Time-of-use electricity price driven charging cluster elastic charging method and system

Abstract

The application relates to the technical field of charging control, and discloses a charging cluster elastic charging method and system based on time-of-use electricity price driving. First, future time-of-use electricity prices are predicted, and charging time period constraints of vehicles are obtained, based on which elastic charging times of the vehicles are determined, and initial charging tasks are generated. During task execution, charging pile loads are monitored in real time, when the load of any charging pile is lower than a threshold value, the charging pile is marked as a schedulable charging pile, and charging tasks that have not been completed and whose original execution time period electricity prices are higher than current time electricity prices are screened in the elastic charging time. If there is a schedulable charging pile, the target task is adjusted from the original execution time period to the current time to start execution. The application triggers task adjustment by dynamically monitoring load states and electricity price differences, only migrates charging tasks when the electricity price is reduced, can effectively reduce charging costs, improves charging equipment utilization, and fully guarantees user vehicle use time constraints.

Description

Technical Field

[0001] This invention relates to the field of charging control technology, and in particular to a flexible charging method and system for charging clusters based on time-of-use pricing. Background Technology

[0002] Time-of-use (TOU) pricing mechanisms are adopted to guide users to charge during off-peak hours and balance grid load. Some solutions predict future electricity prices and allocate charging time slots to vehicles based on cost minimization principles, such as scheduling charging tasks entirely within a continuous time window with the lowest predicted electricity prices. These methods treat charging tasks as indivisible, continuous units, and once the initial plan is determined, it is not adjusted, failing to respond to dynamic changes in electricity prices over time. When electricity prices exhibit multi-peak fluctuations or when user usage time happens to punctuate low-price periods, it is impossible to migrate charging tasks from subsequent high-price periods to even lower-price periods, thus limiting cost optimization capabilities.

[0003] More importantly, the lack of a response mechanism for the load status of charging piles means that after the initial plan is executed, it is impossible to detect whether the equipment is currently idle or to use idle periods to execute other charging tasks in advance. This leads to an imbalance where the equipment is overloaded in some periods and idle in others, making it difficult to improve overall charging efficiency and equipment utilization.

[0004] In addition, charging tasks are generally strongly tied to the execution time of charging piles. Once a task is assigned, it is locked within a specific time interval, lacking the flexibility to dynamically adjust the execution time without changing the physical connection of the vehicle. Summary of the Invention

[0005] Therefore, the purpose of this invention is to overcome the problems of static allocation of charging tasks in charging clusters, which cannot respond to fluctuations in electricity prices and changes in load, resulting in high charging costs and low equipment utilization. The invention provides a flexible charging method and system for charging clusters based on time-of-use pricing. By monitoring the load status of charging piles, incomplete charging tasks are adjusted from their original execution time period to the current time period. The adjustment is triggered only when the electricity price at the current time is lower than the electricity price during the original execution time period, thereby reducing charging costs, improving equipment utilization, and ensuring users' vehicle usage time constraints.

[0006] In a first aspect, to solve the above-mentioned technical problems, the present invention provides a method for flexible charging of charging clusters based on time-of-use pricing, comprising: Predict time-of-use electricity prices for future periods to obtain time-series electricity price data; Obtain the charging time constraints of vehicles connected to each charging pile, and determine the flexible charging time of the vehicles based on the electricity price time series data and the charging time constraints. An initial charging task is generated based on the flexible charging time and the rated charging power of the charging pile; Perform the initial charging task and monitor the load of the charging pile; If the load of any of the charging piles is below a threshold, the charging pile is marked as a schedulable charging pile; and incomplete charging tasks are filtered out during the elastic charging time; target tasks are filtered out from the incomplete charging tasks; the electricity price of the original execution period of the target task is higher than the current electricity price; In response to the existence of the schedulable charging pile, the charging task corresponding to the target task is adjusted from the original execution time period to start execution at the current time.

[0007] Preferably, when generating the initial charging task, the method further includes: dividing the initial charging task into multiple sub-tasks according to a power step size, each sub-task including a charging power and an execution period; within the elastic charging time, based on the charging power and the execution period, allocating each sub-task to different time periods of the charging pile for execution; when any sub-task is completed or interrupted, adjusting the charging power and execution period of the remaining sub-tasks so that the total charging amount of the initial charging task is completed within the elastic charging time.

[0008] Preferably, the initial charging task is divided into multiple sub-tasks according to the power step size, including: obtaining the total charging amount of the initial charging task; calculating the shortest continuous charging time of the initial charging task based on the total charging amount and the rated charging power; determining the maximum number of subdivisible sub-tasks of the initial charging task based on the shortest continuous charging time and the total duration of the flexible charging time; dividing the total charging amount into multiple sub-charging amounts according to the maximum number of subdivisible sub-tasks; generating a corresponding sub-task for each sub-charging amount, wherein the charging power contained in each sub-task is equal to the rated charging power; and allocating execution time slots to each sub-task in chronological order within the flexible charging time, such that the execution time slots of adjacent sub-tasks are consecutive, and the sum of the execution time slots of each sub-task is equal to the shortest continuous charging time.

[0009] Preferably, predicting time-of-use electricity prices for future periods and obtaining time-series electricity price data includes: arranging historical time-of-use electricity price data in chronological order to obtain a historical electricity price sequence; performing STL decomposition on the historical electricity price sequence; the STL decomposition uses locally weighted regression to decompose the historical electricity price sequence into a trend component, a periodic component, and a random component; the periodic component includes a daily period and a weekly period; performing linear extrapolation on the trend component to obtain a trend prediction value for future periods; extending the periodic component cyclically over time to obtain a periodic prediction value for future periods; establishing an autoregressive model for the random component and predicting a random prediction value for future periods based on the autoregressive model; and superimposing the trend prediction value, periodic prediction value, and random prediction value to obtain a predicted time-of-use electricity price value for future periods.

[0010] Preferably, the historical time-of-use electricity price data is in hourly time granularity, including the start and end times of each hourly time interval and the corresponding historical electricity price; the predicted time-of-use electricity price for future periods is in hourly time granularity, including the start and end times of each future hourly time interval and the corresponding predicted electricity price.

[0011] Preferably, the selection of the target task further includes time-of-use electricity price preprocessing, which includes: predicting discrete electricity price data in units of time; the discrete electricity price data includes the start time, end time, and corresponding electricity price of each time interval; identifying the boundary time of adjacent time intervals in the discrete electricity price data, and obtaining the predicted electricity price on both sides of the boundary time; when the execution period of the charging task crosses the boundary time, calculating the duration ratio of the execution period on both sides of the boundary time, and calculating the weighted average electricity price of the execution period based on the duration ratio and the predicted electricity price; using the weighted average electricity price as the electricity price corresponding to the execution period for selecting the target task.

[0012] Preferably, determining the flexible charging time of the vehicle based on the electricity price time series data and the charging period constraints includes: determining the rechargeable time interval of the vehicle based on the charging period constraints; identifying continuous time intervals in the electricity price time series data where the electricity price is lower than a threshold and marking them as candidate low-price intervals; performing an intersection operation on the rechargeable time intervals and each of the candidate low-price intervals to obtain at least one candidate flexible sub-interval; if any candidate flexible sub-interval is greater than or equal to the shortest continuous charging time, then the candidate flexible sub-interval with the longest duration is taken as the flexible charging time.

[0013] Preferably, if all of the candidate flexible sub-intervals are less than the shortest continuous charging time, then the continuous period with the lowest electricity price within the rechargeable time interval shall be taken as the flexible charging time.

[0014] Preferably, when there are multiple target tasks, task matching is performed according to the following priority order: First priority: select the target task with the largest difference between the current electricity price and the electricity price during the original execution period of the target task; Second priority: select the target task with the shortest remaining flexible charging time window.

[0015] Secondly, to solve the above-mentioned technical problems, the present invention provides a flexible charging system for charging clusters based on time-of-use pricing, comprising: The electricity price forecasting module is used to predict the time-of-use electricity price for future periods and obtain time-series electricity price data; The vehicle information acquisition module is used to acquire the charging time constraints of vehicles connected to each charging pile; The flexible time determination module is used to determine the flexible charging time of the vehicle based on the electricity price time series data and the charging time period constraints. The task generation module is used to generate an initial charging task based on the flexible charging time and the rated charging power of the charging pile. A load monitoring module is used to detect the load of the charging pile when the initial charging task is performed; The scheduling and discrimination module is used to mark the charging pile as a schedulable charging pile when the load of any of the charging piles is lower than a threshold; and to filter out incomplete charging tasks within the elastic charging time; and to filter out target tasks from the incomplete charging tasks. The task adjustment module is used to adjust the charging task corresponding to the target task to start execution at the current time.

[0016] Compared with the prior art, the above-described technical solution of the present invention has the following advantages: First, each scheduling adjustment is triggered by the current electricity price being lower than the price during the original execution period, and the adjustment direction always shifts the task from a future high-price period to the current low-price period. Therefore, each adjustment reduces the cost of the corresponding charging task. As the system continuously monitors and repeats this adjustment process every time idle resources become available, the execution time of charging tasks will continuously converge towards lower-price periods, resulting in a significant decrease in overall charging costs.

[0017] Secondly, by monitoring and marking schedulable charging piles, it is possible to promptly detect and utilize every segment of idle capacity, bringing forward charging tasks originally planned for the future to the current idle period. This dynamic filling mechanism smooths out the load curve of the charging cluster, effectively utilizing potentially idle periods and alleviating potentially overloaded periods by bringing forward tasks, thus significantly improving equipment utilization.

[0018] Furthermore, since the scheduling decision is based on real-time monitoring of the load status and the current electricity price, and does not rely on accurate prediction of future states, this method has a natural tolerance for uncertainties such as electricity price prediction deviations, user time changes, and temporary equipment failures.

[0019] Finally, compared to methods that schedule charging tasks as a whole within a continuous period, this invention can utilize multiple low-price periods simultaneously, significantly reducing overall charging costs and solving the problem that the overall scheduling of charging tasks cannot adapt to multi-peak electricity prices.

[0020] In summary, the flexible charging method and system for charging clusters based on time-of-use pricing described in this invention monitors the load status of charging piles and adjusts unfinished charging tasks from their original execution time period to the current time period. Moreover, the adjustment is triggered only when the electricity price at the current time is lower than the electricity price during the original execution time period, thereby reducing charging costs, improving equipment utilization, and ensuring users' vehicle usage time constraints. Attached Figure Description

[0021] To make the content of this invention easier to understand, the invention will be further described in detail below with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of a preferred embodiment of the flexible charging method for charging clusters based on time-of-use pricing in the present invention; Figure 2 This is a flowchart illustrating the generation of the initial charging task in a preferred embodiment of the present invention. Figure 3 This is a flowchart illustrating the splitting of multiple subtasks in a preferred embodiment of the present invention; Figure 4 This is a flowchart illustrating the prediction of time-of-use electricity prices for future periods in a preferred embodiment of the present invention; Figure 5 This is a flowchart of the time-of-use electricity pricing preprocessing in a preferred embodiment of the present invention; Figure 6 This is a flowchart for determining the flexible charging time of a vehicle in a preferred embodiment of the present invention; Figure 7 This is a structural block diagram of a time-of-use pricing-driven flexible charging system for charging clusters, as described in a preferred embodiment of the present invention. Detailed Implementation

[0022] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0023] This invention provides a flexible charging method and system for charging clusters based on time-of-use pricing. The core concept is to monitor the load status of each charging pile in the charging cluster, identify idle resources, and use the price difference as an economic incentive signal to dynamically shift unfinished charging tasks from future high-price periods to the current low-price periods for execution, thereby achieving the dual goals of reducing charging costs and improving equipment utilization.

[0024] Example 1: This embodiment of the invention discloses a flexible charging method for charging clusters based on time-of-use pricing, including: S100: Predict time-of-use electricity prices for future periods and obtain electricity price time-series data; S200: Obtain the charging time constraints of vehicles connected to each charging pile, and determine the flexible charging time of the vehicles based on the electricity price time series data and the charging time constraints. S300 generates an initial charging task based on the flexible charging time and the rated charging power of the charging pile; S400: Performs the initial charging task and monitors the load on the charging station; S500. If the load of any charging pile is lower than the threshold, mark the charging pile as a schedulable charging pile; and filter out incomplete charging tasks within the elastic charging time; filter out target tasks from the incomplete charging tasks; the electricity price of the original execution period of the target task is higher than the current electricity price. S600: In response to the existence of a schedulable charging pile, the charging task corresponding to the target task is adjusted from the original execution time period to start execution at the current time.

[0025] The following combination Figure 1 The specific implementation process will be explained in detail.

[0026] Step S100: Obtain time-of-use (TOU) electricity price data for future time periods as the basis for dispatch decisions. Specifically, historical TOU electricity price data is obtained through power trading platforms, grid data service platforms, or third-party electricity price data service providers. This data is granular in minutes, including the start and end times of each 15-minute time interval and the corresponding historical electricity price. The historical data is arranged in chronological order to form a historical electricity price sequence.

[0027] In order to extract the patterns of electricity price changes from historical data, a time series decomposition method is used to process the historical electricity price series. The decomposition method divides the original electricity price series into three components with different physical meanings: trend component, periodic component, and random component.

[0028] For the trend component, a linear extrapolation method is used, which calculates the trend value for the future period by extending the rate of change of the trend component along the same direction in historical periods. For the periodic component, a cyclic extension method is used, which repeats historical periodic patterns according to their inherent period length to obtain the periodic value for the future period. For the stochastic component, an autoregressive model is established to predict the stochastic value for the future period based on the historical value sequence of the stochastic component itself. Finally, the trend prediction, periodic prediction, and stochastic prediction values ​​are superimposed to obtain the time-of-use electricity price prediction for the future period at a minute-by-minute granularity.

[0029] Step S200: Determine the flexible charging time for each vehicle connected to the charging pile. Flexible charging time refers to the most suitable time window for performing the charging task while meeting the user's vehicle usage time constraints. The user's vehicle usage time constraints are reflected in the charging time period constraints, which include at least the current time the vehicle is connected to the charging pile and the user's expected usage time.

[0030] The available charging time range for the vehicle is determined based on the current time and the expected usage time, which is the time period from the current time to the expected usage time. Then, continuous time intervals with lower electricity prices are identified in the electricity price time series data. The available charging time intervals are matched with these low-price intervals, and the time periods that meet both the user's time constraints and have lower electricity prices are selected as flexible charging times.

[0031] Determining flexible charging times requires considering two factors: first, the electricity price level, selecting periods with lower electricity prices to reduce charging costs; and second, the duration, ensuring the selected period can accommodate the shortest continuous charging time required to complete the charging task. This shortest continuous charging time is calculated by dividing the vehicle's total charging demand by the charging station's rated charging power. If the length of the low-price period within the available charging time range is insufficient, the continuous period with the lowest electricity price within the available charging time range is used as the flexible charging time, ensuring charging is completed within a limited time and at the lowest possible cost.

[0032] Step S300: Generate an initial charging task based on the flexible charging time and the rated charging power of the charging pile. The essence of the initial charging task is to allocate the charging demand of each vehicle to a specific time slot within its flexible charging time and specify the charging power for each time slot. Specifically, using the rated charging power of the charging pile as the charging power, arrange consecutive charging time slots for each vehicle within the flexible charging time, ensuring that the length of each charging time slot is exactly equal to the shortest continuous charging time required to complete charging. If the total length of the flexible charging time is greater than the shortest continuous charging time, the charging task is scheduled to be executed within the continuous sub-interval with the lowest electricity price within the flexible charging time. Through the above method, an initial charging plan is generated for each connected vehicle, including the corresponding charging pile, charging start time, charging end time, and charging power.

[0033] Step S400: Monitor the load status of each charging pile. The load status can be the ratio of the current output power of the charging pile to the rated power, or a Boolean value indicating whether the charging pile is currently in operation.

[0034] In step S500, when the current load of any charging pile is detected to be lower than a preset threshold, such as a load rate below 10% or the charging pile being completely idle, the charging pile is marked as a schedulable charging pile. The essence of this marking action is to confirm that there is currently unused charging capacity in the system, providing a physical basis for the early execution of subsequent tasks. The load threshold can be set according to the actual operating characteristics of the charging cluster, for example, set to 15% of the rated power or a fixed power value of 5kW.

[0035] While marking available charging stations, incomplete charging tasks are filtered within the flexible charging time; incomplete charging tasks refer to tasks that have not yet started execution. From these incomplete tasks, target tasks are further filtered based on the condition that the electricity price during the original execution period of the target task is higher than the current electricity price. This condition ensures that every scheduling adjustment has a positive benefit: only when the current electricity price is lower will the charging task be moved from the originally scheduled high-price period to the current time, thereby inevitably reducing the charging cost of the task. The original execution period electricity price refers to the electricity price corresponding to the execution period allocated to the charging task in the initial charging plan, and the current electricity price refers to the electricity price corresponding to the current time of the monitored available charging station; the comparison between these two electricity prices determines whether the filtering condition is met.

[0036] In step S600, in response to the existence of at least one schedulable charging pile, the charging task corresponding to the target task is adjusted from its original execution time period to the current time. This adjustment is performed on the charging pile to which the vehicle was originally connected, meaning the charging task is always completed on the charging pile physically connected to it. Specifically, the system cancels or postpones the planned charging task during its original execution time period and assigns a new charging task to the charging pile where the task is located at the current time, allowing the charging task to begin execution from the current time. After the adjustment, the status information of the charging task is updated, marking the charging plan corresponding to the original execution time period as canceled, and the adjusted charging plan begins execution at the current time.

[0037] In the above embodiments, when generating the initial charging task, if the flexible charging time window is long (e.g., 8 hours at night) and the electricity price fluctuates within the time window with multiple peaks such as low-high-low, the overall scheduling can only arrange the charging task within a continuous time period, either selecting the early low-price period or the later low-price period, and cannot simultaneously utilize two or more dispersed low-price periods, resulting in limited electricity price adaptability. Furthermore, when the charging task fails to complete as planned due to equipment failure, temporary user interruption, or other reasons, the overall scheduling lacks a fine-grained adjustment mechanism for the remaining charging capacity, which may result in the charging task not being completed within the flexible charging time, affecting user vehicle use.

[0038] To address the issues of the overall scheduling of charging tasks being unable to adapt to multi-peak electricity prices and lacking dynamic correction capabilities in the event of execution anomalies, a method for subtask decomposition and elastic scheduling of charging tasks is further provided based on the above-mentioned embodiments. The following combines... Figures 2-3 The specific implementation methods will be described in detail.

[0039] When generating the initial charging task, it is divided into multiple subtasks according to the power step size. The power step size refers to the charging power unit corresponding to each subtask, and its value can be the same as the rated charging power of the charging pile, or it can be an integer fraction of the rated charging power. Each subtask contains two basic attributes: charging power and execution period. The charging power represents the power level when the subtask is executed, and the execution period represents the specific time interval allocated to the subtask within the elastic charging time.

[0040] The specific implementation process of subtask decomposition is as follows: First, obtain the total charging amount of the initial charging task. This total charging amount represents the total amount of battery power that the vehicle needs to charge, in kilowatt-hours (kWh). Simultaneously, obtain the rated charging power of the charging pile. This rated charging power represents the maximum power that the charging pile can continuously output under normal operating conditions, in kilowatts (kW). Based on the total charging amount and rated charging power, calculate the shortest continuous charging time for the initial charging task. The calculation formula is: the shortest continuous charging time equals the total charging amount divided by the rated charging power. For example, if the total charging amount is 60 kWh and the rated charging power is 60 kW, then the shortest continuous charging time is 1 hour; if the total charging amount is 30 kWh and the rated charging power is 60 kW, then the shortest continuous charging time is 0.5 hours.

[0041] Next, the total duration of the flexible charging time is obtained, representing the time from the earliest allowed start time to the latest required completion time. Based on the shortest continuous charging time and the total duration of the flexible charging time, the maximum number of subdivisions for the initial charging task is determined. The principle for calculating the maximum number of subdivisions is: while ensuring that the sum of the execution periods of all subtasks equals the shortest continuous charging time, the charging task is divided into as many subtasks as possible, with each subtask executed consecutively. Specifically, the maximum number of subdivisions is equal to the total duration of the flexible charging time divided by the shortest continuous charging time, rounded down, and then multiplied by a coefficient related to the power step size. When the power step size equals the rated charging power, the maximum number of subdivisions is the integer part of the total duration of the flexible charging time divided by the shortest continuous charging time. This design ensures that all subtasks after being divided can be completed sequentially within the flexible charging time, without exceeding the user-defined vehicle usage time boundaries.

[0042] Then, the total charging capacity is divided into multiple sub-charging capacities according to the maximum divisible number. The division can be equal, meaning each sub-charging capacity equals the total charging capacity divided by the maximum divisible number; or it can be unequal, according to a preset allocation ratio, such as prioritizing larger sub-charging capacities for periods with lower electricity prices. A corresponding sub-task is generated for each sub-charging capacity, and each sub-task contains a charging power equal to the rated charging power of the charging station. Since the charging power is constant, the execution time of each sub-task is determined by dividing its sub-charging capacity by the rated charging power.

[0043] Finally, within the elastic charging time, execution periods are allocated to each subtask sequentially. The allocation rule is as follows: the execution period of the first subtask begins at the start of the elastic charging time and continues for its corresponding duration; the execution period of the second subtask begins immediately after the end of the first subtask, and so on. In this way, the execution periods of adjacent subtasks are connected end-to-end, forming a continuous time chain, and the sum of the execution periods of all subtasks equals the shortest continuous charging time. This end-to-end allocation method ensures the continuity of charging tasks on the timeline, avoiding charging interruptions or wasted time due to gaps in the time frame.

[0044] After the subtasks are assigned, each subtask is executed according to the assignment results. During the flexible charging period, each subtask is assigned to different time periods of the charging pile for execution. Since the execution periods of each subtask are consecutive on the timeline, the charging pile executes different subtasks sequentially at different time periods, which appears as a continuous charging process overall, but is internally divided into multiple independent subtask units.

[0045] During subtask execution, the execution status of each subtask is monitored. When any subtask is completed, its completion information is recorded, including actual completion time and actual charging amount. When any subtask is interrupted due to device failure, temporary user interruption, communication anomaly, or other reasons, the interruption event is detected and the charging amount completed at the time of interruption is retrieved. Regardless of whether a subtask is completed normally or interrupted abnormally, the adjustment process for the remaining subtasks is triggered.

[0046] The adjustment process first calculates the remaining charging capacity, which is the total charging capacity of the initial charging task minus the sum of the actual charging capacity of all completed sub-tasks. Then, the remaining duration of the flexible charging time is obtained, i.e., the time from the current moment to the end of the flexible charging time. Based on the remaining charging capacity and remaining duration, the charging power and execution time allocation scheme for the remaining sub-tasks are recalculated. Specifically, if the remaining duration is sufficient, the rated charging power remains unchanged, and the remaining sub-tasks are redistributed to consecutive time slots within the remaining duration, ensuring that the execution time slots of each sub-task are consecutive. If the remaining duration is insufficient, the charging power of the remaining sub-tasks is increased to be higher than the rated charging power but not exceeding the maximum allowable power of the charging pile, thereby shortening the required execution time and ensuring that all remaining charging capacity is completed within the flexible charging time. If the remaining duration is severely insufficient, a warning message is sent to the user, indicating that charging may not be completed before the expected usage time. Through the above dynamic adjustments, the total charging capacity of the initial charging task is ensured to be completed within the flexible charging time.

[0047] In this embodiment, the preset value of the power step size can be flexibly configured according to the actual application scenario. For vehicles with large battery capacity, a larger power step size can be used to reduce the number of subtasks and reduce system overhead; for scenarios with frequent electricity price fluctuations, a smaller power step size can be used to improve the accuracy of electricity price adaptation.

[0048] Compared to methods that schedule the entire charging task within a continuous time period, this embodiment can utilize multiple low-cost time periods simultaneously, significantly reducing overall charging costs. Furthermore, the fine-grained breakdown of subtasks allows for adjustments to the execution time of only some subtasks without moving the entire charging task, reducing computational overhead. Moreover, by dividing continuous high-power charging into multiple subtasks, the natural intervals between them provide the battery with a buffer time for heat dissipation and chemical equilibrium. Even if subtasks are consecutive, the switching process of charging power itself produces slight intervals, which helps reduce battery temperature rise during prolonged high-power charging and extends battery cycle life.

[0049] Based on the above embodiments, a method for predicting time-of-use electricity prices and a method for preprocessing electricity price boundaries are further provided. The following combines... Figures 4-5 The specific implementation process will be explained in detail.

[0050] First, predict time-of-use (TOU) electricity prices for future periods to obtain time-series electricity price data. Historical TOU electricity price data is obtained from power trading platforms, grid data service platforms, or third-party electricity price data service providers. This historical TOU electricity price data is granular in minutes, including the start and end times of each 15-minute time interval and the corresponding historical electricity price. For example, a data record can be represented as: start time 0:00, end time 0:15, historical electricity price 0.30 yuan / kWh. The obtained historical TOU electricity price data is arranged in chronological order to form a historical electricity price sequence. This sequence is a one-dimensional array that increases with time, with each element corresponding to the electricity value at the top of a 15-minute interval.

[0051] The historical electricity price series is decomposed using STL (Sustainable Time Series Decomposition). STL stands for Seasonal Trend Decomposition Based on Locally Weighted Regression, a robust time series decomposition method. STL decomposition uses locally weighted regression as the core smoothing tool to decompose the historical electricity price series into three components with different physical meanings. The first component is the trend component, reflecting the overall trend of electricity prices over a longer period, such as the year-on-year increase or decrease in electricity prices due to changes in fuel costs, policy adjustments, or changes in supply and demand. The second component is the periodic component, reflecting the regular fluctuations in electricity prices within a fixed time period. In this embodiment, the periodic component includes daily and weekly cycles: the daily cycle is based on a 24-hour period, reflecting the peak, flat, and trough patterns of electricity prices within each day, such as lower electricity prices at night, higher prices during the day, and potential peak electricity prices in the evening; the weekly cycle is based on a 168-hour period (24 hours multiplied by 7 days), reflecting the price differences between weekdays and weekends. The third component is the random component, which reflects random fluctuations that cannot be explained by the trend component and the periodic component, such as instantaneous abnormal fluctuations in electricity prices caused by extreme weather, sudden equipment failures, or temporary policy adjustments.

[0052] STL decomposition employs locally weighted regression for smoothing. Locally weighted regression is a nonparametric regression method that selects several neighboring data points for each target point, assigns different weights based on their distance, and then performs a weighted polynomial fit. Data points closer to the target point have higher weights, while those farther away have lower weights. In this way, locally weighted regression can smooth the data while preserving local characteristics. STL decomposition extracts trend and periodic components through an inner loop iteration and uses robust weights in an outer loop to reduce the interference of outliers on the decomposition results.

[0053] After completing the STL decomposition, predictions are made for the three components separately: For the trend component, a linear extrapolation method is used. Linear extrapolation refers to extending the trend component in the same direction along the historical time period to calculate the trend value for the future time period based on the rate of change of the trend component over a historical period. Specifically, a linear function of the trend component is fitted, its slope is calculated, and then extended into the future according to this slope to obtain the predicted trend values ​​for each future time period. For the periodic component, a cyclic extension method is used. Cyclic extension refers to repeating the periodic component according to its inherent period length to obtain the periodic values ​​for the future time period. Since the periodic component has a stable periodicity, historical periodic patterns are directly copied to future time periods: daily periodic patterns are copied in 24-hour cycles, and weekly periodic patterns are copied in 168-hour cycles. For the stochastic component, an autoregressive model is established for prediction. The autoregressive model assumes that the random value at the current moment can be represented by a linear combination of random values ​​from several past moments plus a random error term. The coefficients of the autoregressive model are estimated based on historical random component data, and then the model is used to predict the stochastic values ​​for the future time period.

[0054] Finally, the trend forecast, periodic forecast, and random forecast are superimposed to obtain the time-of-use electricity price forecast for future periods. These time-of-use electricity price forecasts for future periods are also in minute-level time granularity, including the start and end times of each future 15-minute time interval and the corresponding predicted electricity price, forming complete time-series electricity price data.

[0055] Since electricity price data is discrete, and the execution period of charging tasks may span the boundary moments of adjacent time intervals, the comparison of electricity prices is distorted. In order to solve this problem, the following solution is proposed.

[0056] The process involves acquiring discrete electricity price data, calculated in minutes, and then calculating the hourly average price for each hour. This average price is the sum of the prices for the four 15-minute intervals within that hour, forming hourly time-series electricity price data to reduce the magnitude of data spanning adjacent time intervals. This discrete price data includes the start and end times of each time interval and the corresponding predicted price. For example, if the electricity prices for the four 15-minute intervals from 0:00 to 1:00 are 0.35 yuan for 0:00-0:15, 0.30 yuan for 0:15-0:30, 0.30 yuan for 0:30-0:45, and 0.25 yuan for 0:45-1:00, then the hourly average price for that interval is 0.30 yuan. If the price is 0.30 yuan for the 0:00-1:00 interval and 0.50 yuan for the 1:00-2:00 interval, with 1:00 as the boundary time, the calculation is more efficient.

[0057] Identify the boundary moments of adjacent time intervals in discrete electricity price data, and obtain the predicted electricity prices on both sides of the boundary moment. Taking the aforementioned 0:00-1:00 and 1:00-2:00 as examples, the boundary moment is 1:00, the electricity price on the left is 0.30 yuan, and the electricity price on the right is 0.50 yuan.

[0058] When the execution period of a charging task spans a boundary time, for example, the execution period is from 0:50 to 1:10, which spans the boundary time of 1:00, the electricity price of 0.30 yuan in the 0:00-1:00 interval or 0.50 yuan in the 1:00-2:00 interval cannot be simply used to represent the electricity price of the entire execution period, because both of these approaches will produce significant errors.

[0059] The correct approach is to calculate the percentage of time spent on either side of the boundary time. In the example of 0:50 to 1:10, the total duration of the execution period is 20 minutes. The duration on the left side of the boundary time (0:50 to 1:00) is 10 minutes, accounting for 50%; the duration on the right side of the boundary time (1:00 to 1:10) is also 10 minutes, accounting for 50%. Then, calculate the weighted average electricity price based on the percentage of time spent and the predicted electricity prices on both sides of the boundary time: the weighted average electricity price equals the percentage of time spent on the left side multiplied by the electricity price on the left side plus the percentage of time spent on the right side multiplied by the electricity price on the right side, i.e., 50% × 0.30 yuan + 50% × 0.50 yuan = 0.40 yuan. This 0.40 yuan is the weighted average electricity price for this execution period, which more accurately reflects the actual charging cost than simply using 0.30 yuan or 0.50 yuan.

[0060] The calculated weighted average electricity price is used as the electricity price for the execution period, and is used for subsequent electricity price comparisons and target task selection. For execution periods that fall entirely within a time interval and do not cross any boundary moments, the electricity price corresponding to that interval is used directly, without the need for weighted average calculation.

[0061] Through the above-mentioned time-of-use pricing preprocessing, when the electricity price during the original execution period of the target task is higher than the current electricity price, the electricity value used is subject to boundary smoothing to ensure the accuracy of the electricity price comparison.

[0062] Based on the above embodiments, a method for determining the flexible charging time is further provided, which is described below in conjunction with... Figure 6 The specific implementation process will be explained in detail.

[0063] First, the available charging time interval for the vehicle is determined based on charging time constraints. These constraints include the vehicle's current connection time and the user's planned usage time. The current connection time refers to the moment the vehicle physically connects to the charging station, while the planned usage time is the moment the user intends to use the vehicle. The available charging time interval is the period between the current connection time and the planned usage time. Within this interval, the vehicle can charge at any time, but must complete the entire charging process before the planned usage time to ensure uninterrupted normal vehicle use.

[0064] Secondly, the system identifies continuous time intervals in the electricity price time-series data where the electricity price is below a preset threshold, and marks these intervals as candidate low-price intervals. The electricity price time-series data consists of the previously predicted time-of-use electricity price data for future periods, including the start and end times of each hourly time interval and the corresponding predicted electricity price. The preset threshold can be set according to actual needs; for example, it can be set as the median of the current day's electricity price, the average value of the electricity price time-series data, or a fixed electricity price level.

[0065] The system iterates through the electricity price time series data, identifying all consecutive time intervals where the electricity price at every point in time within the interval is below a preset threshold. These consecutive intervals are then marked as candidate low-price intervals. For example, if the electricity price time series data shows that the price is 0.28 yuan from 2:00 to 3:00, 0.26 yuan from 3:00 to 4:00, 0.27 yuan from 4:00 to 5:00, and 0.29 yuan from 5:00 to 6:00, and the preset threshold is 0.30 yuan, then 2:00 to 6:00 constitutes a consecutive candidate low-price interval.

[0066] Then, the intersection operation is performed on the rechargeable time interval and each candidate low-price interval. The intersection operation means taking the common part of the two time intervals. For each candidate low-price interval, the overlap between it and the rechargeable time interval is calculated to obtain a candidate flexible sub-interval. For example, if the rechargeable time interval is 0:00 to 5:00 and the candidate low-price interval is 2:00 to 6:00, then the intersection of the two is 2:00 to 5:00, which is one candidate flexible sub-interval. If the rechargeable time interval intersects with multiple candidate low-price intervals, then multiple candidate flexible sub-intervals are obtained. For example, if the rechargeable time interval is 0:00 to 8:00, and there are two candidate low-price intervals, 2:00 to 4:00 and 6:00 to 7:00, then the system obtains two candidate flexible sub-intervals: 2:00-4:00 and 6:00-7:00.

[0067] After obtaining candidate elastic sub-intervals, the optimal one is selected as the elastic charging time. The selection logic performs a distribution process based on whether the duration of the candidate elastic sub-interval meets the charging demand.

[0068] First, determine the vehicle's shortest continuous charging time. This shortest continuous charging time is calculated based on the vehicle's total charging demand and the charging station's rated charging power. The formula is: Shortest continuous charging time equals total charging demand divided by rated charging power. For example, if the vehicle needs 60 kWh of charging and the charging station's rated power is 60 kW, then the shortest continuous charging time is 1 hour.

[0069] Iterate through all candidate elastic sub-intervals and determine whether the duration of each candidate elastic sub-interval is greater than or equal to the shortest continuous charging duration. If at least one candidate elastic sub-interval satisfies this condition, it indicates that one or more complete low-price windows exist within the rechargeable time interval, and the length of these windows is sufficient to accommodate all charging tasks. In this case, the candidate elastic sub-interval with the longest duration is selected as the elastic charging time. The reason for selecting the longest duration is that a longer low-price window means a greater time margin to schedule charging tasks. Even if there are slight fluctuations in electricity prices within the window, the sub-interval with the lowest electricity price within the window can be selected for charging, thereby further reducing charging costs. At the same time, a longer window also provides greater scheduling flexibility, facilitating advance adjustments to tasks in subsequent steps.

[0070] If the duration of all candidate flexible sub-intervals is less than the shortest continuous charging duration, it means that no low-price window within the available charging time interval can meet the full charging demand in terms of duration. This situation usually occurs when the user's expected vehicle usage time is tight, or when there are only scattered low-price periods within the available charging time interval, and each period is insufficient in length. In this case, it is impossible to schedule the charging task completely within any low-price interval, so a backup strategy is required.

[0071] The backup strategy is to select the consecutive period with the lowest electricity price within the available charging time interval as the flexible charging time. Specifically, a window with a length equal to the shortest consecutive charging time is slid within the available charging time interval, the average electricity price of each window is calculated, and the window with the lowest average electricity price is selected as the flexible charging time. If the total duration of the available charging time interval is less than the shortest consecutive charging time, the entire available charging time interval is used as the flexible charging time, and a warning is issued to the user, indicating that it may not be possible to complete the full charging before the expected usage time.

[0072] Through the above-mentioned diversion process, the optimal flexible charging time can be determined in different scenarios: when there is a low-price window of sufficient duration, the low-price window is fully utilized to reduce charging costs; when there is no low-price window of sufficient duration, the charging time with the lowest relative cost is selected within a limited time to complete the charging and ensure the user's vehicle usage needs.

[0073] In this embodiment, the solution achieves the joint solution of user time constraints and electricity price constraints by intersecting the rechargeable time interval and the candidate low price interval. By introducing the shortest continuous charging time as a threshold, the problem is divided into two scenarios and selection strategies are designed for each. The longest duration priority rule reserves optimization space for subsequent flexible scheduling, ensuring that the charging task is completed on time and reducing charging costs.

[0074] Based on the above embodiments, in the actual operation scenarios of screening target tasks, especially when the electricity price is in a downward trend or the system has just switched from a busy state to an idle state, there are often multiple charging tasks that meet the screening conditions at the same time, resulting in multiple target tasks coexisting.

[0075] When there are multiple target tasks, it is not possible to adjust all target tasks to be executed at the current time at the same time.

[0076] To address the matching issue among multiple target tasks, this embodiment employs a two-level priority order. The first priority is to select the target task with the largest difference between the current electricity price and the electricity price during the original execution period. The difference between the current electricity price and the electricity price during the original execution period reflects the unit electricity price reduction that can be achieved by advancing the charging task from its original execution period to the current time. A larger difference means a greater reduction in charging costs and more significant economic benefits from advancing the task. Therefore, prioritizing the target task with the largest price difference ensures that each scheduling adjustment generates maximum immediate cost-benefit.

[0077] When multiple target tasks share the same first priority (i.e., their electricity price differences are equal or similar), the second priority is activated: the task with the shortest remaining flexible charging time window is selected. The remaining flexible charging time window refers to the length of time from the current moment to the end of the target task's flexible charging period. A shorter remaining flexible charging time window means a tighter time constraint for the task. If not adjusted in time, it may not be possible to find a suitable low-price period to complete charging within the remaining time, or it may be forced to execute during a higher-price period. Therefore, the target task with the shortest remaining flexible charging time window is prioritized to ensure that time-sensitive tasks receive priority scheduling opportunities.

[0078] The target tasks are sorted according to the above priority order to form a task matching queue. Then, based on the available power capacity of the schedulable charging piles, target tasks are sequentially retrieved from the queue, and the charging tasks corresponding to the target tasks are adjusted from their original execution time to the current time. Each time a target task is adjusted, the remaining available power capacity of the schedulable charging piles is updated. When the remaining available capacity is insufficient to handle the full charging power of the next target task, the system stops the current round of matching, and the remaining target tasks wait for the next scheduling cycle.

[0079] Example 2: Refer to Figure 7 As shown, this embodiment of the invention provides a flexible charging system for charging clusters driven by time-of-use pricing, comprising: The electricity price forecasting module is used to predict the time-of-use electricity price for future periods and obtain time-series electricity price data; The vehicle information acquisition module is used to acquire the charging time constraints of vehicles connected to each charging pile; The flexible time determination module is used to determine the flexible charging time of the vehicle based on the electricity price time series data and charging period constraints. The task generation module is used to generate initial charging tasks based on the flexible charging time and the rated charging power of the charging pile. The load monitoring module is used to detect the load of the charging pile when performing the initial charging task; The scheduling and discrimination module is used to mark a charging pile as a schedulable charging pile when the load of any charging pile is lower than the threshold; and to filter out incomplete charging tasks within the elastic charging time; and to filter out target tasks from the incomplete charging tasks. The task adjustment module is used to adjust the charging task corresponding to the target task to start execution at the current time.

[0080] The embodiments of the present invention are used to implement the time-of-use pricing-driven flexible charging of charging clusters in any of the above embodiments. They have the same inventive concept and the same beneficial effects as Embodiment 1, and will not be repeated here.

[0081] In summary, the flexible charging method and system for charging clusters based on time-of-use pricing described in this invention monitors the load status of charging piles and adjusts unfinished charging tasks from their original execution time period to the current time period. Moreover, the adjustment is triggered only when the electricity price at the current time is lower than the electricity price during the original execution time period, thereby reducing charging costs, improving equipment utilization, and ensuring users' vehicle usage time constraints.

[0082] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A flexible charging method for charging clusters based on time-of-use pricing, characterized in that, include: Predict time-of-use electricity prices for future periods to obtain time-series electricity price data; Obtain the charging time constraints of vehicles connected to each charging pile, and determine the flexible charging time of the vehicles based on the electricity price time series data and the charging time constraints. An initial charging task is generated based on the flexible charging time and the rated charging power of the charging pile; Perform the initial charging task and monitor the load of the charging pile; If the load of any of the charging piles is below a threshold, the charging pile is marked as a schedulable charging pile; and incomplete charging tasks are filtered out during the elastic charging time; target tasks are filtered out from the incomplete charging tasks; the electricity price of the original execution period of the target task is higher than the current electricity price; In response to the existence of the schedulable charging pile, the charging task corresponding to the target task is adjusted from the original execution time period to start execution at the current time. Generating an initial charging task includes: dividing the initial charging task into multiple sub-tasks according to a power step size, each sub-task including a charging power and an execution period; and, within the elastic charging time, allocating each sub-task to different time periods of the charging pile for execution based on the charging power and the execution period. The process of filtering the target tasks also includes time-of-use electricity price preprocessing, which includes: predicting discrete electricity price data in units of time; the discrete electricity price data includes the start time, end time, and corresponding electricity price of each time interval; identifying the boundary times of adjacent time intervals in the discrete electricity price data, and obtaining the predicted electricity price on both sides of the boundary time; when the execution period of the charging task crosses the boundary time, calculating the duration ratio of the execution period on both sides of the boundary time, and calculating the weighted average electricity price of the execution period based on the duration ratio and the predicted electricity price; and using the weighted average electricity price as the electricity price corresponding to the execution period for filtering the target tasks.

2. The flexible charging method for charging clusters based on time-of-use pricing as described in claim 1, characterized in that, When generating the initial charging task, the method further includes: when any of the sub-tasks is completed or interrupted, adjusting the charging power and execution period of the remaining sub-tasks so that the total charging amount of the initial charging task is completed within the elastic charging time.

3. The flexible charging method for charging clusters based on time-of-use pricing as described in claim 2, characterized in that, The initial charging task is divided into multiple sub-tasks according to the power step size, including: Obtain the total charging amount of the initial charging task, and calculate the shortest continuous charging time of the initial charging task based on the total charging amount and the rated charging power; The maximum number of initial charging tasks that can be split is determined based on the shortest continuous charging duration and the total duration of the elastic charging time. The total charging capacity is divided into multiple sub-charging capacities according to the maximum divisible number, and a corresponding sub-task is generated for each sub-charging capacity. The charging power contained in each sub-task is equal to the rated charging power. During the elastic charging time, execution time slots are allocated to each subtask in chronological order, so that the execution time slots of adjacent subtasks are consecutive, and the sum of the execution time slots of each subtask is equal to the shortest continuous charging time.

4. The flexible charging method for charging clusters based on time-of-use pricing as described in claim 1, characterized in that, To predict time-of-use electricity prices for future periods, obtain time-series electricity price data, including: Arrange historical time-of-use electricity price data in chronological order to obtain a historical electricity price sequence; The historical electricity price series is subjected to STL decomposition; the STL decomposition uses locally weighted regression to decompose the historical electricity price series into trend components, periodic components, and stochastic components; the periodic components include daily and weekly periods; The trend components are linearly extrapolated to obtain the trend prediction values ​​for future periods; The periodic components are cyclically extended over time to obtain the periodic prediction values ​​for future time periods; An autoregressive model is established for the random components, and random predicted values ​​for future time periods are predicted based on the autoregressive model. By superimposing the trend forecast, periodic forecast, and random forecast values, the time-of-use electricity price forecast for future periods is obtained.

5. The flexible charging method for charging clusters based on time-of-use pricing as described in claim 4, characterized in that, The historical time-of-use electricity price data is granular in hours, including the start and end times of each hourly time interval and the corresponding historical electricity price; the predicted time-of-use electricity price for future periods is granular in hours, including the start and end times of each future hourly time interval and the corresponding predicted electricity price.

6. The method for flexible charging of charging clusters based on time-of-use pricing as described in claim 1, characterized in that, Based on the electricity price time-series data and the charging time constraints, the flexible charging time of the vehicle is determined, including: Based on the charging time constraints, the available charging time range for the vehicle is determined; In the electricity price time series data, identify continuous time intervals where the electricity price is below a threshold and mark them as candidate low price intervals; Perform an intersection operation between the rechargeable time interval and each of the candidate low-price intervals to obtain at least one candidate elastic sub-interval; If any of the candidate elastic sub-intervals is greater than or equal to the shortest continuous charging time, then the candidate elastic sub-interval with the longest duration shall be the elastic charging time.

7. The flexible charging method for charging clusters based on time-of-use pricing as described in claim 6, characterized in that, If all of the candidate flexible sub-intervals are less than the shortest continuous charging time, then the continuous period with the lowest electricity price within the rechargeable time interval shall be taken as the flexible charging time.

8. The method for flexible charging of charging clusters based on time-of-use pricing as described in claim 1, characterized in that, When multiple target tasks exist, task matching is performed according to the following priority order: First priority: Select the target task with the largest difference between the current electricity price and the electricity price during the original execution period of the target task; Second priority: Select the task with the shortest remaining elastic charging time window of the target task.

9. A time-of-use pricing-driven flexible charging system for charging clusters, used to execute the method according to any one of claims 1-8, characterized in that, The system includes: The electricity price forecasting module is used to predict the time-of-use electricity price for future periods and obtain time-series electricity price data; The vehicle information acquisition module is used to acquire the charging time constraints of vehicles connected to each charging pile; The flexible time determination module is used to determine the flexible charging time of the vehicle based on the electricity price time series data and the charging time period constraints. The task generation module is used to generate an initial charging task based on the flexible charging time and the rated charging power of the charging pile. A load monitoring module is used to detect the load of the charging pile when the initial charging task is performed; The scheduling and discrimination module is used to mark the charging pile as a schedulable charging pile when the load of any of the charging piles is lower than a threshold; and to filter out incomplete charging tasks within the elastic charging time; and to filter out target tasks from the incomplete charging tasks. The task adjustment module is used to adjust the charging task corresponding to the target task to start execution at the current time; Generating an initial charging task includes: dividing the initial charging task into multiple sub-tasks according to a power step size, each sub-task including a charging power and an execution period; and, within the elastic charging time, allocating each sub-task to different time periods of the charging pile for execution based on the charging power and the execution period.

Images