User charging scheduling system and method

Abstract

The application discloses a user charging planning system and method, comprising a data acquisition module for acquiring multi-source data; a cloud intelligent planning module for performing three-level optimization processing based on the multi-source data, the three-level optimization processing comprising: first-level long-term prediction planning for predicting a charging demand window and a minute-level energy consumption sequence of a future trip based on the multi-source data; second-level short-term fine planning for running a multi-objective optimization algorithm to generate a plurality of candidate planning schemes in combination with the multi-source data and the minute-level energy consumption sequence when the charging demand window is approaching; and third-level real-time dynamic adjustment for monitoring state deviation in the process of performing charging planning control based on any candidate planning scheme, and triggering a re-planning algorithm to generate an adjustment scheme based on the state deviation; and a charging execution module for performing charging planning control. The application can realize multi-source data fusion analysis and support multi-dimensional target optimization, and improves charging planning efficiency.

Description

Technical Field

[0001] This invention belongs to the field of electric vehicle charging planning technology, specifically relating to a user charging planning system and method. Background Technology

[0002] With the widespread application of electric vehicles, charging planning technology faces certain technical obstacles, mainly in the following aspects: Information silos are a prominent problem, characterized by the fragmentation of key data sources such as vehicle operating status, real-time availability of charging piles, power grid load fluctuations, and user travel arrangements. The lack of a unified data fusion and analysis framework results in a weak and fragmented information foundation upon which charging planning relies, making it difficult to support comprehensive and accurate decision-making.

[0003] Limitations of static planning: Existing solutions mostly rely on historical data or fixed parameters for offline calculations, which cannot dynamically respond to real-time environmental disturbances such as changes in traffic congestion, sudden changes in charging pile status, and time-of-use electricity price adjustments. This causes the planning results to quickly become invalid in actual implementation, forcing users to interrupt their trips and intervene manually.

[0004] The problem of singular optimization objectives: Most technologies focus only on the single dimension of shortest distance or lowest electricity cost, failing to take into account multiple dimensions such as time consumption and user personalized preferences, resulting in a poor user experience.

[0005] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of the aforementioned background technology and provide a user charging planning system and method that can achieve multi-source data fusion analysis and support multi-dimensional target optimization, thereby improving charging planning efficiency.

[0007] The technical solution adopted in this invention is: a user charging planning system and method. The data acquisition module is used to acquire multi-source data from multiple source interfaces; The cloud-based intelligent planning module is used to perform three-level optimization processing based on multi-source data. The three-level optimization processing includes: Level 1 long-term forecasting planning is used to predict future travel charging demand windows and minute-level energy consumption sequences based on multi-source data; Secondary short-term fine-grained planning is used to generate multiple candidate planning schemes that include charging stations, charging time, and charging cost when the charging demand window is approaching, by combining multi-source data and minute-level energy consumption sequences with multi-objective optimization algorithms. The three-level real-time dynamic adjustment is used to monitor the state deviation during the charging planning control process based on any candidate planning scheme, and to trigger the replanning algorithm to generate an adjustment scheme based on the state deviation. The charging execution module is used to perform charging planning control based on the candidate planning scheme or the adjustment scheme.

[0008] Furthermore, the multi-source data includes at least vehicle battery data, vehicle driving data, user trip data, charging pile status data, power grid load data, and traffic information data.

[0009] Furthermore, the charging demand window and minute-level energy consumption sequence for predicting future trips based on multi-source data include: Extract user historical trip data, trip destinations and times marked in the trip calendar, weather data, road type data, and user driving behavior scores from multi-source data; The extracted data is input into a time-series neural network model based on an attention mechanism to obtain minute-level energy consumption sequences for future trips and predicted SOC values ​​at the trip's end. The predicted SOC value at the end of the trip is compared with a preset safe power threshold. If the predicted SOC value at the end of the trip is lower than the safe power threshold, then mark the charging demand window; The charging demand window is defined as the length of time that covers the required charging amount, starting from the departure time of the trip marked in the trip calendar and extending backwards to the end point.

[0010] Furthermore, the step of combining multi-source data and minute-level energy consumption sequences to run a multi-objective optimization algorithm to generate multiple candidate planning schemes includes: Construct an objective function that includes electricity costs, time costs, battery depreciation costs, and grid stress costs; Constraints are determined based on the multi-source data and minute-level energy consumption sequences. An improved constrained multi-objective particle swarm optimization algorithm is used to solve the objective function under the given constraints to obtain the Pareto optimal solution set. Multiple candidate planning schemes are generated based on the Pareto optimal solution set. The candidate planning schemes include the location of the charging station, the charging start time, and the charging power curve.

[0011] Furthermore, the objective function is constructed as a multi-objective optimization problem, expressed as: min F(x)=(C_elec(x),T_total(x),D_bat(x),G_grid(x)); Wherein, F(x) is a vector objective containing multiple scalar cost components; x is a decision variable including charging station selection, charging start time, and charging power curve; min F(x) represents minimizing the vector objective F(x) in the Pareto optimal sense, and the feasible optimal solutions constitute the Pareto optimal solution set; C_elec(x) is the electricity cost, T_total(x) is the total time cost, D_bat(x) is the battery loss cost calculated based on the battery degradation model, and G_grid(x) is the grid pressure cost calculated based on grid load forecast.

[0012] Furthermore, the constraints include: Battery SOC safety constraint: Based on the minute-level energy consumption sequence, calculate the estimated SOC value at each time point along the journey, and eliminate candidate solutions that would cause the SOC to fall below the preset safety lower limit; Trip time window constraint: restricts the charging start time and charging duration to fall within the charging demand window; Physical performance constraints of charging piles: The charging power curve is limited to the maximum output power of the charging pile and the maximum power that the vehicle battery can accept.

[0013] Furthermore, the optimization algorithm introduces dynamic inertia weights to balance global search and local development capabilities, and employs a file maintenance mechanism based on congestion distance to screen non-dominated solutions.

[0014] Furthermore, the monitoring of state deviations, based on the state deviation triggering a replanning algorithm to generate an adjustment scheme, includes: The system acquires data reflecting the actual operating status from the multi-source data in real time and calculates the deviation between the data and the expected value of the current execution plan. When the deviation exceeds a preset threshold, an abnormal event is determined to have occurred, and the following replanning algorithm is triggered: Perform rolling time-domain optimization, starting from the current moment, fix the immutable parts that have already occurred, and quickly re-optimize the journey in the future short time domain to generate a preliminary adjustment plan; When the preliminary adjustment plan does not involve multi-user competition for scarce charging pile resources, the preliminary adjustment plan will be determined as the final adjustment plan.

[0015] Furthermore, when the preliminary adjustment plan involves multi-user competition for scarce charging pile resources, the following steps are performed: In response to the detection of competition among multiple users for the scarce charging pile resources, a coordination algorithm based on an auction mechanism is initiated. Obtain bidding parameters and extract user priority, urgency of demand, and willingness to pay for each competing user; Calculate the allocation weights based on the preset bidding rules and the bidding parameters to calculate the resource allocation weights for each user. Perform resource allocation, lock the scarce charging pile resources to high-weight users based on the resource allocation weight, and push alternative adjustment schemes to low-weight users to replace the preliminary adjustment scheme as the final adjustment scheme.

[0016] Furthermore, the execution of charging planning control includes: executing charging pile reservation, navigation route update, and charging start control.

[0017] Furthermore, the charging execution module is also used to achieve automatic identity authentication, automatic charging start, and contactless payment settlement after the vehicle arrives at the charging pile through the vehicle-to-pile communication protocol, forming a closed-loop process of planning-reservation-navigation-charging-payment.

[0018] Furthermore, it also includes a feedback learning module, which is used to acquire actual data and planning data of the corresponding planning stage during the execution of charging planning control, calculate difference data based on actual data and planning data, and optimize the configuration parameters of the cloud intelligent planning module according to the difference data to improve the accuracy of user charging planning. The configuration parameters include at least the fusion weights of each parameter input into the temporal neural network model and the parameters of each cost component in the objective function.

[0019] A user charging planning method based on the above-mentioned user charging planning system includes the following steps: Acquire multi-source data; Based on the multi-source data, the following three-level optimization process is performed: Level 1 long-term forecasting planning is used to predict future travel charging demand windows and minute-level energy consumption sequences based on multi-source data; Secondary short-term fine-grained planning is used to generate multiple candidate planning schemes by combining multi-source data and minute-level energy consumption sequences when the charging demand window is approaching. The three-level real-time dynamic adjustment is used to monitor the state deviation during the charging planning control process based on any candidate planning scheme, and to trigger the replanning algorithm to generate an adjustment scheme based on the state deviation. Based on the candidate planning scheme or the adjustment scheme, charging planning control is performed.

[0020] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the user charging planning method as described above.

[0021] A computer program product includes a computer program that, when run, implements the user charging planning method as described above.

[0022] The beneficial effects of this invention are as follows: This invention's data acquisition module obtains multi-source heterogeneous data from multiple interfaces, including vehicle battery data, driving behavior data, trip plans, charging pile status data, grid load data, and traffic information. This provides a real-time and comprehensive input foundation for cloud-based planning, ensuring the accuracy and timeliness of subsequent charging planning decisions. The first-level long-term predictive planning, based on multi-source data, outputs future trip charging demand windows and minute-level energy consumption sequences through time-series analysis, pre-determining charging time boundaries and energy gaps, achieving forward-looking demand prediction. The second-level short-term refined planning, as the charging demand window approaches, combines multi-source data and minute-level energy consumption sequences to run a multi-objective optimization algorithm, generating multiple candidate solutions. This ensures that the solution meets user preferences and is physically feasible. The three-level real-time dynamic adjustment continuously monitors state deviations during solution execution (such as actual energy consumption exceeding expectations or sudden charging pile failures). Once the deviation exceeds the threshold, a replanning algorithm is triggered to generate an adjustment plan. Through local rapid correction or resource coordination, trip interruptions are avoided, improving the reliability and fault tolerance of the trip under complex road conditions. The charging execution module automatically completes charging pile reservation, navigation route update, and charging start control based on candidate or adjusted plans, transforming cloud-based decisions into seamless physical operations. This achieves full-process automation from planning to execution, significantly improving charging efficiency and user convenience. Attached Figure Description

[0023] Figure 1 This is a system principle block diagram of the present invention. Detailed Implementation

[0024] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments to facilitate a clear understanding of the present invention, but these descriptions do not constitute a limitation on the present invention.

[0025] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0026] Furthermore, references to "one embodiment" or "some embodiments" in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, phrases such as "in one embodiment," "in some embodiments," "in other embodiments," and "in still other embodiments" appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized.

[0027] like Figure 1 As shown, this application proposes a user charging planning system, including: The data acquisition module is used to acquire multi-source data from multiple source interfaces in real time; The cloud-based intelligent planning module is used to perform three-level optimization processing based on multi-source data. This three-level optimization processing includes: Level 1 long-term forecasting planning is used to predict future travel charging demand windows and minute-level energy consumption sequences based on multi-source data; Secondary short-term fine-grained planning is used to generate multiple candidate planning schemes that include charging stations, charging time, and charging cost when the charging demand window is approaching (such as the day before or several hours before). It combines multi-source data and minute-level energy consumption sequences to run a multi-objective optimization algorithm. The three-level real-time dynamic adjustment is used to monitor the state deviation during the charging planning control process based on any candidate planning scheme, and to trigger the replanning algorithm to generate an adjustment scheme based on the state deviation. The charging execution module is used to perform charging planning control based on the candidate planning scheme or adjustment scheme.

[0028] This application aims to overcome the limitations of existing technologies, such as data fragmentation, rigid planning, single objective, and disconnected execution, by integrating multi-source data acquisition, cloud-based intelligent three-level optimization processing, and closed-loop control of charging execution. This provides a more intelligent, dynamic, and user-friendly electric vehicle charging planning solution.

[0029] Multi-source data refers to a collection of data from various sources used for charging planning. This data is obtained through multiple interfaces, including vehicle-mounted terminals, user mobile terminals, charging pile operator platforms, power grid dispatching systems, and traffic information platforms, and is regularly uploaded to the cloud. This multi-source data provides the foundational information for the system to conduct comprehensive and accurate charging planning.

[0030] Three-level optimization refers to a hierarchical and dynamic planning strategy adopted by this system, which aims to predict, plan and adjust charging demand in stages and at multiple levels.

[0031] The charging demand window refers to the time period during future trips when a vehicle will need to be charged. This window is determined based on predictions of future vehicle energy consumption and considerations of battery safety, aiming to provide a time frame for subsequent refined planning. Minute-level energy consumption sequences are data sequences formed by predicting the vehicle's energy consumption per minute during future trips; these sequences provide refined information on vehicle energy consumption.

[0032] Multi-objective optimization algorithms are mathematical optimization methods that can simultaneously consider multiple conflicting or interrelated objective functions and seek the optimal solution set. Candidate planning schemes refer to a set of selectable charging plans generated by the system, each containing specific information such as charging stations, charging time, and charging costs. These schemes provide users with diverse choices and provide a basis for subsequent real-time adjustments.

[0033] State deviation refers to the difference between the actual operating state of the vehicle and the expected value of the currently executed charging plan. Monitoring this deviation is crucial for triggering real-time dynamic adjustments by the system to ensure the adaptability of the plan. The replanning algorithm is an algorithm used by the system to quickly generate a new adjustment plan when state deviation is detected and an abnormal event is determined. This algorithm is designed to cope with unforeseen circumstances and ensure the continued effectiveness of the charging plan.

[0034] The charging planning workflow of this application begins with the user setting a travel plan or the system automatically identifying potential travel demand based on historical patterns. At this point, the data acquisition module gathers data from multiple sources, and the cloud-based intelligent planning module then initiates a first-level prediction, outputting energy consumption curves and a safe power window with a time granularity of minutes. Subsequently, it enters the second-level optimization stage. Under the constraints of the energy consumption curve and the window, the algorithm weighs electricity costs, time costs, battery degradation, and grid load, generating several differentiated candidate solutions, such as economic priority, time priority, and balanced solutions. After user confirmation, the system enters the execution monitoring stage, with a third-level dynamic adjustment mechanism on standby. In case of sudden congestion or charging station malfunctions, the system uses rolling time-domain optimization to quickly generate adjustment solutions or coordinates multi-vehicle resource competition through an auction mechanism. Finally, the charging execution module completes the closed-loop operation from reservation to seamless payment.

[0035] The user charging planning system of this application realizes a complete closed loop from macro demand prediction to micro real-time control by constructing a layered architecture of cloud-edge-device collaboration. That is, it can make predictions, fine planning and real-time adjustments based on real-time dynamic information, thereby providing a highly adaptable charging planning service with a consistent user experience.

[0036] In one embodiment, this application further clarifies that the multi-source data includes at least vehicle battery data, vehicle driving data, user trip data, charging pile status data, power grid load data, and traffic information data.

[0037] Specifically, the vehicle battery data refers to the real-time status information of the electric vehicle battery, such as the current state of charge (SOC), state of health (SOH), battery temperature, maximum charge / discharge power, and battery cycle life. The vehicle driving data includes information such as vehicle speed, acceleration, braking frequency, mileage, average energy consumption, and driving mode (e.g., economy mode, sport mode). This data can be used to analyze user driving habits and predict future energy consumption. The user trip data involves the user's historical travel records, frequently used routes, destination preferences, travel time patterns, and future trip information preset in the user's trip calendar, such as destination, departure time, and estimated arrival time. The charging pile status data includes the real-time availability status of charging piles (idle, occupied, faulty), charging pile type (fast charging, slow charging), maximum charging power, charging cost, geographical location, and historical usage rate. The power grid load data refers to the real-time load situation of the power grid in a specific area or time period, load forecast data, and electricity price information (e.g., peak-valley electricity prices), used to assess the impact of charging behavior on the power grid and optimize charging costs. The traffic information data includes real-time traffic conditions (congestion, smooth traffic), traffic accidents, road construction, and weather conditions (such as temperature, rainfall, and wind speed). This data is used to predict travel time, energy consumption changes, and charging needs.

[0038] Through the above technical solution, this application clarifies the specific composition of multi-source data, enabling the cloud-based intelligent planning module to obtain more comprehensive, refined, and real-time input information. By introducing this specific multi-source data, the prediction accuracy, optimization effect, and adaptability of the user's charging planning system are improved, making the generated charging plan more intelligent, efficient, and practical.

[0039] In one embodiment, this application further proposes the charging demand window and minute-level energy consumption sequence based on multi-source data to predict future trips, including: Extract user historical trip data, trip destinations and times marked in the trip calendar, weather data, road type data, and user driving behavior scores from multi-source data; The extracted data is input into a time-series neural network model based on an attention mechanism to obtain minute-level energy consumption sequences for future trips and predicted SOC values ​​at the trip's end. The predicted SOC value at the end of the trip is compared with a preset safe power threshold. If the predicted SOC value at the end of the trip is lower than the safe power threshold, then mark the charging demand window; The charging demand window is defined as the length of time that covers the required charging amount, starting from the departure time of the trip marked in the trip calendar and extending backwards to the end point.

[0040] The process includes: extracting historical trip data (retrieving similar origin-destination trip records completed by the user within a past period, including statistical features such as mileage, time, and average energy consumption); extracting trip destinations and timestamps from the trip calendar (analyzing synchronized calendar events to identify entries containing geolocation tags and specific timestamps as trigger anchors for the current plan); extracting weather data (calling meteorological service interfaces to obtain meteorological elements such as temperature, precipitation probability, wind speed, and visibility along the route); extracting road type data (distinguishing the proportion of roads of different levels, such as highways, urban roads, and mountain roads, based on navigation route matching results); and extracting a user driving behavior score (a quantitative score representing the user's driving intensity, derived from historical driving data using clustering algorithms).

[0041] Attention-based temporal neural network models are a deep learning architecture that assigns different attention weights to different input features. For example, when faced with the combined effects of heavy rain and highway driving, the model automatically increases the weight allocation to these two factors that lead to high energy consumption. The predicted SOC at the end of the trip is the percentage of battery charge remaining after reaching the destination, as output by the model. The safe charge threshold is a preset conservative value, such as 20%, to prevent over-discharge of the battery. The time length for backtracking to complete the required charging amount refers to the time buffer allowed from the departure time to charge to a set charge level (such as full charge or enough charge to travel a certain distance).

[0042] Specifically, the execution process of the aforementioned first-level long-term predictive planning is as follows: First, the user's travel intention is located. Then, historical travel characteristics, current vehicle status, external environmental parameters, etc. are input into the trained attention mechanism temporal neural network. The model outputs a minute-level energy consumption curve that changes over time. The final SOC is obtained by integrating along this curve. If the value is lower than the safety threshold, the system determines that there is a charging demand. Using the departure time as the endpoint, the latest time window that must start charging is calculated backward. For example, if 30 kWh of charging is required and fast charging takes 30 minutes, the window is from 30 minutes before departure to the departure time.

[0043] Through the aforementioned technical solution, this application can accurately extract key information closely related to vehicle energy consumption and user trips from massive amounts of multi-source data. It then utilizes a temporal neural network model based on an attention mechanism to make high-precision predictions of minute-level energy consumption sequences and the state of charge (SOC) at the trip's end. This refined predictive capability, combined with the determination of safe energy thresholds, effectively avoids delayed or excessive charging planning due to inaccurate predictions. Simultaneously, by defining the charging demand window back to the trip's departure time, it ensures that charging activities are completed in a timely manner before the trip begins. This provides users with more accurate and reliable long-term charging demand warnings, improving the predictability and effectiveness of user charging planning and providing clear optimization boundaries for subsequent refined scheduling.

[0044] In one embodiment, this application further proposes to combine multi-source data and minute-level energy consumption sequences to run a multi-objective optimization algorithm to generate multiple candidate planning schemes, including: Construct an objective function that includes electricity costs, time costs, battery depreciation costs, and grid stress costs; Constraints are determined based on the multi-source data and minute-level energy consumption sequences. An improved constrained multi-objective particle swarm optimization algorithm is used to solve the objective function under the given constraints to obtain the Pareto optimal solution set. Multiple candidate planning schemes are generated based on the Pareto optimal solution set. The candidate planning schemes include the location of the charging station, the charging start time, and the charging power curve.

[0045] Specifically, the core of the multi-objective optimization algorithm lies in constructing an objective function that incorporates electricity costs, time costs, battery degradation costs, and grid stress costs. This function quantifies the various objectives to be optimized and aims to comprehensively consider the multi-dimensional needs of users during the charging process. Electricity costs refer to the fees paid by users for charging, typically related to the charging amount and electricity prices (which may vary with time and location). Time costs encompass the total time spent by users during charging, including travel time to the charging station, queuing time, and actual charging time. Battery degradation costs are an indicator of the impact of charging behavior on battery life; for example, excessively high charging power or charging at extreme temperatures may accelerate battery aging. Grid stress costs reflect the impact of charging behavior on grid load, aiming to guide users to charge when grid load is low to support stable grid operation. By incorporating these interrelated and potentially conflicting costs into the objective function, the optimization algorithm can weigh the trade-offs between different objectives, thereby generating a more comprehensive optimization solution.

[0046] The constraints determined based on the multi-source data and minute-level energy consumption sequences are restrictions that must be satisfied in the optimization problem, ensuring that the generated candidate planning schemes are feasible and safe in practical applications. These constraints collectively define the search space of the optimization algorithm, ensuring the practicality of the final solution.

[0047] An improved constrained multi-objective particle swarm optimization algorithm is employed to solve the objective function while satisfying the aforementioned constraints. Particle swarm optimization (PSO) is a swarm intelligence-based optimization algorithm that excels at finding optimal solutions in complex search spaces. The improved constrained multi-objective PSO algorithm is used to more effectively handle multi-objective optimization problems and complex constraints. Multi-objective means the algorithm needs to optimize multiple objectives simultaneously (such as the four costs mentioned above), rather than a single objective. Constraints mean the algorithm strictly adheres to pre-defined constraints during the search process, avoiding the generation of infeasible solutions. The improved PSO algorithm simulates the foraging behavior of a flock of birds, iteratively adjusting the position and velocity of particles (i.e., candidate solutions) to gradually approach the optimal solution set.

[0048] After the optimization algorithm solves the objective function and satisfies the constraints, it obtains a Pareto optimal solution set, which represents a set of solutions that cannot be further improved on any of the four cost objectives. Multiple candidate planning schemes are generated based on this Pareto optimal solution set. This involves selecting several evenly distributed solutions from the set and transforming them into multiple candidate planning schemes (i.e., forming differentiated charging strategies) that include charging station locations, charging times, and charging power curves, allowing users to choose according to their preferences. Each candidate scheme represents a trade-off between different objectives. For example, one scheme might sacrifice lower electricity costs for longer charging times; another scheme might offer faster charging speeds but higher electricity costs. Each candidate planning scheme includes detailed information such as the specific location of the charging station, the suggested charging start time, and a dynamic charging power curve. The charging power curve refers to the strategy of changing charging power over time during the charging process. This allows the system to dynamically adjust the charging rate based on battery status, grid load, or user demand, thereby achieving more refined charging management. Providing multiple candidate schemes allows users to choose according to their preferences and real-time conditions, increasing the flexibility and adaptability of the planning.

[0049] Through the above technical solution, this application overcomes the limitations of traditional charging planning, which suffers from a single objective and insufficient consideration of constraints. By constructing a comprehensive objective function that includes electricity costs, time costs, battery wear costs, and grid pressure costs, the system can comprehensively evaluate the merits of charging schemes and guide the optimization algorithm to intelligently balance economic efficiency, battery health, and grid friendliness. Constraints determined based on multi-source data and minute-level energy consumption sequences ensure that all generated candidate schemes meet the safety and feasibility requirements of actual operation, avoiding unrealistic planning. An improved constrained multi-objective particle swarm optimization algorithm efficiently finds a diverse set of high-quality non-dominated solutions—multiple candidate planning schemes—within a complex search space. These candidate schemes not only include charging station locations and charging start times but also provide refined charging power curves, allowing the charging process to be dynamically adjusted according to actual conditions, thus achieving more flexible and efficient charging management. Ultimately, users can choose the scheme that best meets their current needs from these multi-dimensional optimized candidate schemes, improving the intelligence level of charging planning and user satisfaction. This application constructs a multi-objective optimization framework and solves the Pareto optimal solution set, which can generate a set of differentiated solutions that achieve optimal balance in all four cost dimensions while preserving the independence of each cost objective, thereby improving the flexibility and user adaptability of charging planning.

[0050] In one embodiment, this application further proposes that the objective function be constructed as a multi-objective optimization problem, expressed as: min F(x)=(C_elec(x),T_total(x),D_bat(x),G_grid(x)); Wherein, F(x) is a vector objective containing multiple scalar cost components; x is a decision variable including charging station selection, charging start time, and charging power curve; min F(x) represents minimizing the vector objective F(x) in the Pareto optimal sense, and the feasible optimal solutions constitute the Pareto optimal solution set; C_elec(x) is the electricity cost, T_total(x) is the total time cost, D_bat(x) is the battery loss cost calculated based on the battery degradation model, and G_grid(x) is the grid pressure cost calculated based on grid load forecast.

[0051] Specifically, constructing the objective function as a multi-objective optimization problem means that electricity cost, time cost, battery loss cost, and grid pressure cost are treated as four components of the vector objective, rather than a single scalar, in order to preserve the independence of each objective. The mathematical meaning of the objective function is as follows: The decision variable x represents a specific charging scheme, including the selection of charging stations, the charging start time, and the charging power curve. The range of values ​​is limited by the physical constraints and market rules of the charging plan, and can be expressed as x=(s,t_s,P(t)), where s is the charging station identifier, t_s is the charging start time, and P(t) is the charging power curve. The selection of charging stations determines the spatial attributes of the charging location and the electricity price system. The charging start time determines the temporal attributes of the charging period and the grid load status. The charging power curve determines the power distribution and duration of the charging process. The three together determine the cost and efficiency performance of the final scheme. The vector objective F(x) consists of four components, each of which is a function of x. min means minimizing F(x) in the Pareto optimal sense, that is, finding a set of solutions such that none of these solutions can improve another cost component without worsening at least one cost component. This set of solutions is the Pareto optimal solution set, which represents the best trade-off among the four cost objectives.

[0052] Electricity cost refers to the fee a user pays to complete charging. Its calculation is typically based on the charging station's real-time electricity price, time-of-use pricing strategy, service fees, and the required charging capacity. Total time cost encompasses the time a user spends throughout the charging process, including but not limited to travel time to the charging station, waiting time at the station, and the actual charging time. This cost aims to quantify the user's time value and the inconvenience of the charging process. Battery degradation cost, calculated based on battery degradation models, measures the impact of charging behavior on vehicle battery life. This cost calculation relies on advanced battery degradation models (e.g., Arrhenius-based lifespan decay models or commonly used semi-empirical aging models), which consider factors such as charging current, battery temperature during charging, battery state of charge (SOC) range, and number of charging cycles. Grid stress cost, calculated based on grid load forecasts, reflects the impact of charging behavior on grid stability and load balance. This cost is calculated based on real-time or predicted grid load data. Each cost component can be conventionally calculated using well-known charging billing rules, route navigation and queuing prediction algorithms, lithium-ion battery degradation models, and regional grid load assessment methods.

[0053] The solution process for the aforementioned multi-objective optimization problem begins with the initialization of the decision variable x. Based on the charging demand window and minute-level energy consumption sequence obtained from the first-level long-term forecast, the system generates an initial set of decision variables containing multiple candidate charging stations, charging times, and power curves. Subsequently, each decision variable x is substituted into the four component functions of the vector objective F(x) to calculate the corresponding electricity cost, total time cost, battery loss cost, and grid pressure cost. Then, an improved constrained multi-objective particle swarm optimization algorithm is used to iteratively update and sort all decision variables under the constraints of battery safety, time window, and charging pile physical performance. Pareto optimal solution sets are gradually selected, and finally, multiple uniformly distributed solutions are extracted from the solution set and transformed into candidate planning schemes containing charging station locations, charging start times, and charging power curves.

[0054] This application defines the objective function as a multi-objective vector optimization problem, which enables the optimization process to objectively reflect the inherent conflict between various cost objectives. The generated Pareto optimal solution set provides users with a truly multi-dimensional trade-off choice, greatly improving the objectivity and scientific nature of charging planning.

[0055] In one embodiment, this application further proposes the following constraints: Battery SOC safety constraint: Based on the minute-level energy consumption sequence, calculate the estimated SOC value at each time point along the journey, and eliminate candidate solutions that would cause the SOC to fall below the preset safety lower limit; Trip time window constraint: restricts the charging start time and charging duration to fall within the charging demand window; Physical performance constraints of charging piles: The charging power curve is limited to the maximum output power of the charging pile and the maximum power that the vehicle battery can accept.

[0056] Specifically, the battery SOC safety constraint aims to ensure that the vehicle's battery charge (SOC) remains above a safe threshold throughout the entire trip, regardless of the charging plan. The system utilizes minute-level energy consumption sequences obtained during the first-level long-term prediction planning phase to simulate the vehicle's energy consumption under different candidate plans, thereby predicting the battery SOC at various points in time during the trip. If a candidate plan is found to cause the battery SOC to fall below a preset safety threshold (e.g., below 10% or 20%), that plan will be immediately eliminated to prevent the vehicle from breaking down due to depleted battery power, ensuring driving safety.

[0057] The travel time window constraint ensures that charging activities align with users' actual travel schedules and time preferences. During the long-term forecasting and planning phase, the system identifies the charging demand window for users' future trips. This constraint requires that the suggested charging start time and duration in all generated candidate plans must be fully included within this charging demand window. This avoids the system proposing charging solutions at times when it is inconvenient or unnecessary for the user, thus improving the practicality of the planning and user acceptance.

[0058] The physical performance constraints of the charging stations are designed to ensure the physical feasibility and equipment safety of the charging process. Each charging station has a fixed maximum output power, and the vehicle's Battery Management System (BMS) dynamically limits its maximum acceptable charging power based on the battery's current state (such as SOC and temperature). This constraint requires that the charging power curve set in the candidate planning scheme must never exceed the smaller of the actual maximum output power of the charging station and the actual maximum acceptable power of the vehicle battery, preventing damage to the charging equipment or vehicle battery due to overload and ensuring the stability and safety of the charging process.

[0059] This application constructs a safe and reliable solution space by setting multi-level physical and logical constraints, ensuring that all output planning schemes are not only theoretically optimal, but also absolutely feasible and executable in the physical world, thereby improving the safety, reliability and practicality of charging planning.

[0060] In one embodiment, this application further proposes that the optimization algorithm introduces dynamic inertia weights to balance global search and local development capabilities, and adopts an archive maintenance mechanism based on congestion distance to screen non-dominated solutions.

[0061] Specifically, the optimization algorithm, such as an improved constrained multi-objective particle swarm optimization algorithm, introduces dynamic inertia weights during the iteration process. Inertia weight is a crucial parameter in particle swarm optimization, determining the degree to which a particle retains its current velocity in the next iteration. To balance global search and local exploitation capabilities, this inertia weight is not a fixed value but is dynamically adjusted based on the algorithm's iteration progress or the current population state. Simultaneously, to effectively manage and filter non-dominated solutions generated by the multi-objective optimization algorithm, this application employs a file maintenance mechanism based on crowding distance. In multi-objective optimization problems, a set of non-dominated solutions is typically obtained, where these solutions do not dominate each other across objectives. The file maintenance mechanism stores and manages these non-dominated solutions. Crowding distance is an indicator of the sparsity of solutions in the objective space, i.e., the distance between a solution and its neighboring solutions. When the number of non-dominated solutions stored in the file reaches a preset upper limit, some solutions need to be removed from the file. The crowding distance-based mechanism prioritizes removing solutions with smaller crowding distances, i.e., those solutions that are more densely packed in the objective space, thus retaining solutions with larger crowding distances. In this way, the mechanism can ensure that the non-dominated solution set stored in the archive has good diversity and distribution, and avoid the solution set being too concentrated in a certain area, thereby providing users or subsequent decision-making modules with more comprehensive and representative candidate planning schemes.

[0062] This application improves the solution efficiency and quality by making targeted improvements to the standard optimization algorithm, ensuring that users can obtain truly diverse and high-quality candidate planning schemes, thereby improving the quality of charging planning and user experience.

[0063] In one embodiment, this application further proposes a specific implementation method for monitoring state deviations and triggering a replanning algorithm to generate an adjustment scheme during the charging planning control process based on any candidate planning scheme. This includes: The system acquires data reflecting the actual operating status from the multi-source data in real time and calculates the deviation between the data and the expected value of the current execution plan. When the deviation exceeds a preset threshold, an abnormal event is determined to have occurred, and the following replanning algorithm is triggered: Perform rolling time-domain optimization, starting from the current moment, fix the immutable parts that have already occurred, and quickly re-optimize the journey in the future short time domain to generate a preliminary adjustment plan; When the preliminary adjustment plan does not involve multi-user competition for scarce charging pile resources, the preliminary adjustment plan will be determined as the final adjustment plan.

[0064] Specifically, to ensure the adaptability and robustness of the charging plan, the system acquires data reflecting the actual operating status from the multi-source data in real time and calculates the deviation between this data and the expected value of the current execution plan. This real-time data may include, but is not limited to, the vehicle's real-time state of charge (SOC), actual driving speed, actual energy consumption, the real-time availability of charging stations, charging power, real-time grid load data, and real-time traffic information data. The system compares this real-time acquired data with the corresponding expected value in the currently executing charging plan. For example, it compares the vehicle's current actual SOC with the expected SOC at that moment on the planned path, or it compares the actual arrival time at the charging station with the planned expected arrival time, thereby calculating the deviation between the two.

[0065] To avoid frequent and unnecessary replanning due to minor fluctuations, the system sets a preset threshold. This threshold can be configured based on historical data analysis, system stability requirements, or user experience preferences. When the real-time calculated deviation exceeds this preset threshold, the system determines that there is a significant difference between the current operating state and the planned scheme, constituting an abnormal event requiring intervention. At this time, the system will automatically trigger a subsequent replanning algorithm. This replanning algorithm adopts a rolling time-domain optimization strategy, using the current moment as the new planning starting point, fixing the parts that have already occurred and cannot be changed, such as the mileage already traveled, the electricity already consumed, or the charging stage already completed. Then, the system quickly re-optimizes the journey only within a predetermined short time domain in the future, such as the next 15 minutes, 30 minutes, or until the next expected charging point, thereby generating a preliminary adjustment plan. After generating the preliminary adjustment plan, the system will evaluate whether the plan will cause multi-user competition for scarce charging pile resources. When the preliminary adjustment plan does not involve multi-user competition for scarce charging pile resources, the preliminary adjustment plan will be directly adopted and determined as the final adjustment plan. This means that, given sufficient resources, the system can respond quickly and directly to state deviations, providing an immediate and feasible solution.

[0066] This application, by introducing real-time monitoring and rolling time-domain optimization, endows the system with powerful dynamic response capabilities, ensures the continuity of charging planning and the stability of user experience, effectively copes with the uncertainty of the external environment, and thus maintains the effectiveness and optimization of charging planning.

[0067] In one embodiment, this application further proposes that when the initial adjustment scheme involves multi-user competition for scarce charging pile resources, the system will perform a series of coordination steps.

[0068] First, in response to the detection of competition among multiple users for the scarce charging pile resources, the system initiates a coordination algorithm based on an auction mechanism. The purpose of this coordination algorithm is to provide a fair, transparent, and efficient allocation mechanism when multiple users simultaneously require limited charging pile resources. The system continuously monitors the real-time status of charging piles, reservation status, and charging requests submitted by users. When multiple users' initial adjustment plans simultaneously target the same limited set of charging pile resources marked as scarce by the system (e.g., during specific peak hours, in specific geographical locations, or at specific charging power levels, demand far exceeds supply), the system automatically triggers this auction-based coordination algorithm.

[0069] Subsequently, the system acquires bidding parameters, specifically including the user priority, urgency of need, and willingness to pay of each competing user. These bidding parameters are key indicators for assessing the intensity and willingness of users to demand scarce resources. User priority can be set based on preset rules such as user membership level, historical credit record, and subscription service type (e.g., VIP service, emergency vehicle service), reflecting the user's overall value or special needs in the system. Urgency of need is quantified by analyzing real-time data such as the user's current battery charge (SOC), remaining range, destination distance, estimated arrival time, and important events marked in the trip calendar (e.g., meetings, flights). For example, users with low battery or who are about to miss important trips have a higher urgency of need. Willingness to pay can be obtained based on the user's preset payment limit in the system, historical charging consumption habits, or additional payment willingness actively entered by the user when competition occurs. This can be a direct monetary bid or a higher charging fee or longer waiting time that the user is willing to accept.

[0070] Next, the system calculates the allocation weights based on preset bidding rules and the bidding parameters. The preset bidding rules are typically a multi-factor weighted model; for example, allocation weights can be obtained by multiplying parameters such as user priority, urgency of need, and willingness to pay by corresponding weighting coefficients and then summing them. These weighting coefficients can be dynamically adjusted based on operational strategies, market supply and demand, or user feedback.

[0071] Finally, the system performs resource allocation, locking scarce charging pile resources to high-weight users based on the resource allocation weights, and pushing alternative adjustment schemes to low-weight users to replace the initial adjustment scheme as the final adjustment scheme. For example, the system will match the user with the highest allocation weight (or the top N ranked users) with scarce charging pile resources and immediately lock the charging pile to ensure their charging plan is executed. For low-weight users who fail to obtain scarce charging pile resources, the system will not simply cancel their charging plan, but will quickly generate and push alternative adjustment schemes based on their current status and available non-scarce resources. These alternative schemes may include recommending nearby non-scarce charging piles, adjusting charging time, or suggesting that the user go to other areas to charge. The generation of alternative adjustment schemes can again utilize rolling time-domain optimization algorithms, but its goal is to find suboptimal but feasible alternatives.

[0072] This application improves user experience and overall system robustness by introducing an auction-based coordination algorithm. This mechanism not only optimizes resource utilization efficiency but also enhances users' trust in the charging planning system, especially in emergency situations, providing more flexible and user-friendly solutions.

[0073] In one embodiment, this application further proposes specific steps for performing charging planning control, including performing charging station reservation, navigation route update and charging start control.

[0074] The purpose of implementing charging station reservations is to ensure the availability of charging stations and avoid situations where users arrive at a charging station only to find no stations available. Specifically, the system can communicate with the charging station operation platform or charging service provider's interface, and after determining the target charging station and charging time, send a reservation request to that charging station to reserve a specific charging station. After a successful reservation, the system will send a confirmation message to the user and may provide a reservation code or QR code for user authentication during charging.

[0075] Navigation route updates provide users with real-time, accurate navigation guidance to their target charging station. The system sends the selected charging station location information to the vehicle's built-in navigation system or the navigation application on the user's mobile device. The navigation system calculates and displays the optimal driving route based on real-time traffic conditions, road information, and other factors. If new adjustments are generated during the charging planning and control process due to real-time dynamic adjustments, the navigation route will be updated accordingly to guide the user to the new target charging station or the adjusted charging location.

[0076] Charging start control aims to automatically or semi-automatically initiate the charging process upon the user's arrival at the charging station, enhancing convenience. This control can be implemented in various ways. For example, the system can interact with the charging station through a communication interface (such as a vehicle-to-charging station communication protocol) to complete necessary authentication and send charging commands. Furthermore, the system will control the charging station to output the corresponding power according to the preset charging power curve in the planning scheme, ensuring that the charging process is executed according to the predetermined plan.

[0077] Through the above technical solutions, this application ensures the smooth implementation of the charging plan, improves the convenience and efficiency of user charging, and thus optimizes the overall user charging experience.

[0078] In one embodiment, this application further proposes that the charging execution module is also used to achieve automatic identity authentication, automatic charging start, and contactless payment settlement after the vehicle arrives at the charging pile via the vehicle-to-pile communication protocol. The execution results of the automatic identity authentication, automatic charging start, and contactless payment settlement are fed back to the cloud-based intelligent planning module to optimize the charging planning for the subsequent trip, forming a closed loop of the entire process from planning, reservation, navigation, charging to payment.

[0079] Specifically, the charging execution module can automatically trigger a series of operations after the vehicle arrives at the charging station. After the vehicle arrives at the charging station, the system can automatically sense that the vehicle has arrived at the predetermined charging location through vehicle positioning modules (such as GPS), visual recognition systems, or near-field communication modules on the charging station side (such as Bluetooth, RFID), and send an arrival signal to the charging execution module.

[0080] The vehicle-to-charging-pile communication protocol is a standard specification for data exchange and command transmission between vehicles and charging piles. This protocol allows for the establishment of a secure communication link between vehicles and charging piles to transmit data such as vehicle charging needs, battery status, user identity information, and charging pile availability and power output. Through this protocol, intelligent interaction and collaborative control between vehicles and charging piles can be achieved.

[0081] The automatic identity authentication refers to the system automatically identifying the user's identity and verifying their charging permissions after the vehicle establishes communication with the charging station, without requiring manual operation from the user. The automatic charging start means that after successful identity authentication, the charging station automatically begins charging based on system presets or vehicle requests, without requiring the user to manually press a start button or swipe a card. The seamless payment settlement after charging means that after charging is completed, the system automatically calculates and deducts the fee, without requiring any payment operation from the user.

[0082] Through the aforementioned technical solution, after the vehicle arrives at the charging station, the charging execution module can utilize the vehicle-to-charging station communication protocol to achieve automatic identity authentication, automatic charging start, and seamless payment settlement after charging. This reduces manual operation by the user during the charging process and avoids delays or errors that may be caused by human intervention. It tightly integrates planning, reservation, navigation, charging, and payment into a highly automated, seamless closed-loop execution process. This automated and seamless experience greatly improves the convenience and efficiency of user charging, making the entire charging process smoother and more worry-free, and further optimizing the overall efficiency of the user charging planning system.

[0083] In one embodiment, this application further proposes a user charging planning system, which includes a feedback learning module. This module continuously collects and analyzes system operation data to achieve adaptive optimization of the planning model and parameters. Specifically, the feedback learning module acquires actual data during the charging planning control process and planning data for the corresponding planning stage. Based on the actual data and planning data, it calculates the difference data and optimizes the configuration parameters of the cloud-based intelligent planning module according to the difference data. This ensures that the planning system can continuously improve its performance over time and with environmental changes, thereby improving the accuracy of user charging planning. The configuration parameters include at least the fusion weights of each parameter input to the temporal neural network model and the parameters of each cost component in the objective function.

[0084] Specifically, actual data refers to the real-time or near-real-time operational data acquired by the charging execution module during the charging planning and control process through vehicle sensors, charging pile communication interfaces, user feedback, etc., such as actual charging amount, actual charging time, actual energy consumption, actual driving route, and actual electricity cost. Planning data refers to the predicted and set values ​​used by the cloud-based intelligent planning module when generating candidate planning schemes or adjusting schemes, such as predicted energy consumption, estimated charging time, estimated charging cost, and planned charging power curve.

[0085] Discrepancy data refers to the deviation or inconsistency between actual and planned data. Methods for calculating discrepancy data include, but are not limited to, direct subtraction, calculating relative error, and percentage deviation. This discrepancy data can intuitively reflect the predictive accuracy of the planning model in a specific scenario and the effectiveness of the planning scheme's implementation.

[0086] Optimizing configuration parameters refers to adjusting the model parameters within the cloud-based intelligent planning module using calculated discrepancies and specific optimization algorithms (such as gradient descent, reinforcement learning, and genetic algorithms). For example, when a persistent deviation is found between predicted and actual energy consumption, the parameters related to energy consumption prediction in the time-series neural network model can be adjusted. This process aims to reduce future planning deviations and make the planning results closer to actual needs and operational conditions.

[0087] This application endows the system with self-learning capabilities through a closed-loop feedback learning mechanism, enabling the system to better adapt to dynamically changing external environments and user behaviors. The planning accuracy continuously improves over time, ensuring robustness and efficiency in long-term operation.

[0088] In one embodiment, this application proposes a user charging planning method based on the above-described user charging planning system, the method comprising the following steps: Acquire multi-source data; Based on the multi-source data, the following three-level optimization process is performed: Level 1 long-term forecasting planning is used to predict future travel charging demand windows and minute-level energy consumption sequences based on multi-source data; Secondary short-term fine-grained planning is used to generate multiple candidate planning schemes by combining multi-source data and minute-level energy consumption sequences when the charging demand window is approaching. The three-level real-time dynamic adjustment is used to monitor the state deviation during the charging planning control process based on any candidate planning scheme, and to trigger the replanning algorithm to generate an adjustment scheme based on the state deviation. Based on the candidate planning scheme or the adjustment scheme, charging planning control is performed.

[0089] This application overcomes the problems of one-sided planning basis, static planning's inability to adapt to real-time dynamic factors, single optimization goal, and disconnect between planning and execution caused by information silos by combining first-level long-term predictive planning, second-level short-term detailed planning, and third-level real-time dynamic adjustment in a hierarchical and dynamic manner. It achieves the effect of providing adaptive, multi-objective collaborative and closed-loop execution charging planning services, realizes the intelligentization of the entire process from data collection to planning execution, and improves the accuracy, adaptability and user experience of charging planning.

[0090] The user charging planning method of this application will be explained in detail below with specific examples: User Wang, whose vehicle is a pure electric SUV (battery capacity 80kWh, current SOC 50%), plans to drive from point A to point B over the weekend (total distance approximately 300 kilometers) and hopes to arrive before 16:00 in the afternoon.

[0091] Step 1: The night before the trip, the system executes long-term forecasts and preliminary planning.

[0092] The system synchronized Wang's itinerary with his calendar and, combined with the weather forecast (sunny the next day, light breeze), predicted a base energy consumption of approximately 55 kWh for the entire trip, with a state of charge (SOC) of approximately 18% upon arrival. Given the low SOC upon arrival and the user's preference for battery maintenance, the system initially determined that a recharge was necessary during the journey.

[0093] Step 2: On the morning of the travel day, the system executes a short-term, detailed plan.

[0094] Acquire real-time data: charging station status at service areas along the route, electricity price at 10:00 AM (average section), and real-time traffic conditions (smooth). Run a multi-objective optimization algorithm to generate two recommended solutions: Option A (Time Priority): Charge the battery at high power for 20 minutes at the midpoint of the journey (Service Area C) to 65%, which is expected to increase the travel time by 25 minutes. The total time is about 4 hours, the electricity cost is 35 yuan, and the battery degradation is moderate.

[0095] Option B (Cost and Battery Maintenance Prioritized): Charge the battery to 90% at a nearby partner charging station (off-peak electricity rates) before departure, eliminating the need for charging during the trip. The total trip time is approximately 3.5 hours (due to high battery levels at departure), with an electricity cost of only NT$15 (utilizing off-peak electricity), resulting in minimal battery degradation.

[0096] Mr. Wang chose Option B, and the system automatically reserved a charging station near his home between 00:30 and 02:00, and set the vehicle to automatically complete charging within this time period.

[0097] Step 3: During the trip, the system monitors and dynamically adjusts in real time.

[0098] After the vehicle departed, a sudden traffic accident caused highway congestion, resulting in an actual energy consumption increase exceeding predictions. The system monitored in real-time and determined that, based on the current trend, the State of Charge (SOC) would drop below 10% upon arrival at point B, triggering an alarm and initiating a replanning process. A rolling time-domain optimization algorithm quickly calculated and recommended a 10-minute emergency fast charge at the next exit (Service Area D), switching the navigation target to that charging station. The system pre-emptively located and locked onto an upcoming vacant charging station at that service area via V2X. Upon arrival, the vehicle automatically parked in the charging space, and the onboard terminal sent a start command to the charging station via Bluetooth to begin charging.

[0099] Step 4: Charging complete, closed loop ends.

[0100] The charging automatically stopped after reaching the preset target (SOC 65%), and the cloud-based seamless payment was completed. The system updated the trip's energy consumption data to optimize Mr. Wang's personal model. Navigation continued to guide him to point B, and he ultimately arrived ahead of schedule.

[0101] It should be noted that the execution process of the above methods and steps is based on the same concept as the system embodiments of this application. For details on the specific implementation process, functions, and technical effects, please refer to the system embodiments section, and they will not be repeated here. The descriptions of each embodiment have their own emphasis; parts not detailed or recorded in a certain embodiment can be referred to the relevant descriptions of other embodiments.

[0102] It is understood that those skilled in the art will clearly recognize that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application.

[0103] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps in the above-described method embodiments.

[0104] This application provides a computer program product, including a computer program, which, when run on an electronic device, enables the electronic device to implement the steps described in the various method embodiments above.

[0105] In applications, computer programs include computer program code, which can be in the form of source code, object code, executable files, or some intermediate form. Computer-readable media can include at least: any entity or device capable of carrying computer program code to an electronic device, recording media, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media. Examples include USB flash drives, portable hard drives, magnetic disks, or optical discs. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electrical carrier signals or telecommunication signals.

[0106] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0107] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application. Content not described in detail in this specification belongs to the prior art known to those skilled in the art.

Claims

1. A user charging planning system, characterized in that, include: The data acquisition module is used to acquire data from multiple sources; The cloud-based intelligent planning module is used to perform three-level optimization processing based on multi-source data. The three-level optimization processing includes: Level 1 long-term forecasting planning is used to predict future travel charging demand windows and minute-level energy consumption sequences based on multi-source data; Second-level short-term fine-grained planning is used to generate multiple candidate planning schemes by combining multi-source data and minute-level energy consumption sequences when the charging demand window is approaching; The three-level real-time dynamic adjustment is used to monitor the state deviation during the charging planning control process based on any candidate planning scheme, and to trigger the replanning algorithm to generate an adjustment scheme based on the state deviation. The charging execution module is used to perform charging planning control based on the candidate planning scheme or the adjustment scheme.

2. The user charging planning system according to claim 1, characterized in that: The multi-source data includes at least vehicle battery data, vehicle driving data, user trip data, charging pile status data, power grid load data, and traffic information data.

3. The user charging planning system according to claim 1, characterized in that, The charging demand window and minute-level energy consumption sequence based on multi-source data for predicting future trips include: Extract user historical trip data, trip destinations and times marked in the trip calendar, weather data, road type data, and user driving behavior scores from multi-source data; The extracted data is input into a time-series neural network model based on an attention mechanism to obtain minute-level energy consumption sequences for future trips and predicted SOC values ​​at the trip's end. The predicted SOC value at the end of the trip is compared with a preset safe power threshold. If the predicted SOC value at the end of the trip is lower than the safe power threshold, then mark the charging demand window; The charging demand window is defined as the length of time that covers the required charging amount, starting from the departure time of the trip marked in the trip calendar and extending backwards to the end point.

4. The user charging planning system according to claim 1, characterized in that, The process of combining multi-source data and minute-level energy consumption sequences to run a multi-objective optimization algorithm generates multiple candidate planning schemes, including: Construct an objective function that includes electricity costs, time costs, battery depreciation costs, and grid stress costs; Constraints are determined based on the multi-source data and minute-level energy consumption sequences. An improved constrained multi-objective particle swarm optimization algorithm is used to solve the objective function under the premise of satisfying the constraints to obtain the Pareto optimal solution set; Multiple candidate planning schemes are generated based on the Pareto optimal solution set.

5. The user charging planning system according to claim 4, characterized in that, The objective function is constructed as a multi-objective optimization problem, expressed as: min F(x)=(C_elec(x),T_total(x),D_bat(x),G_grid(x)); Wherein, F(x) is a vector objective containing multiple scalar cost components; x is a decision variable including charging station selection, charging start time, and charging power curve; min F(x) represents minimizing the vector objective F(x) in the Pareto optimal sense, and the feasible optimal solutions constitute the Pareto optimal solution set; C_elec(x) is the electricity cost, T_total(x) is the total time cost, D_bat(x) is the battery loss cost calculated based on the battery degradation model, and G_grid(x) is the grid pressure cost calculated based on grid load forecast.

6. The user charging planning system according to claim 4, characterized in that, The constraints include: Battery SOC safety constraint: Based on the minute-level energy consumption sequence, calculate the estimated SOC value at each time point along the journey, and eliminate candidate solutions that would cause the SOC to fall below the preset safety lower limit; Trip time window constraint: restricts the charging start time and charging duration to fall within the charging demand window; Physical performance constraints of charging piles: The charging power curve is limited to the maximum output power of the charging pile and the maximum power that the vehicle battery can accept.

7. The user charging planning system according to claim 4, characterized in that: The optimization algorithm introduces dynamic inertia weights to balance global search and local development capabilities, and employs a file maintenance mechanism based on congestion distance to screen for non-dominated solutions.

8. The user charging planning system according to claim 1, characterized in that, The monitored state deviation, based on the state deviation triggering a replanning algorithm to generate an adjustment scheme, includes: The system acquires data reflecting the actual operating status from the multi-source data in real time and calculates the deviation between the data and the expected value of the current execution plan. When the deviation exceeds a preset threshold, an abnormal event is determined to have occurred, and the following replanning algorithm is triggered: Perform rolling time-domain optimization, starting from the current moment, fix the immutable parts that have already occurred, and quickly re-optimize the journey in the future short time domain to generate a preliminary adjustment plan; When the preliminary adjustment plan does not involve multi-user competition for scarce charging pile resources, the preliminary adjustment plan will be determined as the final adjustment plan.

9. The user charging planning system according to claim 8, characterized in that, When the preliminary adjustment plan involves multi-user competition for scarce charging pile resources, the following steps are performed: In response to the detection of competition among multiple users for the scarce charging pile resources, a coordination algorithm based on an auction mechanism is initiated. Obtain bidding parameters and extract user priority, urgency of demand, and willingness to pay for each competing user; Calculate the allocation weights based on the preset bidding rules and the bidding parameters to calculate the resource allocation weights for each user. Perform resource allocation, lock the scarce charging pile resources to high-weight users based on the resource allocation weight, and push alternative adjustment schemes to low-weight users to replace the preliminary adjustment scheme as the final adjustment scheme.

10. The user charging planning system according to claim 1, characterized in that: The execution of charging planning control includes: executing charging pile reservation, navigation route update and charging start control.

11. The user charging planning system according to claim 1, characterized in that: The charging execution module is also used to achieve automatic identity authentication, automatic charging start, and contactless payment settlement after the vehicle arrives at the charging station through the vehicle-to-charging station communication protocol, forming a closed-loop process of planning-reservation-navigation-charging-payment.

12. The user charging planning system according to claim 1, characterized in that: It also includes a feedback learning module, which is used to acquire actual data and planning data of the corresponding planning stage during the execution of charging planning control, calculate difference data based on actual data and planning data, and optimize the configuration parameters of the cloud intelligent planning module according to the difference data to improve the accuracy of user charging planning.

13. A user charging planning method based on the user charging planning system of claim 1, characterized in that: Acquire multi-source data; Based on the multi-source data, the following three-level optimization process is performed: Level 1 long-term forecasting planning is used to predict future travel charging demand windows and minute-level energy consumption sequences based on multi-source data; Second-level short-term fine-grained planning is used to generate multiple candidate planning schemes by combining multi-source data and minute-level energy consumption sequences when the charging demand window is approaching; The three-level real-time dynamic adjustment is used to monitor the state deviation during the charging planning control process based on any candidate planning scheme, and to trigger the replanning algorithm to generate an adjustment scheme based on the state deviation. Based on the candidate planning scheme or the adjustment scheme, charging planning control is performed.

14. A computer-readable storage medium storing a computer program thereon, characterized in that: When the computer program is executed by the processor, it implements the user charging planning method as described in claim 13.

15. A computer program product, comprising a computer program, characterized in that: When the computer program is run, it implements the user charging planning method as described in claim 13.

Images