Data-driven method for configuring electric vehicle light storage charging station

Abstract

The present application relates to the technical field of charging station control, and more particularly to a data-driven electric vehicle light storage charging station configuration method, which comprises the following steps: constructing a charging starting sequence of each vehicle in a behavior time window of a target date; determining a charging regularity of the vehicle in the behavior time window of the target date according to the difference between adjacent time points in the charging starting sequence and the charging frequency of the vehicle in the behavior time window; dividing the vehicle into a common user vehicle and an uncommon user vehicle according to the charging regularity of the vehicle; and processing a common user vehicle charging quantity sequence and an uncommon user vehicle charging quantity sequence in the behavior time window of the target date using a prediction algorithm respectively to obtain a common user vehicle charging prediction quantity and an uncommon user vehicle charging prediction quantity, so as to control the target energy storage quantity of an energy storage system in the light storage charging station on the prediction date and effectively improve the configuration accuracy of the electric vehicle light storage charging station.

Description

Technical Field

[0001] This invention relates to the field of charging station control technology. More specifically, this invention relates to a data-driven method for configuring electric vehicle photovoltaic-storage charging stations. Background Technology

[0002] With the widespread adoption of electric vehicles, electric vehicle charging stations, as a new type of power load unit, are experiencing rapid growth in both number and scale. Photovoltaic-storage charging stations, combining photovoltaic power generation and energy storage systems, can cope with peak loads on the traditional power grid, improve the absorption capacity of renewable energy, and reduce operating costs. In the operation of photovoltaic-storage charging stations, the rational configuration of the energy storage system is crucial for achieving economical and reliable operation.

[0003] The capacity configuration of an energy storage system directly impacts the performance and efficiency of a power station. If the capacity is too small, it may be unable to meet the charging needs of all users when photovoltaic power generation is insufficient or charging demand surges, leading to service interruptions and affecting user experience. Conversely, if the capacity is too large, it will result in excessively high investment costs for energy storage equipment, wasted idle resources, and unnecessary cycle losses in batteries, significantly increasing operating costs and reducing overall economic efficiency. Therefore, accurately predicting future charging loads is a crucial prerequisite for optimizing the configuration of energy storage systems.

[0004] However, existing load forecasting methods often exhibit insufficient accuracy when dealing with the complex load fluctuations of photovoltaic-storage charging stations. Traditional forecasting models (such as exponential smoothing) typically treat the total charging load of a charging station as a single, holistic time series for analysis and modeling. However, in actual user charging load behavior, there are significant differences in charging time and visit frequency between regular users and non-regular users. Traditional models do not distinguish between these two groups during forecasting, resulting in an inability to accurately capture the essential characteristics of load changes and low accuracy in electricity load forecasting. Therefore, how to differentiate the behavioral habits of different groups to improve the accuracy of electricity load forecasting is a problem that needs to be solved. Summary of the Invention

[0005] To address the aforementioned technical problem of how to differentiate the behavioral habits of different groups to improve the accuracy of electricity load forecasting, this invention proposes a data-driven method for configuring electric vehicle photovoltaic-storage charging stations. This method includes the following steps:

[0006] Taking any day as the target date, a behavioral time window is constructed from the historical charging records starting from the target date. The start charging time of each vehicle in the behavioral time window is sorted in ascending order to obtain the time points in the charging start sequence of each vehicle. Based on the differences between adjacent time points in the charging start sequence of the vehicle on the target date and the number of times the vehicle is charged in the behavioral time window, the charging regularity of the vehicle in the behavioral time window of the target date is determined. According to the charging regularity, the vehicles are divided into regular user vehicles and non-regular user vehicles. The daily charging amount of regular user vehicles and non-regular user vehicles in the behavioral time window of the target date is obtained to obtain the charging amount sequence of regular user vehicles and the charging amount sequence of non-regular user vehicles in the behavioral time window. The charging amount sequence of regular user vehicles and the charging amount sequence of non-regular user vehicles in the behavioral time window of the target date are processed by prediction algorithms to obtain the predicted charging amount of regular user vehicles and the predicted charging amount of non-regular user vehicles on the predicted day of the target date, so as to control the target energy storage capacity of the energy storage system in the photovoltaic-storage charging station on the predicted day.

[0007] This invention improves the accuracy of energy storage control by predicting the target energy storage capacity of a photovoltaic-energy storage charging station and then controlling its energy storage. When predicting the target energy storage capacity, this invention calculates charging regularity indicators through hierarchical modeling of user behavior and decouples the stable load of regular users from the random fluctuations of non-regular users based on these indicators, avoiding the misjudgment of trends by traditional single models. Furthermore, by predicting separately for regular and non-regular users, this invention achieves accurate predictions based on their respective electricity consumption data, improving the accuracy of energy storage demand prediction, reducing the risk of overcharging or power shortages, and effectively enhancing the accuracy of charging capacity allocation for electric vehicle photovoltaic-energy storage charging stations.

[0008] According to the data-driven electric vehicle photovoltaic-storage charging station configuration method provided by the present invention, the step of constructing a behavioral time window for the target date in historical charging records starting from the target date further includes: collecting vehicle charging records daily, wherein the charging records include at least the vehicle identification code, charging start timestamp, and charging amount; converting the charging start timestamp into a numerical timestamp to obtain the start charging time; and setting the next day's record of the target date as the day to be predicted for the target date.

[0009] This invention provides a high-quality data foundation for subsequent pattern analysis by standardizing timestamps.

[0010] According to the data-driven electric vehicle photovoltaic-storage charging station configuration method provided by the present invention, determining the charging regularity of the vehicle within a target date behavior time window includes:

[0011] ;

[0012] For the target date action time window The charging patterns of individual vehicles For the first The number of times a vehicle charges within the target date and time window. The maximum number of charges within the target date and time window. For the target date action time window The absolute value variance of the difference between adjacent time points in the vehicle charging start sequence This represents the maximum absolute variance of the differences between adjacent time points in the charging start sequence of all vehicles within the target date time window. It is an exponential function with base e.

[0013] This invention provides a precise method for calculating charging regularity indicators. By combining charging frequency and charging time stability, abstract user behavior can be transformed into quantifiable values. It features data-driven, objective calculations for user segmentation, providing a basis for accurate user segmentation.

[0014] According to the data-driven electric vehicle photovoltaic-storage charging station configuration method provided by the present invention, the step of classifying vehicles into regular user vehicles and non-regular user vehicles based on the charging regularity of the vehicles includes: sorting the charging regularity of all vehicles in the target date behavior time window in ascending order, obtaining the maximum difference between adjacent charging regularities; using the mean of adjacent charging regularities corresponding to the maximum difference as a segmentation threshold; and obtaining regular user vehicles and non-regular user vehicles based on the comparison result between the charging regularity of the vehicles and the segmentation threshold.

[0015] This invention provides a precise method for calculating segmentation thresholds. By finding the maximum interval in the distribution of regular indicators, the classification threshold is automatically determined. This method can adapt to the distribution characteristics of user behavior at different charging stations and at different times, thereby improving the accuracy of user segmentation.

[0016] According to the data-driven electric vehicle photovoltaic-storage charging station configuration method provided by the present invention, the step of obtaining regular user vehicles and non-regular user vehicles based on the comparison result of the charging regularity of vehicles and the segmentation threshold includes: if the charging regularity of a vehicle is greater than the segmentation threshold, then the vehicle is a regular user vehicle; otherwise, the vehicle is a non-regular user vehicle.

[0017] According to the data-driven electric vehicle photovoltaic-storage charging station configuration method provided by the present invention, the predicted charging volume of regular users and the predicted charging volume of non-regular users on the target date are obtained by using an exponential smoothing algorithm.

[0018] According to the data-driven electric vehicle photovoltaic-storage charging station configuration method provided by the present invention, the step of processing the charging volume sequence of regular user vehicles and the charging volume sequence of non-regular user vehicles in the behavioral time window of the target date using prediction algorithms respectively further includes: obtaining the week number of the day to be predicted for the target date; fitting the non-regular user vehicle charging volume corresponding to the number of days in the behavioral time window of the target date that have the same week number as the day to be predicted to obtain the periodic total charging volume of non-regular user vehicles for the day to be predicted; and obtaining the target energy storage capacity for the day to be predicted based on the predicted charging volume of regular user vehicles, the predicted charging volume of non-regular user vehicles, and the periodic total charging volume of non-regular user vehicles for the day to be predicted.

[0019] According to the data-driven electric vehicle photovoltaic-storage charging station configuration method provided by the present invention, obtaining the target energy storage capacity for the predicted day includes:

[0020] ;

[0021] The target energy storage capacity for the predicted date. Forecast of charging volume for regular user vehicles on the predicted date. Predicted charging volume for non-user vehicles on the forecast date. The total periodic charging amount for non-user vehicles on the predicted date. This is the function for finding the maximum value.

[0022] This invention provides an accurate method for calculating the target energy storage capacity for the predicted day. By compensating for the periodicity that may exist in the load of extraordinary users, it takes into account both the short-term growth trend of the load and the periodic peaks, thereby effectively avoiding the problem of insufficient energy storage caused by underestimating the sudden charging demand of extraordinary users and improving the accuracy and reliability of the prediction results.

[0023] According to the data-driven configuration method for electric vehicle photovoltaic-storage charging stations provided by the present invention, the method for controlling the target energy storage capacity of the energy storage system in the photovoltaic-storage charging station on a predicted day includes: acquiring the energy storage capacity in the energy storage system; generating a control signal based on the difference between the energy storage capacity and the target energy storage capacity on the predicted day; inputting the control signal into the energy storage system to adjust the energy storage capacity in the energy storage system to the target energy storage capacity; and reducing the charging power and entering trickle charging mode in response to the energy storage capacity in the energy storage system reaching the target energy storage capacity.

[0024] According to the data-driven configuration method for electric vehicle photovoltaic-storage charging stations provided by the present invention, the method further includes controlling the target energy storage capacity of the energy storage system in the photovoltaic-storage charging station on the predicted day, and then performing real-time anomaly monitoring on the charging process of the electric vehicle photovoltaic-storage charging station.

[0025] The present invention has the following beneficial effects:

[0026] Based on the above technical solutions, the data-driven configuration method for electric vehicle photovoltaic-storage charging stations provided by this invention improves the accuracy of energy storage control by predicting the target energy storage capacity of the charging station and then controlling its energy storage. When predicting the target energy storage capacity, this invention calculates charging regularity indicators through hierarchical modeling of user behavior and decouples the stable load of regular users from the random fluctuations of non-regular users based on these indicators, avoiding the misjudgment of trends by traditional single models. Furthermore, by predicting separately for regular and non-regular users, this invention can accurately predict energy storage demand based on their respective electricity consumption data, improving the accuracy of energy storage demand prediction, reducing the risk of overcharging or power shortage, and effectively improving the accuracy of charging capacity configuration for electric vehicle photovoltaic-storage charging stations. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating the steps of a data-driven method for configuring electric vehicle photovoltaic-storage charging stations according to an embodiment of the present invention. Detailed Implementation

[0028] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0029] Please see Figure 1 , Figure 1 This is a flowchart illustrating the steps of a data-driven electric vehicle photovoltaic-energy storage charging station configuration method according to an embodiment of the present invention. This method constructs charging regularity indicators by quantitatively analyzing the charging behavior of each vehicle in historical data to dynamically identify and separate regular users from non-regular users. Then, based on the load characteristics of the two types of users, a differentiated prediction strategy is adopted to accurately obtain the target energy storage capacity. The method specifically includes the following steps:

[0030] S1: Take any day as the target date, obtain the predicted date of the target date, and retrieve the historical charging records of the photovoltaic and energy storage charging station on the target date.

[0031] It's important to note that accurate load analysis and prediction for charging stations require comprehensive and detailed historical operational data. Each photovoltaic-storage charging station typically has an independent Data Management System (DMS), which records and stores key information from all operational processes. Therefore, data retrieval and extraction from the photovoltaic-storage charging station's DMS can provide a data foundation for subsequent analysis.

[0032] For example, in an embodiment of the present invention, obtaining historical charging records of a photovoltaic-storage charging station includes: collecting vehicle charging records daily, wherein the charging records include at least a vehicle identification code, a charging start timestamp, and a charging amount; converting the charging start timestamp into a numerical timestamp to obtain the start charging time; and recording the next day of the target date as the predicted day of the target date.

[0033] When collecting vehicle charging records, a one-day cycle is used, that is, the collection starts at 0:00 and ends at 23:59, and the corresponding week number is obtained for each day.

[0034] Specifically, you can retrieve the total charging volume corresponding to the field by querying with the Structured Query Language (SQL) or calling the data access API provided by the system, and obtain the daily charging records by using the date as the primary key.

[0035] The vehicle identification number (VIN) serves as a unique ID for each vehicle, providing access to a set of charging records for each vehicle. The charging start timestamp may include both date and time, potentially leading to lower accuracy in subsequent differential calculations. Therefore, it is possible to... The function converts the charging start timestamp into a numeric timestamp.

[0036] Understandably, when predicting the target energy storage capacity required on the day to be predicted for the target date, the data for the target date is the closest data. Therefore, a behavioral time window can be constructed from the historical charging records starting from the target date. Within this behavioral time window, the charging behavior of each vehicle can be analyzed to distinguish between regular users and non-regular users, where the target date is the latest ongoing day.

[0037] S2: Construct a behavioral time window for the target date from the historical charging records, starting from the target date, to obtain the time points in the charging start sequence of each vehicle; determine the charging regularity of the vehicle in the behavioral time window based on the differences between adjacent time points in the charging start sequence of the vehicle on the target date and the number of times the vehicle is charged in the behavioral time window.

[0038] When constructing a behavioral time window for the target date in historical charging records, starting from the target date, the day before the date to be predicted can be used as the cutoff point, and 30 complete days can be traced back to form the time window required for the analysis.

[0039] It's important to note that, in order to effectively distinguish between regular users with stable charging behavior and non-regular users with random behavior, a quantitative indicator is needed to objectively evaluate the regularity of user charging behavior. This indicator should comprehensively reflect two core dimensions of user charging: the stability of charging frequency and the consistency of charging time intervals. Regular users typically charge at the same charging station at similar times and with high frequency, making their behavior patterns highly predictable.

[0040] Based on this, embodiments of the present invention can analyze the charging behavior patterns of each vehicle within a behavioral time window of a target date to obtain its regularity.

[0041] For example, the start charging time of each vehicle in the behavior time window is sorted in ascending order to obtain the time points in the charging start sequence of each vehicle.

[0042] Understandably, if a vehicle is charged less than or equal to one time within the target date time window, it is considered a non-user and can be directly marked. Therefore, the following steps only apply to vehicles that have been charged more than one time. The variance of the difference between adjacent time points in the charging start sequence for each vehicle reflects the fluctuation of the charging start time. The specific calculation method is as follows:

[0043] ;

[0044] For the target date action time window The absolute value variance of the difference between adjacent time points in the vehicle charging start sequence For the first The number of times a vehicle charges within the target date and time window. For the target date action time window The first vehicle charging start sequence The absolute value of the difference between adjacent time points in a group. For the target date action time window The mean of the absolute values ​​of the differences between adjacent time points in the vehicle charging start sequence.

[0045] For example, the results obtained according to the above steps The charging regularity of vehicles within the target date time window is determined, and the specific calculation method is as follows:

[0046] ;

[0047] For the target date action time window The charging patterns of individual vehicles For the first The number of times a vehicle charges within the target date and time window. The maximum number of charges within the target date and time window. For the target date action time window The absolute value variance of the difference between adjacent time points in the vehicle charging start sequence This represents the maximum absolute variance of the differences between adjacent time points in the charging start sequence of all vehicles within the target date time window. It is an exponential function with base e.

[0048] in, The first time window reflecting the target date behavior The relative charging frequency of a vehicle. The larger this value, the more frequently the vehicle is charged within the time window corresponding to the target date, and more frequently than other vehicles.

[0049] The first time window reflecting the target date behavior The relative instability of charging for each vehicle; the smaller this value, the more unstable the charging becomes within the target date and time window. The more stable the charging intervals for each vehicle, the better.

[0050] In summary, if The larger and The smaller the value, the more likely it is that the target date / behavior is in the time window. The more frequently a vehicle charges compared to other vehicles and the more stable the charging intervals, the more likely that the vehicle is a regular user of the charging station, has a high degree of dependence on the charging station, and its behavior is highly predictable, thus the stronger the charging regularity. Conversely, if a vehicle charges less frequently and the charging intervals are more stable, its regularity is considered weaker.

[0051] Based on the above steps, the charging regularity of each vehicle within the target date behavior time window can be obtained. The higher the charging regularity, the greater the likelihood that the vehicle is a frequent user vehicle. Based on this, the vehicles can be classified.

[0052] S3: Based on the charging regularity of the vehicles, divide the vehicles into regular user vehicles and non-regular user vehicles, obtain the daily charging amount of regular user vehicles and non-regular user vehicles in the behavioral time window of the target date, and obtain the charging amount sequence of regular user vehicles and non-regular user vehicles in the behavioral time window.

[0053] It should be noted that urban photovoltaic-storage charging stations are often located in areas with high user concentration, exhibiting a clear differentiation in charging behavior. One group consists of regular users with stable travel patterns, whose charging frequency is high and highly regular; the other group consists of non-regular users who visit randomly, whose charging behavior is sporadic and dispersed in time. After calculating the charging regularity indicators for all vehicles, these indicators typically exhibit a bimodal distribution, meaning that highly regular users and less regular users naturally form two clusters. Therefore, this embodiment of the invention can divide vehicles into regular and non-regular user groups by setting a dynamic threshold, thereby separating their respective historical daily charging volume sequences and laying the foundation for subsequent differentiated predictions.

[0054] For example, in this embodiment of the invention, classifying vehicles into regular user vehicles and non-regular user vehicles based on their charging patterns includes: sorting the charging patterns of all vehicles in the target date time window in ascending order and obtaining the maximum difference between adjacent charging patterns; using the mean of adjacent charging patterns corresponding to the maximum difference as a segmentation threshold; and obtaining regular user vehicles and non-regular user vehicles based on the comparison result between the vehicle's charging patterns and the segmentation threshold.

[0055] The resulting segmentation threshold is the inflection point between highly regular and lowly regular groups, and this segmentation can maximize the differences between classes.

[0056] For example, in an embodiment of the present invention, the determination of regular user vehicles and non-regular user vehicles based on the comparison result of the charging regularity of the vehicle and the segmentation threshold includes: if the charging regularity of the vehicle is greater than the segmentation threshold, then the vehicle is a regular user vehicle; otherwise, the vehicle is a non-regular user vehicle.

[0057] Specifically, vehicles with charging regularity indicators greater than the segmentation threshold are marked as regular users, and their vehicle identification numbers are stored in the regular user vehicle identification number set. Correspondingly, vehicles with charging regularity indicators less than or equal to the threshold are marked as non-regular users, and their vehicle identification numbers are stored in the non-regular user vehicle identification number set.

[0058] S4: Use prediction algorithms to process the regular user vehicle charging sequence and the non-regular user vehicle charging sequence in the behavioral time window of the target date, respectively, to obtain the predicted regular user vehicle charging volume and the predicted non-regular user vehicle charging volume for the day to be predicted on the target date.

[0059] For example, in an embodiment of the present invention, an exponential smoothing algorithm is used to obtain the predicted charging amount of regular users' vehicles and the predicted charging amount of non-regular users' vehicles for the target date.

[0060] The steps of using the exponential smoothing algorithm to obtain the predicted charging volume of regular users' vehicles and the predicted charging volume of non-regular users' vehicles for the target date can be implemented by existing technologies, and will not be described in detail here in this embodiment of the invention.

[0061] It should be noted that the prediction algorithm can be used to obtain the predicted charging volume of regular users' vehicles and the predicted charging volume of non-regular users' vehicles on the target date.

[0062] However, regular users are highly predictable, so smoothing algorithms can be used directly to predict energy storage capacity. Unregistered users, on the other hand, may experience a weekend / holiday effect, where a large number of randomly arriving vehicles charge on weekends, causing significant fluctuations in charging volume with a recurring weekly pattern, while the changes are relatively gradual on normal working days. When predicting the charging volume sequence of unregistered users, relying solely on smoothing algorithms will ignore the influence of these periodic external factors, leading to an underestimation of energy storage capacity.

[0063] Therefore, when forecasting the electricity consumption of extraordinary users, it is necessary to consider and compare the time-series estimates of the smoothing algorithm with the historical trends of the same period to ensure that the electricity consumption forecast of extraordinary users can meet the demand.

[0064] For example, in this embodiment of the invention, the prediction algorithm is used to process the charging sequence of regular user vehicles and the charging sequence of non-regular user vehicles in the behavioral time window of the target date, and the method further includes: obtaining the week number of the day to be predicted for the target date; fitting the charging volume of non-regular user vehicles corresponding to the number of days in the behavioral time window of the target date that are the same as the week number of the day to be predicted to obtain the total periodic charging volume of non-regular user vehicles for the day to be predicted; and obtaining the target energy storage capacity for the day to be predicted based on the predicted charging volume of regular user vehicles, the predicted charging volume of non-regular user vehicles, and the total periodic charging volume of non-regular user vehicles for the day to be predicted.

[0065] The least squares method can be used for fitting, and the specific implementation steps can be achieved by existing technology. The embodiments of the present invention will not be described in detail here.

[0066] For example, to illustrate the number of days in the target date behavior time window that are the same as the week number of the day to be predicted: If the week number of the day to be predicted is Wednesday, then all Wednesdays are obtained in the target date behavior time window, which is the number of days in the target date behavior time window that are the same as the week number of the day to be predicted.

[0067] For example, in an embodiment of the present invention, the target energy storage capacity for the predicted day is obtained, and the specific calculation method is as follows:

[0068] ;

[0069] The target energy storage capacity for the predicted date. Forecast of charging volume for regular user vehicles on the predicted date. Predicted charging volume for non-user vehicles on the forecast date. The total periodic charging amount for non-user vehicles on the predicted date. This is the function for finding the maximum value.

[0070] Among these, the charging behavior of extraordinary users is sudden and uncontrollable. Therefore, when planning energy storage systems, it is necessary to maintain as much reserve capacity as possible, for example, some extraordinary users may charge during peak hours outside of weekends. Therefore, the embodiments of the present invention... The maximum value is taken between the predicted charging volume of non-user vehicles on the predicted date and the total periodic charging volume of non-user vehicles on the predicted date, to ensure that demand can be met under both periodic peaks and trend growth, thereby improving the reliability of the prediction results.

[0071] The target energy storage capacity for the predicted day can be accurately obtained by following the above steps. Adjusting the energy storage capacity for the predicted day based on the target energy storage capacity can effectively improve the accuracy of controlling the energy storage system in the photovoltaic-energy storage charging station.

[0072] S5: Based on the predicted charging volume of regular users' vehicles and the predicted charging volume of non-regular users' vehicles on the forecast date, control the target energy storage capacity of the energy storage system in the photovoltaic-storage charging station on the forecast date.

[0073] It should be noted that the target energy storage capacity obtained from the above steps is the core parameter guiding the operation of the energy storage system. The control system of the photovoltaic-storage charging station will intelligently adjust the storage and utilization strategies of photovoltaic power generation based on this target to achieve energy conservation, emission reduction, and optimized operation.

[0074] For example, in an embodiment of the present invention, controlling the target energy storage capacity of the energy storage system in the photovoltaic-energy storage charging station on a predicted day includes: acquiring the energy storage capacity in the energy storage system; generating a control signal based on the difference between the energy storage capacity and the target energy storage capacity on the predicted day; inputting the control signal into the energy storage system to adjust the energy storage capacity in the energy storage system to the target energy storage capacity; and reducing the charging power and entering trickle charging mode in response to the energy storage capacity in the energy storage system reaching the target energy storage capacity.

[0075] The trickle charging mode can be set to a charging power equal to 10% of the charging power required to reach the target energy storage capacity. This setting can be customized to meet specific needs. Trickle charging mode maintains the energy storage system at full charge for unforeseen circumstances while minimizing energy loss and battery cycle damage caused by continuous full-power charging. This ensures service efficiency while reducing operating energy consumption and carbon emissions.

[0076] For example, in an embodiment of the present invention, the energy storage system in the photovoltaic-energy storage charging station is controlled to achieve a target energy storage capacity on a predicted day, and then the process further includes: real-time anomaly monitoring of the charging process of the electric vehicle photovoltaic-energy storage charging station.

[0077] As can be seen, in this embodiment of the invention, when configuring an electric vehicle photovoltaic-storage charging station, any day can be used as the target date. A behavioral time window for the target date is constructed from historical charging records, starting from the target date. The start charging time of each vehicle within the behavioral time window is sorted in ascending order to obtain the time points in the charging start sequence for each vehicle. Based on the differences between adjacent time points in the vehicle charging start sequence for the target date and the number of times the vehicle is charged within the behavioral time window, the charging regularity of the vehicle within the behavioral time window for the target date is determined. Based on the charging regularity of the vehicles, they are divided into regular user vehicles and non-regular user vehicles. The daily charging amounts of regular user vehicles and non-regular user vehicles within the behavioral time window for the target date are obtained, resulting in a regular user vehicle charging amount sequence and a non-regular user vehicle charging amount sequence within the behavioral time window. Prediction algorithms are then used to process the regular user vehicle charging amount sequence and the non-regular user vehicle charging amount sequence within the behavioral time window for the target date, respectively, to obtain the predicted charging amounts of regular user vehicles and non-regular user vehicles for the predicted day of the target date. This controls the target energy storage capacity of the energy storage system in the photovoltaic-storage charging station on the predicted day, effectively improving the accuracy of the electric vehicle photovoltaic-storage charging station configuration.

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

Claims

1. A data-driven method for configuring electric vehicle photovoltaic-storage charging stations, characterized in that, include: Take any day as the target date, construct a behavior time window for the target date in the historical charging records, and sort the start charging time of each vehicle in the behavior time window in ascending order to obtain the time point in the charging start sequence of each vehicle. Based on the differences between adjacent time points in the vehicle charging start sequence on the target date and the number of times the vehicle charges within the behavioral time window, the charging regularity of the vehicle within the behavioral time window on the target date is determined, including: ; For the target date action time window The charging patterns of individual vehicles For the first The number of times a vehicle charges within the target date and time window. The maximum number of charges within the target date and time window. For the target date action time window The absolute value variance of the difference between adjacent time points in the vehicle charging start sequence This represents the maximum absolute variance of the differences between adjacent time points in the charging start sequence of all vehicles within the target date time window. It is an exponential function with base e; the charging behavior pattern comprehensively reflects two core dimensions of user charging: the stability of the number of charging cycles and the consistency of the charging time interval; Based on the charging patterns of vehicles, vehicles are divided into regular user vehicles and non-regular user vehicles. The daily charging amount of regular user vehicles and non-regular user vehicles within the behavioral time window of the target date is obtained, resulting in the charging amount sequence of regular user vehicles and the charging amount sequence of non-regular user vehicles within the behavioral time window. The prediction algorithm is used to process the charging sequence of regular users and the charging sequence of non-regular users in the behavioral time window of the target date, respectively, to obtain the predicted charging amount of regular users and the predicted charging amount of non-regular users on the day to be predicted, so as to control the target energy storage capacity of the energy storage system in the photovoltaic-storage charging station on the predicted day.

2. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 1, characterized in that, The process of constructing a behavioral time window for the target date from historical charging records, starting with the target date, previously included: Collect vehicle charging records daily. The charging records include at least the vehicle identification number, charging start timestamp, and charging amount. Convert the charging start timestamp into a numerical timestamp to obtain the charging start time. The next day's record for the target date is used as the predicted day for the target date.

3. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 1, characterized in that, The method of classifying vehicles into regular user vehicles and non-regular user vehicles based on their charging patterns includes: The charging patterns of all vehicles within the target date time window are sorted in ascending order, and the maximum difference between adjacent charging patterns is obtained. The mean of the adjacent charging patterns corresponding to the maximum difference is used as the segmentation threshold. Based on the comparison between the vehicle's charging pattern and the segmentation threshold, regular user vehicles and non-regular user vehicles are identified.

4. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 3, characterized in that, The process of identifying regular user vehicles and non-regular user vehicles based on a comparison of the vehicle's charging regularity with a segmentation threshold includes: If the charging regularity of a vehicle exceeds the segmentation threshold, the vehicle is considered a regular user vehicle; otherwise, the vehicle is considered a non-regular user vehicle.

5. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 1, characterized in that, The exponential smoothing algorithm is used to obtain the predicted charging volume of regular users' vehicles and the predicted charging volume of non-regular users' vehicles for the target date.

6. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 1, characterized in that, The process of processing the regular user vehicle charging volume sequence and the non-regular user vehicle charging volume sequence within the behavioral time window of the target date using prediction algorithms also includes: Obtain the weekday number of the day to be predicted for the target date, and fit the non-user vehicle charging amount corresponding to the number of days in the target date behavior time window that are the same as the weekday number of the day to be predicted to obtain the periodic total charging amount of non-user vehicles for the day to be predicted; based on the predicted charging amount of regular user vehicles, the predicted charging amount of non-user vehicles, and the periodic total charging amount of non-user vehicles for the day to be predicted, obtain the target energy storage capacity for the day to be predicted.

7. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 6, characterized in that, The process of obtaining the target energy storage capacity for the predicted day includes: ； The target energy storage capacity for the predicted date. Forecast of charging volume for regular user vehicles on the predicted date. Predicted charging volume for non-user vehicles on the forecast date. The total periodic charging amount for non-user vehicles on the predicted date. This is the function for finding the maximum value.

8. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 1, characterized in that, The target energy storage capacity of the energy storage system in the photovoltaic-energy storage charging station on the predicted day includes: The system acquires the energy storage capacity in the energy storage system, generates a control signal based on the difference between the energy storage capacity and the target energy storage capacity for the predicted day, and inputs the control signal into the energy storage system to adjust the energy storage capacity in the energy storage system to the target energy storage capacity; in response to the energy storage capacity in the energy storage system reaching the target energy storage capacity, the system reduces the charging power and enters trickle mode.

9. The data-driven configuration method for electric vehicle photovoltaic-storage charging stations according to claim 1, characterized in that, The method for controlling the energy storage system in the photovoltaic-energy storage charging station to achieve the target energy storage capacity on the predicted day also includes: Real-time anomaly monitoring is performed on the charging process of electric vehicle photovoltaic-storage charging stations.

Images