Electric vehicle charging guiding method and system considering charging station selection habit of user
By building user portraits and satisfaction models, combining real-time data and road conditions information, selecting the charging station with the highest user satisfaction for charging guidance, solving the limitations of the charging guidance strategy based on the complete rational assumption in the existing technology, and improving the charging of electric vehicles Satisfaction and operational efficiency.
Patent Information
- Application Number
- CN202411182464.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-12-03
AI Technical Summary
Most of the existing electric vehicle charging guidance strategies are based on the user's completely rational assumptions, and they cannot accurately understand and predict the user's limited rational charging behavior, resulting in limitations in actual applications, affecting charging satisfaction and operational efficiency.
By building user portraits, using SOFM and SHAP methods to cluster the user's time cost and price cost, establish a user satisfaction model, combine real-time data and road conditions information, and select the charging station with the highest user satisfaction for charging guidance.
It improves user charging satisfaction and charging station operation efficiency, realizes quantitative evaluation and optimization of user charging experience, and promotes the popularization and sustainable development of electric vehicles.
Smart Images

Figure CN119066556A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of electric vehicle charging management, and in particular to an electric vehicle charging guidance method and system taking into account a user's habit of selecting a charging station. Background Art
[0002] Against the backdrop of the continued growth of global energy consumption and increasingly serious environmental pollution, electric vehicles (EVs) have received widespread attention and rapid development as an important means to reduce carbon emissions and promote the development of sustainable transportation. By 2023, the number of new energy vehicles in my country has reached 20.41 million, and the number of charging piles has reached 8.596 million. However, despite the increasing popularity of electric vehicles, problems such as unbalanced vehicle-pile ratio and difficulty in charging still exist, restricting the further development of the electric vehicle industry. Therefore, it has become an urgent issue to study and develop effective charging guidance strategies to solve the problem of electric vehicle travel charging, improve user charging satisfaction and the operating efficiency of charging stations.
[0003] At present, many scholars and researchers have conducted in-depth research in the field of charging guidance strategies. For example, the literature [Zhan Hua, Jiang Changxu, Su Qinglie. Electric vehicle charging guidance method based on hierarchical reinforcement learning [J]. Electric Power Automation Equipment, 2022, 42 (10): 264-272.] proposed a charging guidance strategy based on hierarchical reinforcement learning, which effectively solved the problems of EV charging path planning and real-time charging guidance. The literature [Fang Yuxuan, Hu Junjie, Ma Wenshuai. Electric vehicle aggregator master-slave game optimization scheduling strategy considering user willingness [J / OL]. Transactions of the Chinese Society of Electrical Engineering, 1-13.] In view of the uncertainty of user scheduling willingness, an optimization scheduling strategy considering user time anxiety cost and battery loss cost is proposed. Through the optimization scheduling of electric vehicle aggregators master-slave game, the overall economy of users and aggregators is increased, and the economic and safe operation of the power grid is promoted. The literature [Xie Longtao, Xie Shiwei, Chen Kaiyue, et al. Pricing strategy for charging stations in a power-transportation coupling network considering user travel cost budget [J / OL]. Automation of Electric Power Systems, 1-9.] proposed a pricing strategy for charging stations in a power-transportation coupling network considering user travel cost budget, which can reasonably formulate the charging price of the charging station according to the user's travel behavior mechanism.
[0004] However, most of the above literature on charging guidance strategies is based on the assumption that users are completely rational. In fact, users often show limited rational behavior when making charging decisions, which makes the existing charging guidance strategies have certain limitations in practical applications. Therefore, increasing the research on user charging decision-making behavior to more accurately understand and predict user charging behavior is of great significance to improving the practicality and effectiveness of charging guidance strategies. Summary of the invention
[0005] In order to overcome the shortcomings of the above-mentioned prior art, the present invention provides an electric vehicle charging guidance method that takes into account the user's habit of selecting charging stations. It uses the user's time cost and price cost data to make a detailed profile of the user, and identifies the degree of preference of different users for time and price when selecting charging stations; based on the results of user portraits, a satisfaction model for different user groups is constructed. Finally, through actual calculation verification, it can be seen that it has shown significant advantages in improving user charging satisfaction and charging station operation efficiency, which fully proves its effectiveness and applicability.
[0006] The technical solution of the present invention:
[0007] On the one hand, the present invention provides an electric vehicle charging guidance method taking into account the user's habit of selecting a charging station, which is characterized in that it includes the following steps:
[0008] Step S1. Build user portrait:
[0009] The SOFM method is used to perform cluster analysis on the user's historical charging data, with the user's time cost and fee cost as characteristic parameters;
[0010] The clustering results of the SOFM algorithm are adjusted using the SHAP method to obtain the best user portrait results and divide users into different user groups. Each group represents a set of users with similar charging decision preferences.
[0011] Step S2. Establishment of satisfaction model:
[0012] The entropy weight method is used to calculate the weights of time cost and expense cost for each type of user in the process of selecting a charging station;
[0013] According to the calculated weights, the user's time cost and expense cost are converted into user satisfaction indicators to build a user satisfaction model;
[0014] Step S3. Charging guide:
[0015] When an electric vehicle user sends a charging request, the user's satisfaction at different charging stations is calculated based on the user's real-time location information, remaining power, and usage of the charging station, combined with the user portrait results and satisfaction model;
[0016] Select the charging station with the highest user satisfaction for charging guidance and output the charging guidance results.
[0017] Preferably, in the user portrait construction step, the time cost includes the driving time of the user from the departure point to the charging station and the queuing time during the charging process, and the cost cost includes the charging fee, service fee and parking fee.
[0018] Preferably, the satisfaction model uses a cosine distribution function to convert time cost and expense cost into user satisfaction, and the user satisfaction is calculated based on the weights of time cost and expense cost.
[0019] Preferably, the method further includes a step of acquiring real-time traffic information, and acquiring the traffic conditions at different time periods in real time through a digital map interface to accurately calculate the driving time of the user to the target charging station.
[0020] Preferably, the method further includes a charging aggregator profitability evaluation step, in which the profitability of the charging aggregator is evaluated by calculating the unit time profit of the charging pile.
[0021] On the other hand, the present invention also provides a computer system for implementing the above-mentioned electric vehicle charging guidance method, comprising:
[0022] A data collection module for collecting real-time and historical charging data from electric vehicle users;
[0023] User portrait module, used to execute SOFM and SHAP algorithms to profile users;
[0024] Satisfaction calculation module, used to calculate user satisfaction based on portrait results and real-time data;
[0025] A charging guidance module, used to guide the selection of charging stations based on user satisfaction;
[0026] The profit evaluation module is used to evaluate the profit per unit time of the charging pile.
[0027] Compared with the prior art, the present invention has the following beneficial effects:
[0028] 1) By building a user satisfaction model and comprehensively considering time cost and expense cost, the user charging experience can be quantitatively evaluated and optimized, which helps charging station operators better understand user needs and provide personalized charging services, thereby improving user satisfaction and loyalty.
[0029] 2) By implementing the charging guidance model, the problems of difficult and slow charging of electric vehicles are effectively solved, which promotes the popularization and promotion of electric vehicles, reduces dependence on traditional fuel vehicles, and promotes environmental protection and sustainable development.
[0030] 3) By adopting the SOFM algorithm to profile the electric vehicle group, accurate prediction and management of the use of charging facilities can be achieved, which helps charging facility operators to reasonably arrange the construction and layout of charging facilities and improve the utilization rate and operational efficiency of charging facilities. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 It is a user portrait flow chart of the present invention.
[0032] Figure 2 It is a charging guide flow chart of the present invention. DETAILED DESCRIPTION
[0033] The present invention is further explained below in conjunction with the accompanying drawings and embodiments, but the protection scope of the present invention should not be limited thereto.
[0034] A method for guiding electric vehicle charging taking into account a user's habit of selecting a charging station comprises the following steps:
[0035] 1. Data Collection:
[0036] When EV users need charging services, the charging platform collects users' real-time information and historical charging data.
[0037] Real-time information includes: user’s current location coordinates, charging application time, remaining SOC, fast charging station usage, charging station coordinates, charging price, etc.
[0038] Historical charging data includes: charging start time, historical charging power, historical queuing time, historical charging price, user satisfaction evaluation, etc.
[0039] 2. Path planning:
[0040] The platform uploads the user’s current coordinates and the fast charging station’s coordinates to the navigation system.
[0041] The navigation system gives the planned route and arrival time based on real-time traffic conditions.
[0042] The mileage data is transmitted back to the charging platform.
[0043] 3. User portrait creation
[0044] The charging platform creates a charging decision profile based on the user's historical charging station selection behavior based on the user's historical data.
[0045] The portrait is mainly based on time cost and expense cost, where time cost includes driving time and waiting time, and expense cost includes charging cost, service fee and parking fee.
[0046] The SOFM (Self-Organizing Feature Mapping) algorithm is used to divide users into different user clusters based on their charging behavior data. The user's charging decision preference category and the satisfaction reference value of different categories of users are determined through the portrait results. The clustering results of the SOFM algorithm are adjusted using the SHAP method to obtain the best user portrait results, and users are divided into different user groups. Each group represents a collection of users with similar charging decision preferences, such as Figure 1 shown.
[0047] In training, a data Xs is randomly selected from the input space, and then its similarity with each neuron is calculated, and the neuron with the highest similarity is selected as the winning neuron. After the winning neuron is obtained, the weights of the neighboring neurons of the winning neuron are continuously iterated through equation (1), equation (2) and equation (3) according to equation (2) until the final image result is obtained.
[0048] μ j =(x(t)-w j (t)) (1)
[0049] w j (t+1)=w j (t)+α(t)h(j,c)μ j (2)
[0050]
[0051] In the above formula: w j is the weight of neuron j. α(t) is the learning rate; h(j,c) is the domain function of neuron j and neuron c; Z j and Z c are the coordinates of neuron j and neuron c respectively; δ(t) is the field radius.
[0052] 4. Establishment of satisfaction model:
[0053] The entropy weight method is used to calculate the weights of time cost and expense cost for each type of user in the process of selecting a charging station;
[0054] According to the calculated weights, the user's time cost and expense cost are converted into user satisfaction indicators to build a user satisfaction model. Specifically, it includes:
[0055] 4.1 Time cost calculation
[0056] Arrival time: Through the digital map interface, the time and space differences of traffic flow are taken into account to determine the arrival time and detour distance of EVs at the destination at different times.
[0057] Queueing time: Calculate the EV's queueing time based on the remaining charging time of the vehicle currently being charged at the charging pile and the travel time of the user's EV to the charging station:
[0058]
[0059] is the queuing time of user i in charging pile j. For EV i The current remaining charging time of the EV at the charging pile j when the charging request is issued For EV iThe driving time from the current location to charging station j; The number of charging piles.
[0060] Time cost: The time cost of the user is calculated by combining the arrival time and the waiting time:
[0061]
[0062] For EV i The comprehensive time cost to reach charging station j, For EV i Driving time from the current location to charging station j For EV i The time required to wait in line after arriving at charging station j.
[0063] 4.2 Cost Calculation
[0064] Electricity fee: The electricity fee is calculated based on the charging time of the EV at the charging station, the initial load status, the load status after charging, the battery capacity and the charging power.
[0065] Service fee and parking fee: Additional fees are calculated based on the service fee unit price and parking fee unit price of the charging station.
[0066] Cost: Calculate the user's cost by combining electricity charges, service fees and parking fees.
[0067] EV i The cost of charging station j can be expressed as M ij express:
[0068]
[0069] Where: t ij The charging time of user i at charging station j; is the initial load state and the load state after charging; C i is the battery capacity of the i-th EV; P ij is the charging power of user i at charging station j. They are the charging unit price, service fee unit price, and parking fee unit price of charging station j respectively.
[0070] 4.3 User Satisfaction Calculation
[0071] Satisfaction normalization: Use the cosine distribution function to normalize time cost and expense cost into user satisfaction.
[0072] Weight allocation: Assign weights to time cost and expense cost, and calculate the user's comprehensive satisfaction at the charging station.
[0073]
[0074] in and are the minimum and maximum satisfaction thresholds of y respectively. From formula (12), we can get EV i User charging satisfaction G at charging station j ij for:
[0075]
[0076] In formula (12): are the satisfaction of time and cost respectively. β and λ are EV i The weights of the time and cost at charging station j, and β+λ=1.
[0077] 4.4 Charging plan selection
[0078] Profit considerations: Calculate the profit per unit time of the charging pile as a reference factor for selecting charging plans.
[0079] The highest satisfaction principle: select the best charging route for users based on user satisfaction and charging pile profitability.
[0080] The unit time profit of the charging pile can be expressed as:
[0081]
[0082] Formula (13): R s is the total profit cost of the charging station, T s It is the total service time of the charging pile.
[0083] 5. Real-time charging guidance:
[0084] When an electric vehicle user sends a charging request, the user's satisfaction at different charging stations is calculated based on the user's real-time location information, remaining power, and usage of the charging station, combined with the user portrait results and satisfaction model;
[0085] Select the charging station with the highest user satisfaction for charging guidance and output the charging guidance results.
[0086] User satisfaction is used to measure the benefits of EV users. Assume that there are I EV users and J charging stations in the road network. After dividing the users into N categories, each category contains I n EV users. At this time, the user objective function can be expressed as:
[0087]
[0088] Where: x ij EV for users iIs it charging at charging station j? If so, then x ij If it is not charged, x ij is 0.
[0089] The EV's mileage is limited by its remaining charge, so users only consider charging stations within the mileage range when selecting charging stations. The mileage calculation formula is as follows:
[0090]
[0091] D ij ≤D i (16)
[0092] Where: For electric vehicles EV i Load status at departure, C i For EV i The battery capacity, i For EV i Power consumption per kilometer, D i For EV i Maximum mileage; D ij For EV i The distance between the electric vehicle and the matching fast charging pile j requires that the distance between the electric vehicle and the matching fast charging pile does not exceed the maximum drivable mileage of the electric vehicle.
[0093] The number of charging piles in the charging station is fixed, so the sum of the charging power of V users of the charging station at the same time should be less than the total power P l , expressed by the following formula:
[0094]
[0095] Where P l is the total power of the charging station, P v represents the power used by the vth electric vehicle to charge at the charging station. Formula (18) indicates that each electric vehicle can only choose one charging station. When user i chooses charging station j, x ij is 1 otherwise x ij is 0.
[0096] The charging guidance process for EV users is as follows Figure 2 As shown, the specific solution steps are as follows:
[0097] a. Prepare the original data, select part of the data X as the input of the SOFM neural network, and set the number of neurons and the maximum number of iterations.
[0098] b. Initialize all neural weights and prepare for clustering operations.
[0099] c. Run the SOFM neural network and output the clustering result Y.
[0100] d. Use the SHAP method to analyze the importance of features and select the number of clusters whose time cost and expense cost are closest to 0.5 in the profiling process. If the conditions are not met, return to the first step and readjust.
[0101] e. Use the entropy weight method for the selected clusters to calculate the weights of time cost and expense cost.
[0102] f. Set thresholds for time cost and expense cost according to actual conditions, and build satisfaction models for various types of users based on these thresholds.
[0103] g. When the user sends a charging request, the user's total time cost Tsum is updated through the digital interface.
[0104] h. Use the user satisfaction model to calculate the user's satisfaction in the current state.
[0105] i. Provide charging guidance based on user satisfaction, recommend the most suitable charging station to the user, and output the satisfaction results.
[0106] Example:
[0107] 1. EV user charging platform implementation
[0108] In order to realize efficient and intelligent charging services for EV users, a charging platform is constructed, which has the following functions:
[0109] Information collection and upload: When an EV user submits a charging application, the platform first collects the user's real-time information (such as current location, charging application time, remaining SOC, etc.) and historical charging data (such as historical charging power, queuing time, etc.).
[0110] Route planning and navigation: The platform uploads the user's current coordinates and the fast charging station coordinates to the navigation system, which then gives the planned route and arrival time based on real-time traffic conditions.
[0111] Charging decision profile: The platform builds a charging decision profile based on the user's historical data and selects the charging planning route with the highest user satisfaction.
[0112] Assume that there are 100 EV users and 5 charging stations in a city. The battery capacity, remaining power, power consumption per kilometer, and maximum mileage of each EV user are known, and the maximum charging power and location information of each charging station are also determined.
[0113] Data preparation:
[0114] Collect the battery capacity, remaining power, and power consumption per kilometer of 100 EV users.
[0115] Collect the location information, number of charging piles and maximum charging power of 5 charging stations.
[0116] 2. Implementation of electric vehicle group portrait based on SOFM algorithm
[0117] In order to better guide users to charge in an orderly manner, the SOFM algorithm is used to profile the electric vehicle group. The specific steps are as follows:
[0118] Feature construction: The driving time of users to the charging station, the waiting time, the charging power and the charging price are selected as the main features.
[0119] Profiling process: Cluster users through the SOFM algorithm to form different user aggregates. Set the number of neurons to 5 and the maximum number of iterations to 100.
[0120] Run the SOFM neural network and get the clustering result Y, that is, 100 EV users are divided into 5 categories.
[0121] The SHAP method was used to analyze the importance of each feature, and the number of clusters whose importance of time cost and expense cost in the profiling process was closest to 0.5 was selected.
[0122] The entropy weight method is used for the selected clusters to calculate the weights of time cost and expense cost.
[0123] According to the actual situation, set the thresholds of time cost and expense cost for subsequent user satisfaction calculation.
[0124] Output results: Determine the user's charging decision preference category and the satisfaction reference value of different categories of users based on the portrait results.
[0125] 3. User satisfaction model implementation
[0126] In order to measure the user's satisfaction with the charging service, a user satisfaction model was constructed. The specific steps are as follows:
[0127] Time cost calculation: Considering the driving time and waiting time of users to the charging station, the arrival time and detour distance are determined through the digital map interface.
[0128] Cost calculation: including electricity charges, service fees and parking fees, calculated based on charging time and charging station charging standards.
[0129] Satisfaction normalization: Use time cost and expense cost to normalize user satisfaction into a nonlinear function to better characterize user charging decision-making behavior.
[0130] 4. Implementation of charging guidance model
[0131] In order to achieve effective utilization of charging stations and improve user satisfaction, a charging guidance model was constructed. The specific steps are as follows:
[0132] Objective function construction: User satisfaction is used as the objective function, taking into account the number and distribution of EV users and charging stations.
[0133] Constraint setting: including EV mileage constraints, charging station maximum capacity constraints and electric vehicle matching constraints.
[0134] Model solving: The model is solved by dividing the original data, initializing the neural weights, and outputting the clustering results, and finally the charging guidance plan with the highest user satisfaction is obtained.
[0135] Through the above embodiments, the present invention combines SOFM cluster analysis with the SHAP method to construct a portrait of electric vehicle users, thereby improving the accuracy and practicality of user portraits. The cosine distribution function is used to convert time cost and expense cost into user satisfaction indicators, providing a new quantitative basis for charging guidance. At the same time, the real-time road condition information is integrated into the charging guidance process, making the calculation of driving time more accurate and improving the practical application value of charging guidance. The introduction of the charging aggregator profit evaluation step provides a new perspective and tool for the operation and planning of charging stations. In summary, the present invention uses the SOFM algorithm to perform electric vehicle group portraits, construct user satisfaction models and charging guidance models, and realizes efficient and intelligent guidance of charging services for EV users.
Claims
1. A method for guiding electric vehicle charging taking into account the user's habit of selecting a charging station, characterized in that: The following steps are involved: Step S1. Build user portrait: The SOFM method is used to perform cluster analysis on the user's historical charging data, with the user's time cost and fee cost as characteristic parameters; The clustering results of the SOFM algorithm are adjusted using the SHAP method to obtain the best user portrait results and divide users into different user groups. Each group represents a set of users with similar charging decision preferences. Step S2. Establishment of satisfaction model: The entropy weight method is used to calculate the weights of time cost and expense cost for each type of user in the process of selecting a charging station; According to the calculated weights, the user's time cost and expense cost are converted into user satisfaction indicators to build a user satisfaction model; Step S3. Charging guide: When an electric vehicle user sends a charging request, the user's satisfaction at different charging stations is calculated based on the user's real-time location information, remaining power, and usage of the charging station, combined with the user portrait results and satisfaction model; Select the charging station with the highest user satisfaction for charging guidance and output the charging guidance results.
2. The electric vehicle charging guidance method according to claim 1, characterized in that: In the user portrait construction step, the time cost includes the driving time from the departure point to the charging station and the queuing time during the charging process, and the cost cost includes the charging fee, service fee and parking fee.
3. The electric vehicle charging guidance method according to claim 1, characterized in that: The satisfaction model uses a cosine distribution function to convert time cost and expense cost into user satisfaction, and user satisfaction is calculated based on the weights of time cost and expense cost.
4. The electric vehicle charging guidance method according to claim 1, characterized in that: The method also includes a step of obtaining real-time traffic information, which obtains traffic conditions at different time periods in real time through a digital map interface to accurately calculate the driving time of the user to the target charging station.
5. The electric vehicle charging guidance method according to claim 1, characterized in that: It also includes a charging aggregator profitability evaluation step, which evaluates the profitability of charging aggregators by calculating the unit time profit of charging piles.
6. A computer system for implementing the electric vehicle charging guidance method according to any one of claims 1 to 5, comprising: A data collection module for collecting real-time and historical charging data from electric vehicle users; User portrait module, used to execute SOFM and SHAP algorithms to profile users; Satisfaction calculation module, used to calculate user satisfaction based on portrait results and real-time data; A charging guidance module, used to guide the selection of charging stations based on user satisfaction; The profit evaluation module is used to evaluate the profit per unit time of the charging pile.
Citation Information
Cited By
Electric heavy truck trip chain simulation method and system considering charging strategy optimization
CN119963071A
Intelligent charging pile based on user behavior analysis and control method thereof
CN120996494A
Cooling liquid filling and exhausting method for fuel cell system
CN121097119A
