times per day; and low-density areas with fewer charging tasks, such as suburbs with less than 20 charging times per day. This generates a set of historical density zoning labels and identifies the density level of each area.

[0029] Preferably, a set of historical daily charging volume curves is used to predict the peak and trough periods of charging volume starting from the current time. Specifically, based on the set of historical daily charging volume curves, big data algorithms (such as time series analysis and machine learning models) are used to analyze the changing patterns of charging volume within a day, identify the peak charging volume periods (such as evening rush hour) and trough periods (such as early morning) that may occur on the current date (such as today being a weekday or weekend), and arrange them in chronological order to form the first charging demand time series. For example, if the current day is a weekday, the model predicts that today's charging peak is 17:00-20:00 and the trough is 0:00-6:00; if the current day is a weekend, the peak may be delayed to 10:00-12:00 and 15:00-17:00, in order to determine which time periods have high charging demand and which time periods have low demand.

[0030] Preferably, the final step is to perform density zoning prediction for peak and trough periods using a set of historical density zoning identifiers. Specifically, within defined peak and trough periods, the historical density zoning identifier set is used to analyze the changing trends in charging density in different areas during different time periods. For example, historical data shows that high-density areas during weekday evening peak hours (peak periods) are concentrated around office buildings and near subway stations; high-density areas in the early morning (trough periods) may shift to residential areas (users charging at night). This allows for prediction of the regional density level for each time period and identification of charging task distribution density zones. For instance, predicting that from 18:00 to 20:00 (peak period), area A (with concentrated office buildings) is a high-density area, and residential area B is a medium-density area; or predicting that from 0:00 to 5:00 (trough period), area A becomes a low-density area, and residential area B becomes a high-density area (users charging at home). This ensures that mobile charging stations are in the right locations at the right time, avoiding resource waste or insufficient supply.

[0031] Furthermore, step S220 also includes step S221, collecting charging task cancellation records at any historical time; step S222, collecting historical average charging data records for the requesting user corresponding to the charging task cancellation record; and step S223, using the historical average charging data records to perform corresponding time-based charging volume correction on the historical daily charging volume curve set.

[0032] Preferably, charging task cancellation records are collected at any historical moment. Each record includes the cancellation time, the specific time the user canceled the charging request (e.g., 2025-05-20, 18:30:00); the original planned charging location (the location of the charging station the user originally intended to use); and the reason for cancellation (e.g., "charging station malfunction," "excessive waiting time," "temporary change of itinerary"). Then, for each user who canceled a charging task, their historical successfully completed charging records are retrieved, and historical average charging data is calculated. This may include average charging amount (the average amount of kWh charged per charge for the user, e.g., user A averages 30 kWh per charge); and average charging time. The system calculates the user's average charging time per charge, e.g., user B's average charging time is 1 hour per charge; charging time preferences, e.g., the user's preferred charging time periods, e.g., user C often charges between 19:00 and 21:00 on weekdays; finally, it corrects the charging amount at the corresponding moment using the historical average charging data record and the historical daily charging amount curve set. Specifically, on the historical daily charging amount curve, the time point corresponding to the canceled record is found, and the user's historical average charging amount (e.g., 30 degrees) is superimposed on the original charging amount at the corresponding moment to generate the corrected curve, thus obtaining the corrected historical daily charging amount curve, which is closer to the actual needs.

[0033] Furthermore, step S240 also includes step S241, marking each curve in the historical daily charging volume curve set with time factor features, including marking weekdays, holidays, weather, and seasons, to generate time factor identifiers; step S242, clustering the curves according to the time factor identifiers to generate curve sets corresponding to each time factor identifier; step S243, using the current moment as the prediction starting point, predicting the peak to trough period of charging volume based on time-based and year-on-year comparisons based on the curve sets, to generate the first charging demand time series.

[0034] Preferably, each curve in the historical daily charging volume curve set is labeled with time factor features. This involves adding multiple time-dimensional labels to each historical daily charging volume curve, including weekdays, holidays, weather, and seasons, generating time factor identifiers. Different time factors significantly affect charging demand. For example, users are more likely to use electric vehicles on rainy days, potentially increasing charging demand; low temperatures in winter reduce battery range, leading to increased charging frequency; and increased travel during holidays causes a surge in demand for charging stations around scenic spots. Then, all curves are clustered according to their time factor labels. For example, clustering algorithms such as K-means are used, with the time factor labels as feature vectors for clustering. This ensures that curves within the same group have similar charging volume fluctuation patterns, thus forming multiple sets of similar scenario curves corresponding to different time factor identifiers. Each set corresponds to a specific time scenario; for example, the weekday / sunny / summer group includes all curves labeled [weekday, sunny, summer], and the holiday / rainy / winter group includes all curves labeled [holiday, rainy, winter].

[0035] Preferably, the current moment is used as the starting point for prediction. Based on various curve sets, the peak and trough periods of charging volume are predicted based on month-on-month and year-on-year comparisons. Specifically, month-on-month analysis refers to comparing the charging volume trend of the current day with that of the previous day to identify short-term fluctuation patterns. If the charging volume increased at the same time period the previous day, it is predicted that the trend may continue today. Year-on-year analysis refers to selecting curves from a set that are close to the current date in history and analyzing the peak and trough periods of these curves. Then, by combining the month-on-month and year-on-year results, the peak and trough periods of today are predicted and arranged in chronological order to form a prediction time series, and finally the first charging demand time series is obtained.

[0036] Step S300: Taking the first charging demand time sequence as the task objective, perform the second type of charging demand prediction for mobile charging piles based on the charging pile distribution map to generate second charging demand information.

[0037] Preferably, step S300 further includes step S310, performing charging pile density adaptation calculation based on the charging task distribution density partition identifier of the first charging demand time sequence and the location distribution in the charging pile distribution map, and generating a density adaptation index; step S320, if the density adaptation index is greater than or equal to a preset adaptation threshold, performing optimal fitting of charging pile task allocation based on the charging pile distribution map with the first charging demand time sequence as the task target, and determining the identified charging pile and the identified interruption task that caused the charging task to be interrupted; step S330, calculating the minimum charging demand constraint and charging time constraint of the identified charging pile with the identified interruption task, and generating the second charging demand information.

[0038] Preferably, based on the charging task distribution density zoning identifier of the first charging demand time series, and combined with the location distribution in the charging pile distribution map, the density adaptation calculation of charging piles is performed. That is, the high-density areas (such as business districts and office buildings) in the first charging demand time series are spatially matched with the charging pile locations in the charging pile distribution map. For example, if the demand forecast shows that a certain business district (Area A) is a high-density charging area from 18:00 to 20:00, while the charging pile map shows that there are currently only 5 charging piles in this area, the ratio of the predicted demand density to the existing charging pile density is calculated as the density adaptation index, which represents the number of vehicles that each charging pile should serve. For example, if the predicted demand density of Area A is 100 vehicles / square kilometer, that is, 100 vehicles need to be served per square kilometer; and the existing charging pile density of Area A is 20 units / square kilometer, that is, each charging pile needs to serve 5 vehicles; then the adaptation index is 5. If the density adaptation index is greater than or equal to a preset threshold, it means that the number of existing charging piles is insufficient to meet the demand, which may lead to the interruption of charging tasks.

[0039] Preferably, if the density matching index is greater than or equal to the preset matching threshold, based on the charging pile distribution map, the optimal fitting of charging pile task allocation is performed with the first charging demand time sequence as the task objective. Specifically, in the charging pile distribution map, the task in the first charging demand time sequence is simulated to be allocated to the nearest charging pile. For example, 100 charging demands in area A from 18:00 to 20:00 are allocated to 5 surrounding charging piles, each of which needs to serve 20 vehicles. Then, according to the constraints (time constraints and power constraints), the detection is performed. If a charging pile needs to serve within 2 hours, the optimal matching index is determined. With 20 vehicles and an average charging time of 0.5 hours per vehicle, the charging station can serve a maximum of 4 vehicles, and the remaining 16 tasks will be interrupted. If a charging station has only 100 kWh of remaining power, while the total demand for the assigned tasks is 200 kWh, the excess will be interrupted. The charging station that caused the charging task interruption will be identified, such as charging stations ID-001 and ID-002 in area A, which may be interrupted due to excessive load. The interrupted tasks will also be identified, such as the 11th to 20th assigned tasks of ID-005 during the 19:00-20:00 time period.

[0040] Preferably, the minimum charging demand constraint (minimum power requirement) and charging time constraint (latest start time and earliest end time) of the identified charging pile are calculated using the interrupted task. Among the multiple interrupted charging tasks, the minimum amount of electricity required for each vehicle is multiplied by the number of interrupted charging tasks to obtain the total demand constraint, i.e., the minimum power requirement. The minimum charging demand constraint and charging time constraint are then transformed into optimization objectives to generate second charging demand information, thereby providing precise objectives for mobile charging pile scheduling. This includes determining the number of backup charging piles based on the minimum demand constraint, determining the optimal deployment time of charging piles based on the time constraint, and determining high-risk areas for priority reinforcement based on density zoning identifiers, thereby improving user experience and operational efficiency.

[0041] Furthermore, step S300 also includes the first type of charging demand being the charging demand of the mobile charging pile for charging electric vehicle users within the preset area; and the second type of charging demand being the charging demand of the mobile charging pile itself.

[0042] Preferably, the first type of charging demand refers to the demand for mobile charging piles to provide charging services to electric vehicle users within a preset area, with the core being to meet the charging needs of the users; the second type of charging demand refers to the demand for mobile charging piles, as electrical equipment, to return to the operation center or fixed charging point to replenish their power, with the core being to ensure the sustainable service capability of the charging piles.

[0043] Furthermore, step S320 also includes step S321, determining the charging task at each moment based on the first charging demand time sequence, and predicting the charging duration by combining the charging pile charging voltage information to generate a first charging duration time sequence; step S322, combining the charging pile distribution map, the first charging demand time sequence, and the first charging duration time sequence, performing charging pile and task matching under distance priority and time matching to determine unmatched tasks and matched results; step S323, for the unmatched tasks, performing task reallocation under the charging pile's own power decision under distance priority in the matched results to generate the identified charging pile and the identified interrupted task.

[0044] Preferably, the charging task at each moment is determined from the first charging demand time sequence, and the charging time is predicted by combining the charging pile's charging voltage information. Specifically, based on each charging task in the first charging demand time sequence (e.g., user A needs to charge 70 kWh between 18:00 and 19:00), and combined with the charging pile's charging voltage information (e.g., 400V for fast charging piles and 220V for slow charging piles), the charging time is calculated using a formula. For example, if user A uses a 400V fast charging station (efficiency 90%), the charging time = 70 ÷ (400 × 90%) ≈ 0.2 hours (about 12 minutes). Then, the predicted charging times of all tasks are arranged in chronological order to form the first charging time sequence.

[0045] Preferably, by combining the charging pile distribution map, the time sequence of the first charging demand, and the time sequence of the first charging duration, charging pile and task matching is performed under the principle of distance priority and time matching. Specifically, the user is matched with the nearest charging pile. For example, if user B is 1 km away from charging pile C and 3 km away from charging pile D, then charging pile C is matched first. Then, it is checked whether the charging pile is available during the user's demand period. For example, if user B needs to charge from 19:00 to 19:30, but charging pile C has a task available from 18:45 to 19:15, then the matching fails. Then, the unmatched tasks (tasks that cannot be assigned due to distance or time conflict) and the matched results (recording the successfully assigned tasks) are output.

[0046] Preferably, for unmatched tasks, task reallocation is performed based on the charging pile's own power supply decision under the distance priority among the matched results. That is, charging piles that are close to the unmatched task and have sufficient remaining power are identified. Then, it is determined whether the charging pile can take on the new task without affecting the existing task. For example, if charging pile F has a task (requiring 20 kWh) from 19:00 to 19:30, but has a total power of 50 kWh, it can provide an additional 30 kWh to meet user B's needs. Then, a marked charging pile is generated, such as charging pile F being marked as a charging pile that can provide reinforcement. If the subsequent task (such as user G from 20:00 to 20:30) cannot be completed due to insufficient power after charging pile F takes on the new task, then user G is marked as having an interrupted task. This improves the service experience and increases the turnover rate of mobile charging pile equipment.

[0047] Step S400: Based on the second charging demand information, perform a second type of charging scheme decision and generate a second charging decision.

[0048] Step S400 further includes step S410, obtaining the distribution information of the operation centers of mobile charging piles; and step S420, performing charging time matching by combining the second charging demand information and the distribution information of the operation centers to generate the second charging decision.

[0049] Preferably, the distribution information of mobile charging station operation centers is obtained, that is, the geographical coordinates (such as longitude and latitude) of each operation center are recorded. For example, operation center i (coordinates: 116.4810°E, 39.9219°N, located in the city center) and operation center j (coordinates: 121.4737°E, 31.2304°N, located in the suburbs). A power constraint rule is set, that is, the minimum reserved power of the mobile charging station when returning to the operation center (such as the remaining power must be ≥15%), to ensure that it can complete the return charging. Then, the charging time matching is performed by combining the second charging demand information and the operation center distribution information. Specifically, according to the second charging demand information, it is determined when the mobile charging station needs to arrive at the designated area and complete the charging task. According to the distance between the current location of the charging station and the operation center, the return time is calculated. For example, if the distance is 10 kilometers and the average speed is 30 kilometers per hour, it will take 20 minutes. At the same time, combined with the current remaining power of the charging station, it is determined whether the power requirement for returning to the operation center after completing the task is met.

[0050] Preferably, time matching rules are then set, including forward time constraints and reverse time derivation. Specifically, the charging station must arrive at the designated area before the task start time, and the charging task execution time (e.g., 1 hour) + return time (20 minutes) ≤ the total duration supported by the charging station's remaining battery power; the departure time from the operation center = task start time - travel time, and the remaining battery power of the charging station at departure must ≥ (outbound battery power + task battery power + return battery power); This generates a second charging decision, which includes scheduling objects, path planning, and time allocation. The system involves scheduling and power allocation. Scheduling the mobile charging station refers to selecting which mobile charging station will perform the task. Route planning involves finding the optimal route from Operations Center A to Business District A (avoiding congested areas). The schedule is as follows: departure at 17:30, arrival at 18:00, charging service from 18:00 to 19:00, and return to Operations Center A before 19:20 (remaining power ≥ 100 kWh). Power allocation uses 300 kWh for the task; for example, 400 kWh - 20 kWh for the outbound trip - 20 kWh for the return trip = 360 kWh available, with a 60 kWh safety margin. This ensures the operational sustainability of the mobile charging stations, improves the efficiency and accuracy of charging station scheduling, and dynamically balances the supply and demand of mobile charging stations.

[0051] Furthermore, step S400 also includes step S430, determining whether each mobile charging pile operation center has a backup mobile charging pile; step S440, if so, using the backup mobile charging pile to execute the charging decision for the unmatched task, and combining the matched result to execute the charging pile's own charging decision after the charging task is completed according to the preset off-peak charging pile allocation strategy.

[0052] Preferably, when an unmatched task is detected (i.e., user demand remains unmet after the initial matching), a backup charging pile query is triggered. This involves determining whether each mobile charging pile operation center has backup mobile charging piles, including monitoring the charging pile status of each operation center and distinguishing between "in-use charging piles" (currently performing charging tasks or in transit) and "backup charging piles" (idle and fully charged / high-charged, ready for immediate use). For example, if operation center A has 10 charging piles, 6 in use and 4 on standby (remaining charge ≥ 90%), and if an operation center has a backup charging pile (≥ 1), the next scheduling step is initiated; otherwise, other methods (such as cross-regional scheduling or prompting the user to wait) are used. If a mobile charging pile operation center has a backup mobile charging pile, matching is performed, including selecting the backup charging pile of the operation center closest to the unmatched task. Since the backup charging pile is idle, it can immediately depart for the target area after the scheduling instruction is issued, shortening the user's waiting time.

[0053] Preferably, after the matching results are completed, a preset off-peak charging pile allocation strategy is followed. Specifically, an off-peak period is defined, which is the low peak period in the first charging demand time sequence (such as 0:00-6:00 AM) or the off-peak period of the power grid price (such as when electricity is cheaper at night). The optimal time for the charging pile to return to the operation center for charging is set. Then, according to the preset off-peak charging pile allocation strategy, the charging pile makes its own charging decision and performs a status judgment after the task is completed. If the remaining power of the charging pile after completing the task is greater than or equal to the minimum return power (such as 20%), it returns to the operation center. If the remaining power is insufficient, it needs to be charged immediately at the nearest charging pile (such as using a fixed charging pile) or wait for rescue. Then, the charging pile is given priority to start charging during the off-peak period to reduce operating costs (such as taking advantage of the low electricity price at night). For example, if the charging pile completes the task and returns to the operation center at 20:00 with 30% remaining power, the system controls it to delay starting charging at 23:00 (the beginning of the off-peak period). It takes 2 hours to fully charge and is completed at 1:00 AM, becoming a backup charging pile. This achieves a dual optimization of supply and demand balance and operational efficiency. Furthermore, the optimized charging strategy can prevent charging piles from charging during peak grid hours, reducing grid pressure and ensuring that backup charging piles are fully charged before the next peak, thereby improving equipment resource utilization.

[0054] Step S500: The charging of the mobile charging station is controlled by the second charging decision.

[0055] Preferably, the second charging decision is translated into specific charging pile scheduling instructions to regulate the operation status of mobile charging piles in real time, ensuring that charging resources efficiently and accurately meet user needs. Specifically, the charging pile scheduling instructions include the target location, the geographical coordinates of the mobile charging pile to be located, such as a shopping mall parking lot or a temporary road location; departure time and route, specifying when the charging pile should depart from its current location and the optimal route to avoid congested sections; task charge, specifying the amount of charging to be provided to the user (e.g., 20 kWh, 50 kWh) or service duration (e.g., 1 hour); and requiring the charging pile to provide information such as location, remaining charge, and task progress at regular intervals (e.g., every 5 minutes); contingency plans for situations such as charging pile malfunction or user cancellation of the task, such as scheduling a nearby charging pile to take over the task; if the charging pile encounters traffic congestion during its journey and is not expected to arrive on time, the route is automatically replanned, an alternative route is selected, and the arrival time is recalculated. If the time is still insufficient, a nearby charging pile is scheduled to take over the task; during the charging process, if the remaining charge of the charging pile falls below a preset safety threshold (e.g., 15%), the current task is automatically terminated, and a return trip is initiated to ensure that the charging pile can return to the operation center for charging. This enables on-demand scheduling, intelligent response, and efficient circulation of mobile charging pile resources, ultimately improving the reliability, convenience, and charging efficiency of new energy charging services.

[0056] In the above text, refer to Figure 1This paper describes in detail a method for predicting the charging demand of mobile charging stations based on big data analysis according to embodiments of the present invention. Next, we will refer to... Figure 2 This invention describes a mobile charging pile charging demand prediction system based on big data analysis according to an embodiment of the present invention.

[0057] The mobile charging pile charging demand prediction system based on big data analysis according to embodiments of the present invention is used to solve the technical problems existing in the prior art, such as the inability to match user dynamic trajectories, difficulty in adapting to dynamically changing charging demands, and unreasonable allocation of charging resources, resulting in inaccurate charging demand prediction, poor charging service efficiency, and poor user experience. It achieves the technical effects of optimizing charging resource allocation, improving the accuracy of charging demand prediction, and enhancing charging service efficiency and user experience. Figure 2 As shown, the mobile charging pile charging demand prediction system based on big data analysis includes: a charging pile distribution map construction module 10, a first charging demand generation module 20, a second charging demand generation module 30, a second charging decision generation module 40, and a mobile charging pile control module 50.

[0058] The charging pile distribution map construction module 10 is used to acquire the location distribution and remaining power distribution of all mobile charging piles within a preset area in real time and construct a charging pile distribution map; the first charging demand generation module 20 is used to perform a first type of charging demand prediction from peak to trough within the preset area based on big data analysis, starting from the current time, and to perform charging task distribution density partitioning and identification to generate a first charging demand time sequence; the second charging demand generation module 30 is used to perform a second type of charging demand prediction of mobile charging piles based on the charging pile distribution map, with the first charging demand time sequence as the task target, and to generate second charging demand information; the second charging decision generation module 40 is used to perform a second type of charging scheme decision based on the second charging demand information and generate a second charging decision; and the mobile charging pile control module 50 is used to control the charging of the mobile charging piles based on the second charging decision.

[0059] The specific configuration of the second charging demand generation module 30 will be described in detail below. The second charging demand generation module 30 further includes: the first type of charging demand is the charging demand for electric vehicle users within the preset area to receive charging services from the mobile charging pile; the second type of charging demand is the charging demand of the mobile charging pile itself.

[0060] The specific configuration of the first charging demand generation module 20 will be described in detail below. The first charging demand generation module 20 further includes: collecting historical charging datasets of mobile charging piles within the preset area, including arbitrary historical charging data including charging location and charging amount at any historical time; performing charging demand analysis at each time point within a single day based on the historical charging dataset to generate a set of historical daily charging amount curves; performing partitioning based on charging task distribution density based on the charging locations within the historical charging dataset to generate a set of historical density partitioning identifiers; using the set of historical daily charging amount curves to perform peak-to-valley period prediction of charging amount starting from the current time to generate the first charging demand time series; and performing density partitioning prediction of peak-to-valley periods using the set of historical density partitioning identifiers to perform charging task distribution density partitioning.

[0061] The specific configuration of the first charging demand generation module 20 will be described in detail below. The first charging demand generation module 20 further includes: collecting charging task cancellation records at any historical time; collecting historical average charging data records for the requesting users corresponding to the charging task cancellation records; and performing corresponding time-based charging volume correction on the historical daily charging volume curve set using the historical average charging data records.

[0062] The specific configuration of the first charging demand generation module 20 will be described in detail below. The first charging demand generation module 20 further includes: marking each curve in the historical daily charging volume curve set with time factor characteristics, including markings for weekdays, holidays, weather, and seasons, generating time factor identifiers; clustering the curves according to the time factor identifiers to generate curve sets corresponding to each time factor identifier; and using the current moment as the prediction starting point, predicting the peak to trough period of charging volume based on time-based and year-on-year comparisons of the curve sets, generating the first charging demand time series.

[0063] The specific configuration of the second charging demand generation module 30 will be described in detail below. The second charging demand generation module 30 further includes: performing density adaptation calculations for charging piles based on the charging task distribution density partitioning identifier of the first charging demand time sequence and the location distribution in the charging pile distribution map, generating a density adaptation index; if the density adaptation index is greater than or equal to a preset adaptation threshold, performing optimal fitting of charging pile task allocation based on the charging pile distribution map and with the first charging demand time sequence as the task target, determining the identified charging piles and identified interruption tasks that cause charging task interruption; calculating the minimum charging demand constraint and charging time constraint of the identified charging pile based on the identified interruption task, generating the second charging demand information.

[0064] The specific configuration of the second charging demand generation module 30 will be described in detail below. The second charging demand generation module 30 further includes: determining the charging task at each moment based on the first charging demand time sequence; predicting the charging duration by combining the charging pile charging voltage information to generate a first charging duration time sequence; combining the charging pile distribution map, the first charging demand time sequence, and the first charging duration time sequence, performing charging pile and task matching under distance priority and time matching to determine unmatched tasks and matched results; for the unmatched tasks, performing task reallocation under distance priority and charging pile self-power decision-making in the matched results to generate the identified charging pile and the identified interrupted task.

[0065] The specific configuration of the second charging decision generation module 40 will be described in detail below. The second charging decision generation module 40 further includes: acquiring the distribution information of mobile charging pile operation centers; combining the second charging demand information and the operation center distribution information to perform charging time matching, and generating the second charging decision.

[0066] The specific configuration of the second charging decision generation module 40 will be described in detail below. The second charging decision generation module 40 further includes: determining whether there are backup mobile charging piles in each mobile charging pile operation center; if so, executing the charging decision of the unmatched task with the backup mobile charging pile, and combining the matched result to execute the charging pile's own charging decision according to the preset off-peak charging pile allocation strategy after the charging task is completed.

[0067] The mobile charging demand prediction system based on big data analysis provided in this invention can execute the mobile charging demand prediction method based on big data analysis provided in any embodiment of this invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0068] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0069] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for predicting charging demand for mobile charging stations based on big data analysis, characterized in that, include: Real-time acquisition of the location distribution and remaining battery power distribution of all mobile charging stations within a preset area, and construction of a charging station distribution map; Starting from the current moment, the first type of charging demand prediction from peak to trough is performed in the preset area through big data analysis, and the distribution density of charging tasks is partitioned and identified to generate the first charging demand time series. Taking the first charging demand time sequence as the task objective, the second type of charging demand prediction for mobile charging piles is performed based on the charging pile distribution map to generate second charging demand information. Based on the second charging demand information, a second type of charging scheme decision is executed to generate a second charging decision. The charging of the mobile charging pile is controlled by the second charging decision. Starting from the current moment, big data analysis is used to predict the first type of charging demand from peak to trough within the preset area, and the charging task distribution density is partitioned to generate the first charging demand time series, including: Collect historical charging data of mobile charging piles within the preset area, including charging location and charging amount at any historical moment; Based on the historical charging dataset, perform charging demand analysis at each time point within a single day to generate a set of historical daily charging volume curves; Based on the charging locations within the historical charging dataset, a partitioning identifier based on the charging task distribution density is generated to produce a set of historical density partitioning identifiers. Density partitioning prediction for peak to trough periods is performed using the historical density partitioning identifier set to identify the density partitioning of charging task distribution; Using the historical daily charging volume curve set, perform charging volume peak to trough period prediction starting from the current time to generate the first charging demand time series; Specifically, taking the first charging demand time series as the task objective, and based on the charging pile distribution map, a second type of charging demand prediction for mobile charging piles is performed to generate second charging demand information, including: Based on the charging task distribution density partition identifier of the first charging demand time sequence, and combined with the location distribution in the charging pile distribution map, the density adaptation calculation of the charging piles is performed to generate the density adaptation index. If the density adaptation index is greater than or equal to the preset adaptation threshold, based on the charging pile distribution map, the optimal fitting of charging pile task allocation is performed with the first charging demand time sequence as the task objective, and the identified charging piles and identified interrupted tasks that cause the charging task to be interrupted are determined. The minimum charging demand constraint and charging time constraint of the identified charging pile are calculated based on the identified interruption task, and the second charging demand information is generated. Based on the charging pile distribution map, and with the first charging demand time sequence as the task objective, the optimal fitting of charging pile task allocation is performed to determine the identified charging piles and tasks that cause charging task interruption, including: The charging task at each moment is determined based on the first charging demand time sequence, and the charging duration is predicted by combining the charging pile charging voltage information to generate the first charging duration time sequence. Combining the charging pile distribution map, the first charging demand time sequence, and the first charging duration time sequence, perform charging pile and task matching under distance priority and time matching to determine unmatched tasks and matched results; For the unmatched tasks, task reallocation is performed in the matched results based on distance priority and the charging pile's own power level, generating the identified charging pile and the identified interrupted task.

2. The method for predicting mobile charging demand based on big data analysis as described in claim 1, characterized in that, The first type of charging demand is the charging demand of mobile charging piles providing charging services to electric vehicle users within the preset area; the second type of charging demand is the charging demand of the mobile charging piles themselves.

3. The method for predicting mobile charging demand based on big data analysis as described in claim 1, characterized in that, The generation of a set of historical daily charging volume curves also includes: Collect charging task cancellation records at any historical moment; Historical average charging data records are collected for the requesting user corresponding to the charging task cancellation record. The historical average charging data records are used to perform the corresponding time-based charging amount correction of the historical daily charging amount curve set.

4. The method for predicting mobile charging demand based on big data analysis as described in claim 1, characterized in that, Using the historical daily charging volume curve set, perform charging volume peak-to-trough time prediction starting from the current time to generate the first charging demand time series, including: Each curve in the set of historical daily charging volume curves is marked with time factor characteristics, including markings for weekdays, holidays, weather, and seasons, to generate a time factor identifier; Cluster the curves according to the time factor identifier to generate a set of curves corresponding to each time factor identifier; Starting from the current moment, the charging demand time series is generated by predicting the peak to trough period of charging volume based on the time-by-time and year-on-year comparisons of the various curve sets.

5. The method for predicting mobile charging demand based on big data analysis as described in claim 1, characterized in that, Based on the second charging demand information, a second type of charging scheme decision is executed to generate a second charging decision, including: Obtain information on the distribution of mobile charging station operation centers; By combining the second charging demand information and the operation center distribution information, charging time matching is performed to generate the second charging decision.

6. The method for predicting mobile charging demand based on big data analysis as described in claim 1, characterized in that, Also includes: Determine whether each mobile charging station operation center has backup mobile charging stations; If so, the backup mobile charging pile executes the charging decision for the unmatched task, and combined with the matched result, executes the charging pile's own charging decision according to the preset off-peak charging pile allocation strategy after the charging task is completed.

7. A mobile charging pile charging demand prediction system based on big data analysis, characterized in that, The system is used to implement the mobile charging pile charging demand prediction method based on big data analysis as described in any one of claims 1 to 6, and the system includes: The charging pile distribution map construction module is used to obtain the location distribution and remaining power distribution of all mobile charging piles within a preset area in real time, and construct a charging pile distribution map. The first charging demand generation module is used to perform peak-to-trough first-type charging demand prediction within the preset area starting from the current time through big data analysis, and to perform charging task distribution density partitioning and identification to generate the first charging demand time sequence. The second charging demand generation module is used to perform second type of charging demand prediction of mobile charging piles based on the charging pile distribution map, with the first charging demand time sequence as the task objective, and generate second charging demand information. The second charging decision generation module is used to execute a second type of charging scheme decision based on the second charging demand information and generate a second charging decision. A mobile charging pile control module is used to control the charging of the mobile charging pile according to the second charging decision.

Images