Method for optimizing charging and discharging of electric vehicle in real time

Abstract

The invention belongs to the technical field of electric vehicle charging and discharging, and particularly relates to a method for optimizing charging and discharging of an electric vehicle in real time, which comprises the following steps: S1, constructing a time-space characteristic model of a charging demand; s2, constructing a partition dynamic electricity price model based on the real-time voltage level of each node of the power distribution network, and setting a discharge excitation electricity price to form an electricity price collaborative guidance model; s3, dividing the users into rigid users and elastic users; for the elastic user, the discharge response will is quantified based on a fuzzy reasoning model, and a discharge will value is obtained; the discharging willingness value and the battery degradation cost are jointly incorporated into a charging and discharging decision-making model of the elastic user; s4, constructing a multi-target optimization model with the lowest total user cost, the lowest power distribution network load variance and the highest user comprehensive satisfaction as targets; and S5, solving the multi-objective optimization model, and determining a charging and discharging power scheduling plan of the electric vehicle cluster. According to the method, safe operation of the power grid, and collaborative improvement of user economic cost and comprehensive satisfaction can be realized.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of electric vehicle charging and discharging, and particularly relates to a method for real-time optimization of electric vehicle charging and discharging. BACKGROUND

[0002] With the deepening of the "double carbon" strategy, electric vehicles (EVs) as the core carriers of transportation electrification are accelerating the scale of access to the distribution network. However, their disordered and centralized charging behavior, while improving the level of clean energy use at terminals, also poses a serious challenge to the safe and stable operation of the distribution network. A large number of users are accustomed to charging in high-load peak periods such as evening peak or in weak areas of the power grid, resulting in a sharp increase in local node load, an intensification of voltage deviation, and even an out-of-limit risk, forming a typical "peak on peak" phenomenon. This not only amplifies the system peak-valley difference and reduces the efficiency of equipment utilization, but also may cause power quality problems and threaten the overall reliability of the power grid.

[0003] More critically, most current charging and discharging scheduling strategies lack a detailed depiction of users' real response willingness and effective incentive mechanisms. On the one hand, users' enthusiasm for participating in vehicle-to-grid (V2G) is influenced by multiple factors such as battery degradation cost, travel uncertainty, and economic benefit expectations; on the other hand, existing guiding methods often use static or coarse-grained price signals, which are difficult to accurately reflect the actual operating state and adjustment demand of the power grid at different time-space nodes, resulting in low actual participation of users and limited scheduling effectiveness, ultimately restricting the coordinated improvement of user satisfaction and power grid regulation efficiency.

[0004] To address the above challenges, researchers in the field have conducted a number of technical explorations. For example, patent CN119010132A proposes a three-stage optimization method for electric vehicle charging and discharging based on price guidance, which realizes peak clipping and valley filling to some extent through "charge and go" and "orderly charging and discharging" dual-mode selection and dynamic priority mechanism. However, this method highly depends on users' rational decision-making and active cooperation, does not fully model the uncertainty of user behavior, and ignores the coupling effects of traffic travel chains and renewable energy fluctuations, limiting its adaptability in complex urban energy systems. For another example, patent CN119647896A constructs a spatiotemporal flexibility economic dispatching model that integrates traffic networks and distribution networks, effectively balancing the spatial distribution of charging load and reducing user costs, but it does not include battery cycle aging costs caused by V2G discharging in the decision-making framework, ignoring the negative impact of battery wear on users' long-term economic benefits and participation willingness.

[0005] In summary, although existing research has made some progress in multi-network collaboration, dynamic pricing, and load guidance, the following core problems still exist: (1) Rough user behavior modeling - it fails to accurately depict the spatiotemporal distribution characteristics of electric vehicle charging demand based on real travel data, and also lacks quantitative assessment of rigid / elastic user classification and discharge willingness; (2) Imperfect incentive mechanism - existing electricity price models are mostly globally unified or fixed for time periods, making it difficult to achieve "different policies for different areas and pricing according to time", and unable to effectively guide EVs to avoid weak nodes in the spatial dimension and to stagger peak hours or reverse power supply in the time dimension; (3) Lack of economic-technical trade-off - the cost of battery degradation and the benefits of discharge have not been optimized in a coordinated manner, resulting in users facing a dilemma between economic efficiency and battery life when participating in V2G, which inhibits their enthusiasm for discharge.

[0006] Therefore, how to achieve a coordinated improvement in power grid safety, user economic costs, and overall satisfaction has become an urgent problem to be solved. Summary of the Invention

[0007] To address the shortcomings of the existing technologies, this invention provides a method for real-time optimized charging and discharging of electric vehicles, which can achieve a synergistic improvement in grid safety, user economic costs, and overall satisfaction.

[0008] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0009] A method for real-time optimized charging and discharging of electric vehicles includes the following steps:

[0010] S1. Based on the travel chain theory and historical travel data, fit the probability distribution of electric vehicle trip start and end time, driving mileage and parking time, and construct a spatiotemporal characteristic model of electric vehicle charging demand to characterize the charging load distribution of electric vehicles at different time and space nodes and predict their charging behavior.

[0011] S2. Based on the real-time voltage levels of each node in the distribution network, a regional dynamic electricity price model is constructed, and a discharge incentive price is superimposed during peak electricity consumption periods to form a coordinated electricity price guidance model. The regional dynamic electricity price model guides the spatial distribution of charging loads for electric vehicles. The discharge incentive price incentivizes electric vehicles with discharge capabilities to supply power to the distribution network during peak hours. The coordinated electricity price guidance model provides spatiotemporally differentiated charging and discharging price signals to reflect the grid operation status and user incentive levels in different regions and time periods.

[0012] S3. Based on the current SOC of the electric vehicle and the power required for the next leg of the journey, users are divided into rigid users and flexible users; wherein, the rigid users are those whose current SOC is less than a preset charging trigger value, or whose remaining power is less than the power threshold required for the next leg of the journey.

[0013] For flexible users, their willingness to discharge is quantified based on a fuzzy inference model to obtain a discharge willingness value. The discharge willingness value and battery degradation cost are jointly incorporated into the charging and discharging decision model of flexible users to weigh the discharge benefits and battery losses, and determine whether to participate in discharging and the corresponding discharge power level. The battery degradation cost is the cost obtained by economically converting the equivalent capacity decay caused by electric vehicles participating in discharging based on a battery cycle aging model applicable to vehicle-to-grid interaction scenarios.

[0014] S4, based on the spatiotemporal characteristic model of charging demand in S1, the electricity price coordination guidance model in S2, and the charging and discharging decision model in S3, construct a multi-objective optimization model with the goals of minimizing total user cost, minimizing distribution network load variance, and maximizing overall user satisfaction.

[0015] S5. Solve the multi-objective optimization model to determine the charging and discharging power scheduling plan of the electric vehicle cluster, which is used to guide each electric vehicle to perform real-time charging and discharging operations.

[0016] Glossary: ​​SOC stands for State of Charge, which refers to the state of charge of a battery.

[0017] Compared with the prior art, the present invention has the following beneficial effects:

[0018] 1. Accurately depict the spatiotemporal distribution of charging demand, improving the accuracy of load forecasting. Unlike existing methods that often employ static or averaged travel assumptions, this approach uses travel chain theory and historical data to fit the probability distributions of trip start and end times, mileage, and parking duration, constructing a charging demand model with individual differences and spatiotemporal dynamics. This model can more realistically reflect users' actual travel patterns, significantly improving the prediction accuracy of electric vehicle cluster charging load in both time and space dimensions, providing reliable input for subsequent scheduling.

[0019] 2. Implement differentiated pricing guidance based on grid status awareness to optimize load spatial distribution. Unlike traditional uniform time-of-use pricing or simple time-of-use pricing mechanisms, this solution innovatively combines the real-time voltage levels of each node in the distribution network to construct a regional dynamic pricing model, and superimposes discharge incentive pricing during peak hours. This mechanism not only guides users to avoid charging in areas with weak grid connections, alleviating local overload and voltage exceedance issues, but also proactively incentivizes vehicles with V2G capabilities to supply power in reverse during critical periods, effectively suppressing peak-on-peak phenomena and enhancing the resilience of the distribution network.

[0020] 3. Refined user segmentation and quantification of discharge willingness enhance user engagement. Existing scheduling strategies often treat users as a homogeneous group, ignoring the differences in their travel rigidity and response flexibility. This solution categorizes users into rigid and flexible groups based on SOC (State of Occupation) and trip demand. For flexible users, a fuzzy inference model is introduced to quantify their discharge willingness, making scheduling decisions more aligned with users' actual psychological and behavioral characteristics. Compared to a "one-size-fits-all" scheduling approach, this method significantly enhances user acceptance and response reliability.

[0021] 4. Embedded battery degradation costs achieve a reasonable balance between economic efficiency and battery life. Addressing the common problem in existing technologies that neglect the impact of V2G on battery aging, this solution incorporates battery degradation costs based on cycle aging mechanisms into the decision-making model for resilient users. This allows users to consider long-term usage costs while evaluating discharge benefits. This mechanism avoids user resistance due to over-discharge, maintains reasonable economic incentives while ensuring battery health, and enhances the sustainability of the V2G model.

[0022] 5. Unlike single-objective methods that focus solely on peak shaving on the grid side or minimizing user-side costs, this scheme simultaneously optimizes three objectives: total user cost, distribution network load variance, and overall user satisfaction. It generates a charging and discharging scheduling plan that balances technical feasibility and user experience through multi-objective solutions. This systematic and collaborative design effectively bridges the conflict between grid control requirements and individual user interests.

[0023] In summary, this method can achieve a synergistic improvement in power grid safety, user economic costs, and overall satisfaction.

[0024] Preferably, in S1, the spatiotemporal characteristic model of charging demand is constructed based on travel chain theory, including:

[0025] Based on historical travel data, probability distribution models that follow a normal distribution are established for the start and end times, mileage, and parking time of electric vehicles.

[0026] The SOC at the end of the trip is determined based on the driving mileage, and the trip end time is taken as the time when the charging demand occurs; at the same time, a deterministic functional relationship between the SOC, battery capacity, charging power and charging efficiency is established for the charging duration.

[0027] The probability distribution model and the deterministic functional relationship are used together to characterize the charging load distribution of electric vehicles in the spatiotemporal dimension.

[0028] This approach, unlike traditional methods that use fixed or average value assumptions, establishes a probability distribution model for trip start and end times, mileage, and parking duration. This more realistically reflects the randomness and diversity of user travel behavior, avoiding load prediction bias caused by simplistic assumptions. By using the trip end time as the occurrence time of charging demand and establishing a deterministic function of charging duration based on SOC, battery parameters, etc., the dynamic positioning of charging events on the time axis and accurate estimation of their duration are achieved, enhancing the model's temporal resolution and physical rationality.

[0029] 2. By combining probabilistic models with deterministic functions, not only are the uncertainties of user behavior captured, but the physical constraints of the charging process are also preserved, giving the charging load distribution stronger spatial distribution characteristics and temporal evolution patterns, providing high-precision input for subsequent scheduling.

[0030] Preferably, in S2, the zoned dynamic electricity price model guides electric vehicles to rationally distribute charging loads in space by increasing the charging price in areas where the voltage is lower than a preset low voltage value and decreasing the charging price in areas where the voltage is higher than a preset high voltage value.

[0031] The regional dynamic electricity pricing model includes:

[0032]

[0033] In the formula, represents the regional dynamic electricity price of the area where node k is located at time t; h represents the base electricity price adjustment coefficient; This represents the per-unit voltage value of the distribution network node where region k is located at time t; c represents the electricity price at time t; gd The benchmark electricity price; P t th This represents the estimated theoretical average load at time t; This indicates the expected average load.

[0034] The discharge excitation price for:

[0035]

[0036] ε t =1+tanh(g t |δ i,t,dis ·P i,t,dis |);

[0037] In the formula, ε i,t As a motivating factor; g t δ represents the excitation coefficient at period t, and δ represents the rate of change of the ratio of discharge power to excitation price in each time period; i,t,disP is the discharge decision variable, taking a value of 1 when the electric vehicle discharges at time t, and 0 otherwise; i,t,dis This is the rated discharge power.

[0038] This setup has several advantages: 1. Unlike traditional uniform electricity pricing or simple time-of-use pricing, this scheme directly couples voltage levels into the electricity price calculation, creating spatial price differences. Increasing charging prices in weak areas with low voltage can effectively discourage users from concentrating their charging in these areas, avoiding localized overloads and voltage drops, and improving the stability of the distribution network.

[0039] 2. Enhance the real-time response capability of electricity price signals to the grid operating status. Regional dynamic electricity pricing not only depends on voltage levels but also incorporates the changing trend of the system's average load (via P). t th and (Comparison) This model enables dynamic responses of electricity prices in both time and space dimensions. Compared to static or time-based pricing mechanisms, this model can more sensitively reflect the actual operating status of the power grid, improving the accuracy and adaptability of dispatching.

[0040] 3. Establish a differentiated discharge incentive mechanism to stimulate users' enthusiasm for participating in V2G. The discharge incentive price is not a fixed value, but rather varies with the user's discharge decision variable δ. i,t,dis Related to discharge power, through the excitation factor ε i,t This mechanism enables "on-demand incentives." It dynamically adjusts incentive levels to ensure higher returns for discharging during high-value periods, thereby significantly enhancing users' economic motivation to participate in vehicle-to-everything (V2X) interactions.

[0041] Preferably, in S3, the fuzzy inference model takes the discharge incentive price, the current SOC, and the parking duration as input variables. After fuzzification, fuzzy rule inference, and defuzzification, it outputs a discharge response willingness value that represents the degree of user's willingness to participate in discharge.

[0042] Unlike existing methods that often treat users as passive responders or rely solely on price signals, this solution introduces fuzzy reasoning to generate discharge intention values. This effectively captures users' subjective preferences and behavioral tendencies in uncertain environments, making the decision-making model closer to real user psychology and improving the executability of scheduling strategies.

[0043] Preferably, in S3, the battery cycle aging model is:

[0044]

[0045] In the formula, α, β, γ, σ, ∈ are predetermined cyclic aging coefficients; The charge / discharge rate at time t; Ah i,tis the battery's ampere-hour throughput; K is the electric vehicle's cumulative equivalent cycle count within a time period, used to quantify the degree of battery cycle aging; E represents the cyclic aging loss per unit time for the i-th electric vehicle at time t; Bat The rated capacity of the battery; δ i,t,cha P is the charging decision variable, taking a value of 1 when the electric vehicle is charging at time t, and 0 otherwise; i,t,cha P represents the discharge power of an electric vehicle. i,t,dis V represents the rated discharge power; V is the battery's nominal voltage, assumed to remain constant during battery operation; Δt represents the charge / discharge time.

[0046] The battery degradation cost for:

[0047]

[0048] In the formula, BP and SP represent the purchase price and scrap price of the battery, respectively; E n This refers to the energy capacity at the end of the battery's life.

[0049] This setup enables the scientific quantification and economic calculation of the aging and loss of electric vehicle batteries during vehicle-to-grid (V2G) interaction, effectively supporting users' rational decision-making between discharge benefits and battery life, and promoting the transformation of the V2G model from "short-term incentive-driven" to "long-term sustainable participation".

[0050] Preferably, in S3, the charging and discharging decision model for the flexible user includes a discharging decision sub-model and a charging decision sub-model;

[0051] The discharge decision sub-model is used to calculate the discharge benefits of elastic users, and combined with the discharge response intention value generated by the fuzzy inference model, it determines whether elastic users should discharge parametrically.

[0052] The charging decision sub-model is used to calculate the overall charging willingness of flexible users who do not participate in discharging, based on the degree of range anxiety and the influence factor of regional dynamic electricity price, and to determine whether to perform charging operation according to the preset threshold.

[0053] This setup offers several advantages: 1. It supports the precise characterization and guidance of differentiated user behaviors. By setting up independent sub-models for discharging and charging decisions, it can simultaneously address two groups: users participating in V2G and users who only need charging, avoiding biases caused by applying the same strategy to multiple users. This hierarchical decision-making mechanism significantly improves the system's adaptability to heterogeneous user groups.

[0054] 2. Enhance user engagement and sustainability. Discharge decisions consider not only benefits but also battery degradation costs and user preferences, enabling users to make rational choices that align with their own interests. Charging decisions combine range anxiety and electricity price incentives to avoid risks associated with excessively delayed charging. This dual-protection mechanism increases user trust in the dispatch strategy and their willingness to participate long-term.

[0055] Preferably, the discharge benefit is calculated using the following formula:

[0056]

[0057] In the formula, C income P represents the discharge gain; i,t,dis Rated discharge power; Δt is the charge / discharge time; η dis For discharge efficiency;

[0058] When C income When the value of the discharge response intention is greater than 0 and the value of the discharge response intention is greater than the preset intention value, the discharge decision sub-model determines that the elastic user participates in the discharge.

[0059] This setup bases discharge decisions not only on whether the economic benefit is positive, but also on a dual assessment of the user's subjective willingness generated by fuzzy reasoning. This composite criterion of "objective benefit + subjective willingness" effectively mitigates the risk of users refusing to respond due to temporary emotions or non-economic factors, improving the success rate of dispatch command execution. It enables refined economic assessment and intelligent decision-making regarding electric vehicle discharge behavior, effectively balancing the relationship between user economic benefits and battery health, promoting a shift in vehicle-to-grid interaction from "mandatory participation" to "voluntary response," and enhancing the flexibility of system regulation and user satisfaction.

[0060] Preferably, the charging decision sub-model calculates the comprehensive charging intention W of elastic users who do not participate in discharging using the following formula. i,t :

[0061] W i,t =β i,t ·R i,t +(1-β i,t )·η t ;

[0062]

[0063]

[0064] In the formula, β i,t R represents the weighting coefficients; i,t Range anxiety level; η t The factors influencing regional dynamic electricity prices; and These represent the maximum and minimum values ​​of the dynamic electricity price for each zone; when the user's overall charging intention W... i,t If the value is greater than 0.5, the user selects to charge; otherwise, charging is not performed. (SOC) Max The maximum state of charge (SOC) of the i-th electric vehicle. i,t Let D be the real-time SOC of the i-th electric vehicle at time t; i α represents the distance an electric vehicle travels to its next destination. ev This refers to the energy consumption per unit distance of an electric vehicle.

[0065] Unlike traditional methods that rely solely on SOC thresholds or electricity price signals as charging triggers, this approach weights and integrates two core driving factors—range anxiety and electricity price—to construct a comprehensive charging willingness index. This model more comprehensively reflects users' psychological state and economic considerations in real-world travel scenarios, enhancing the scientific rigor of charging decisions. It enables refined and intelligent guidance of flexible user charging behavior, effectively balancing user travel safety needs with grid economic operation goals, promoting the transformation of electric vehicles from passive loads to proactively responding resources, and improving the overall coordination and sustainability of the energy system.

[0066] Preferably, in S4, the objective function of the multi-objective optimization model is:

[0067] min{F A (χ i ),-F B (χ i ),F C (χ i )};

[0068] In the formula, F A (χ i Let F be the economic cost of charging and discharging the i-th electric vehicle. B (χ i F represents the user satisfaction with the charging needs of the i-th electric vehicle. C (χ i ) represents the power distribution network load fluctuation of the i-th electric vehicle;

[0069]

[0070] α n1 =min{SOC f,i,e / SOC f,i,exp ,1};

[0071] α n2 =min{1 / (1+η)} i,charge ),1};

[0072]

[0073] In the formula, N is the total number of electric vehicles; T is the scheduling time period; λ1 and λ2 are weighting coefficients; α n1 Indicates user charging satisfaction; α n2 Indicates battery wear satisfaction; SOC f,i,e This indicates the electric vehicle's battery level when the user arrives at the charging station; SOC f,i,exp Indicates the user's expected battery level when leaving; η i,charge This represents the number of charge / discharge transitions for the i-th electric vehicle during the scheduling process.

[0074] This setup achieves two key advantages: 1. It enables coordinated optimization of goals on both the grid and user sides, improving overall system efficiency. Unlike traditional single-objective methods that focus solely on grid peak shaving or minimizing user costs, this solution integrates user economic costs, charging satisfaction, and grid load fluctuations into a unified objective function, achieving a balance of interests among multiple parties through a multi-objective optimization framework. This systematic design avoids the problem of "paying for one thing at the expense of another," significantly improving the overall coordination of the scheduling strategy.

[0075] 2. Balancing user travel safety and long-term asset protection enhances user engagement. User satisfaction metrics not only include whether the battery level meets expectations upon arrival (charging satisfaction) but also incorporate considerations of battery degradation (the impact of charge / discharge cycles). This ensures that while guiding user responses, the system also respects their concerns about battery life. This mechanism effectively alleviates user resistance to frequent V2G operations and increases long-term engagement.

[0076] 3. By minimizing the load fluctuation target F C This model can proactively suppress load peak-valley differences, reducing voltage fluctuations and equipment overload risks. Especially against the backdrop of large fluctuations in renewable energy output, this smoothing effect helps improve the distribution network's ability to accommodate uncertain power sources and enhances system resilience.

[0077] Preferably, in S4, the multi-objective optimization model satisfies the following constraints: upper and lower limits of electric vehicle power, upper and lower limits of charging and discharging power, maximum load capacity of the distribution network, mutual exclusion of charging and discharging states, and the constraint that the state of charge (SOC) at the time of departure is not lower than the power required for the next trip.

[0078] In S5, the NSGA-III multi-objective optimization algorithm is used to solve the multi-objective optimization model, obtain the Pareto optimal solution set, and determine the charging and discharging power scheduling plan of the electric vehicle cluster from it. Attached Figure Description

[0079] To make the objectives, technical solutions, and advantages of the invention clearer, the invention will now be described in further detail with reference to the accompanying drawings, wherein:

[0080] Figure 1 This is a flowchart of the method;

[0081] Figure 2 This is a schematic diagram illustrating the relationship between the excitation factor and the discharge power in the embodiment.

[0082] Figure 3 This is a schematic diagram of the membership function in the embodiment;

[0083] Figure 4 This is a flowchart of the fuzzy inference process in the embodiment;

[0084] Figure 5 The flowchart shows the solution process of the NSGA-III algorithm in the embodiment. Detailed Implementation

[0085] The following detailed explanation illustrates the specific implementation methods:

[0086] Example:

[0087] like Figure 1 As shown in the figure, this embodiment discloses a method for real-time optimized charging and discharging of electric vehicles, including the following steps:

[0088] S1. Based on the travel chain theory and historical travel data, fit the probability distribution of electric vehicle trip start and end time, driving mileage and parking time, and construct a spatiotemporal characteristic model of electric vehicle charging demand to characterize the charging load distribution of electric vehicles at different time and space nodes and predict their charging behavior.

[0089] The spatiotemporal characteristic model of charging demand is constructed based on travel chain theory and includes:

[0090] Based on historical travel data, probability distribution models that follow a normal distribution are established for the start and end times, mileage, and parking time of electric vehicles.

[0091] The SOC at the end of the trip is determined based on the driving mileage, and the trip end time is taken as the time when the charging demand occurs; at the same time, a deterministic functional relationship between the SOC, battery capacity, charging power and charging efficiency is established for the charging duration.

[0092] The probability distribution model and the deterministic functional relationship are used together to characterize the charging load distribution of electric vehicles in the spatiotemporal dimension.

[0093] Unlike traditional methods that rely on fixed or average assumptions, this approach establishes a probabilistic distribution model for trip start and end times, mileage, and parking duration. This more realistically reflects the randomness and diversity of user travel behavior, avoiding load prediction biases caused by simplistic assumptions. By using the trip end time as the occurrence time of charging demand and establishing a deterministic function for charging duration based on SOC and battery parameters, the dynamic positioning and accurate estimation of charging events on the timeline are achieved, enhancing the model's temporal resolution and physical plausibility. Furthermore, combining the probabilistic model with the deterministic function not only captures the uncertainty of user behavior but also preserves the physical constraints of the charging process, giving the charging load distribution stronger spatial distribution characteristics and temporal evolution patterns, providing high-precision input for subsequent scheduling.

[0094] In practice, the spatiotemporal characteristics of electric vehicle charging demand are constructed based on electric vehicle travel data, including:

[0095] Obtain the start and end times of the electric vehicle's trip:

[0096]

[0097] In the formula, and These represent the start and end times of the j-th segment of the journey, respectively. and Let these represent the mean and variance of the start and end times of the j-th segment of the journey, respectively, with values ​​of [values ​​to be filled in].

[0098] Get the driving range of an electric vehicle:

[0099]

[0100] Always, μ dj and σ dj Let μ and μ represent the mean and variance of the mileage of the j-th segment, respectively. d1 =19.3, σ d1 =1.8; μ d2 =5.5, σ d2 =1.2; μ d3 =8.8, σ d3 =2.1.

[0101] Get the parking time of electric vehicles:

[0102]

[0103] In the formula, and Let A and B represent the mean and variance of the parking time after the j-th segment of the journey, respectively, with values ​​of 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19 ...

[0104] Further details on the charging time of electric vehicles:

[0105]

[0106] In the formula, T cj SOC represents the charging time after the j-th segment of the journey ends. Max This represents the battery's maximum state of charge; η cha P cha These are charging efficiency and charging power, respectively.

[0107] The driving process of an electric vehicle is composed of trip start and end times, parking time and driving distance, and changes in State of Charge (SOC), and integrates multiple factors using travel chain theory. Based on this, by analyzing travel chain data with spatiotemporal characteristics, electric vehicle charging load can be predicted.

[0108] S2. Based on the real-time voltage levels of each node in the distribution network, a regional dynamic electricity price model is constructed, and a discharge incentive price is superimposed during peak electricity consumption periods to form a coordinated electricity price guidance model. The regional dynamic electricity price model guides the spatial distribution of charging loads for electric vehicles. The discharge incentive price incentivizes electric vehicles with discharge capabilities to supply power to the distribution network during peak hours. The coordinated electricity price guidance model provides spatiotemporally differentiated charging and discharging price signals to reflect the grid operation status and user incentive levels in different regions and time periods.

[0109] Traditional time-of-use pricing mechanisms (as shown in Table 1) lack real-time response capabilities and are unable to effectively avoid the "peak-on-peak" problem caused by concentrated charging of electric vehicles. In contrast, dynamic pricing can be adjusted according to the real-time load status of the power grid, guiding the transfer of charging load and thus avoiding the overlap of peak electric vehicle charging times with peak distribution network load.

[0110] Table 1 Time-of-use electricity prices

[0111]

[0112] Given the time-varying nature of the distribution network load, and based on the initial time of electric vehicle connection to charging stations, the theoretical average load P in different time periods t is estimated. t th Set a benchmark electricity price c gd If P t th Below the expected average load for the whole day Electricity prices will be reduced based on the proportion of load shortage.

[0113] In practice, the zoned dynamic electricity price model guides electric vehicles to rationally distribute charging loads in space by increasing charging prices in areas where the voltage is lower than a preset low voltage value and decreasing charging prices in areas where the voltage is higher than a preset high voltage value.

[0114] The regional dynamic electricity pricing model includes:

[0115]

[0116] In the formula, represents the regional dynamic electricity price of the area where node k is located at time t; h represents the base electricity price adjustment coefficient; This represents the per-unit voltage value of the distribution network node where region k is located at time t; c represents the electricity price at time t; gd The benchmark electricity price; P t th This represents the estimated theoretical average load at time t; This indicates the expected average load.

[0117] The discharge excitation price for:

[0118]

[0119] ε t =1+tanh(g t |δ i,t,dis ·P i,t,dis |);

[0120] In the formula, ε i,t As the motivating factor; δ i,t,dis P is the discharge decision variable, taking a value of 1 when the electric vehicle discharges at time t, and 0 otherwise; i,t,dis This is the rated discharge power. t The excitation coefficient at period t represents the rate of change of the ratio of discharge power to excitation price in each time period, such as... Figure 2 As shown, the steeper the slope of the excitation factor curve, the higher the excitation benefit obtained from the discharge power, but all are controlled within a reasonable range of 1-2 times. The specific generation mechanism is as follows:

[0121] When an electric vehicle connects to the power grid, the control system automatically generates an arithmetic sequence starting from 0.15 and decreasing to 0.04 in fixed steps. The number of elements in this sequence is set to half the total number of scheduling periods within the response cycle. Based on the descending priority of the peak load of the distribution network in each time period, the system dynamically allocates the sequence elements to the corresponding excitation coefficient g for that time period. t g is not assigned a value t Set the value to 0.

[0122] Unlike traditional uniform or simple time-of-use pricing, this scheme directly couples voltage levels to price calculations, creating spatial price variability. Increasing charging prices in weak areas with low voltage effectively discourages concentrated charging in these areas, preventing localized overloads and voltage dips, and improving the stability of the distribution network. Furthermore, the regional dynamic pricing not only depends on voltage levels but also incorporates the trend of average system load changes (via P). t th and (Comparison) This model enables dynamic responses of electricity prices in both time and space dimensions. Compared to static or time-based pricing mechanisms, this model more sensitively reflects the actual operating status of the power grid, improving the accuracy and adaptability of dispatching. Furthermore, the discharge incentive price is not a fixed value, but rather varies with the user's discharge decision variable δ. i,t,dis Related to discharge power, through the excitation factor ε i,t This mechanism enables "on-demand incentives." It dynamically adjusts incentive levels to ensure higher returns for discharging during high-value periods, thereby significantly enhancing users' economic motivation to participate in vehicle-to-everything (V2X) interactions.

[0123] S3. Based on the current State of Charge (SOC) of the electric vehicle's battery and the amount of electricity required for the next leg of the journey, users are divided into rigid users and flexible users. Rigid users are those whose current SOC is less than a preset charging trigger value, or whose remaining battery power is less than a threshold value required for the next leg of the journey. The preset charging trigger value is 20%, but those skilled in the art can set it differently depending on the specific usage, which will not be elaborated upon here.

[0124] For flexible users, their willingness to discharge is quantified based on a fuzzy inference model to obtain a discharge willingness value. The discharge willingness value and battery degradation cost are jointly incorporated into the charging and discharging decision model of flexible users to weigh the discharge benefits and battery losses, and determine whether to participate in discharging and the corresponding discharge power level. The battery degradation cost is the cost obtained by economically converting the equivalent capacity decay caused by electric vehicles participating in discharging based on a battery cycle aging model applicable to vehicle-to-grid interaction scenarios.

[0125] In specific implementation, the fuzzy inference model takes the discharge incentive price, the current SOC and the parking duration as input variables. After fuzzification, fuzzy rule inference and defuzzification, it outputs the discharge response willingness value, which represents the degree of user's willingness to participate in discharge.

[0126] To effectively characterize the uncertainty of user discharge response behavior, this invention constructs a user discharge response intention model based on fuzzy reasoning theory. The specific modeling steps are as follows:

[0127] Blur processing:

[0128] First, the input quantities—incentive electricity price, remaining SOC, and parking time—are fuzzified to determine their respective fuzzy subsets. Based on practical experience, the incentive electricity price is divided into three fuzzy subsets: "low (L), medium (M), and high (H)"; the remaining SOC is divided into three fuzzy subsets: "insufficient (SM), moderate (MD), and sufficient (SF)"; and the parking time is divided into three fuzzy subsets: "short (S), medium (M), and long (L)". Based on these three inputs, a fuzzy decision model is constructed, outputting the discharge response intention P, and selecting low discharge response intention (L), low-medium discharge response intention (ML), medium discharge response intention (M), medium-high discharge response intention (MH), and high discharge response intention (H) as fuzzy subsets.

[0129] Establish membership function:

[0130] After inputting the above data, a corresponding membership function is assigned. The choice of membership function is usually determined based on statistical data or experience. Therefore, a Gaussian membership function is selected, and the membership degree corresponding to different fuzzy subsets is calculated using the following formula:

[0131]

[0132] Where σ and c are the standard deviation and mean, respectively, and the membership function curve of the input variable is shown in Figure 1. Figure 3 As shown.

[0133] Formulating fuzzy rules:

[0134] Based on the membership function and the analysis of user behavior characteristics, a fuzzy rule table was formulated, as shown in Table 2. To mathematically describe the fuzzy inference model and obtain the corresponding fuzzy implication relations, the fuzzy implication relations of the fuzzy controller are represented as the union of 27 fuzzy rules, as shown in the following equation:

[0135]

[0136] Table 2 Fuzzy rule table for discharge decision

[0137]

[0138] Fuzzy inference based on the Mamdani algorithm:

[0139] Assuming the sampling step size of the fuzzy controller is k (k = 0, 1, n), and using the Mamdani maximum-minimum synthesis method, the input fuzzy sets C(t), T(h), and the residual SOC are denoted as follows: The output probability is denoted as Where j = 1, 2, ... 27, the membership function corresponding to the fuzzy relation R is as follows:

[0140]

[0141] The fuzzy relations corresponding to the fuzzy rules are shown below:

[0142]

[0143] The corresponding output fuzzy value U can be obtained through fuzzy inference, as shown below:

[0144]

[0145] Deblurring:

[0146] To transform the fuzzy set of discharge response intentions output by fuzzy inference into precise values, a centroid method is used for fuzzification. This method determines the final precise discharge response intention value by calculating the centroid position of the region enclosed by the membership function curve and the horizontal axis. Its mathematical expression is as follows:

[0147]

[0148] in: The input is the Gaussian membership function; For y j The focus. A detailed flowchart is shown below. Figure 4 As shown.

[0149] In specific implementation, the battery cycle aging model is as follows:

[0150]

[0151] In the formula, α, β, γ, σ, ∈ are predetermined cyclic aging coefficients; The charge / discharge rate at time t; Ah i,t is the battery's ampere-hour throughput; K is the electric vehicle's cumulative equivalent cycle count within a time period, used to quantify the degree of battery cycle aging; E represents the cyclic aging loss per unit time for the i-th electric vehicle at time t; Bat The rated capacity of the battery; δ i,t,cha P is the charging decision variable, taking a value of 1 when the electric vehicle is charging at time t, and 0 otherwise; i,t,cha P represents the discharge power of an electric vehicle. i,t,dis V represents the rated discharge power; V is the battery's nominal voltage, assumed to remain constant during battery operation; Δt represents the charge / discharge time.

[0152] The battery degradation cost for:

[0153]

[0154] In the formula, BP and SP represent the purchase price and scrap price of the battery, respectively; E n This refers to the energy capacity at the end of the battery's life.

[0155] This enables the scientific quantification and economic calculation of the aging and loss of electric vehicle batteries during vehicle-to-grid interaction, effectively supporting users' rational decision-making between discharge benefits and battery life, and promoting the transformation of the V2G model from "short-term incentive-driven" to "long-term sustainable participation".

[0156] In specific implementation, the charging and discharging decision model for the flexible user includes a discharging decision sub-model and a charging decision sub-model;

[0157] The discharge decision sub-model is used to calculate the discharge benefits of elastic users, and combined with the discharge response intention value generated by the fuzzy inference model, it determines whether elastic users should discharge parametrically.

[0158] The charging decision sub-model is used to calculate the overall charging willingness of flexible users who do not participate in discharging, based on the degree of range anxiety and the influence factor of regional dynamic electricity price, and to determine whether to perform charging operation according to the preset threshold.

[0159] In this way, by setting up independent sub-models for discharging and charging decisions, the system can simultaneously address two groups: users participating in V2G and users who only need charging, avoiding the bias caused by "one policy for multiple purposes." This hierarchical decision-making mechanism significantly improves the system's adaptability to heterogeneous user groups. Furthermore, the discharging decision not only considers benefits but also integrates battery degradation costs and user willingness, enabling users to make rational choices that align with their own interests through a trade-off. The charging decision combines range anxiety and electricity price incentives to avoid risks associated with excessively delayed charging. This dual-protection mechanism increases user trust in the dispatching strategy and their willingness to participate long-term.

[0160] The discharge benefit is calculated using the following formula:

[0161]

[0162] In the formula, C income P represents the discharge gain; i,t,dis Rated discharge power; Δt is the charge / discharge time; η dis For discharge efficiency;

[0163] When C income When the value of the discharge response intention is greater than 0 and the value of the discharge response intention is greater than the preset intention value, the discharge decision sub-model determines that the elastic user participates in the discharge.

[0164] In this way, the discharge decision is based not only on whether the economic benefit is positive, but also on a dual judgment based on the user's subjective willingness generated by fuzzy reasoning. This composite criterion of "objective benefit + subjective willingness" effectively avoids the risk of users refusing to respond due to temporary emotions or non-economic factors, and improves the success rate of dispatching instructions. It enables refined economic assessment and intelligent decision-making of electric vehicle discharge behavior, effectively balancing the relationship between user economic benefits and battery health, promoting the transformation of vehicle-to-grid interaction from "mandatory participation" to "voluntary response," and improving the flexibility of system regulation and user satisfaction.

[0165] The charging decision sub-model calculates the comprehensive charging intention W of elastic users who do not participate in discharging using the following formula. i,t :

[0166] W i,t =β i,t ·R i,t +(1-β i,t )·η t ;

[0167]

[0168] In the formula, β i,t R represents the weighting coefficients; i,t Range anxiety level; η t The factors influencing regional dynamic electricity prices; and These represent the maximum and minimum values ​​of the dynamic electricity price for each zone; when the user's overall charging intention W... i,t If the value is greater than 0.5, the user selects to charge; otherwise, charging is not performed. (SOC) Max The maximum state of charge (SOC) of the i-th electric vehicle. i,t Let D be the real-time SOC of the i-th electric vehicle at time t; i α represents the distance an electric vehicle travels to its next destination. ev This refers to the energy consumption per unit distance of an electric vehicle.

[0169] Unlike traditional methods that rely solely on SOC thresholds or electricity price signals as charging triggers, this approach weights and integrates two core driving factors—range anxiety and electricity price—to construct a comprehensive charging willingness index. This model more comprehensively reflects users' psychological state and economic considerations in real-world travel scenarios, enhancing the scientific rigor of charging decisions. It enables refined and intelligent guidance of flexible user charging behavior, effectively balancing user travel safety needs with grid economic operation goals, promoting the transformation of electric vehicles from passive loads to proactively responding resources, and improving the overall coordination and sustainability of the energy system.

[0170] Based on the spatiotemporal characteristic model of charging demand in S4, the electricity price coordination guidance model in S2, and the charging and discharging decision model in S3, a multi-objective optimization model is constructed with the goals of minimizing total user cost, minimizing distribution network load variance, and maximizing overall user satisfaction.

[0171] In specific implementation, the objective function of the multi-objective optimization model is:

[0172] min{F A (χ i ),-F B (χ i ),F C (χ i )};

[0173] In the formula, F A (χ i Let F be the economic cost of charging and discharging the i-th electric vehicle. B (χ i F represents the user satisfaction with the charging needs of the i-th electric vehicle. C (χ i ) represents the power distribution network load fluctuation of the i-th electric vehicle;

[0174]

[0175] α n1 =min{SOC f,i,e / SOC f,i,exp ,1};

[0176] α n2 =min{1 / (1+η)} i,charge ),1};

[0177]

[0178] In the formula, N is the total number of electric vehicles; T is the scheduling time period; λ1 and λ2 are weighting coefficients; α n1 Indicates user charging satisfaction; α n2 Indicates battery wear satisfaction; SOC f,i,e This indicates the electric vehicle's battery level when the user arrives at the charging station; SOC f,i,exp Indicates the user's expected battery level when leaving; η i,charge This represents the number of charge / discharge transitions for the i-th electric vehicle during the scheduling process.

[0179] Unlike traditional single-objective methods that focus solely on grid peak shaving or minimizing user costs, this scheme integrates user economic costs, charging satisfaction, and grid load fluctuations into a unified objective function, achieving a balance of interests among multiple parties through a multi-objective optimization framework. This systematic design avoids the problem of "paying for one thing at the expense of another," significantly improving the overall coordination of the scheduling strategy. Furthermore, the user satisfaction index not only includes whether the arrival power meets expectations (charging satisfaction) but also incorporates considerations of battery degradation (the impact of charge-discharge cycles), ensuring that the system guides user responses while respecting their concerns about battery life. This mechanism effectively alleviates user resistance to frequent V2G operations and enhances long-term participation. Moreover, by minimizing the load fluctuation objective F... C This model can proactively suppress load peak-valley differences, reducing voltage fluctuations and equipment overload risks. Especially against the backdrop of large fluctuations in renewable energy output, this smoothing effect helps improve the distribution network's ability to accommodate uncertain power sources and enhances system resilience.

[0180] The objective optimization model satisfies the following constraints: upper and lower limits of electric vehicle battery capacity, upper and lower limits of charging and discharging power, maximum load capacity of the distribution network, mutual exclusion of charging and discharging states, and the constraint that the state of charge (SOC) at the time of departure must not be lower than the battery capacity required for the next trip; specifically as follows:

[0181] Electric vehicle power constraints:

[0182]

[0183] In the formula, SOC i,t This represents the battery capacity at time t; Indicates the minimum and maximum battery capacity;

[0184] Charge and discharge power constraints:

[0185] 0≤P i,t,cha ≤P ch,max ;

[0186] 0≤P i,t,dis ≤P dis,max ;

[0187] Where: P ch,max and P dis,max These represent the maximum charging and discharging power of the electric vehicle, respectively.

[0188] Maximum load constraint of distribution network:

[0189]

[0190] in: This indicates the maximum load that the distribution network can withstand.

[0191] Electric vehicle charge / discharge state constraints:

[0192] δ i,t,cha ·δ i,t,dis =0;

[0193] Departure time SOC constraint:

[0194]

[0195] S5. The NSGA-III multi-objective optimization algorithm is used to solve the multi-objective optimization model to obtain the Pareto optimal solution set, and the charging and discharging power scheduling plan of the electric vehicle cluster is determined from it to guide each electric vehicle to perform real-time charging and discharging operations.

[0196] For the objective function described above in this invention, the NSGA-III algorithm is used for solution. The NSGA-III algorithm is an evolutionary algorithm proposed for high-dimensional multi-objective optimization problems. Its core lies in guiding population distribution through reference points to maintain diversity while ensuring solution convergence. This algorithm inherits the elitist strategy and non-dominated sorting mechanism of NSGA-II in its framework, but introduces reference points and normalization processing in the selection mechanism, significantly improving the optimization performance in high-dimensional objective spaces. Therefore, the NSGA-III algorithm is used to solve multi-objective optimization problems. The flowchart of the NSGA-III algorithm is as follows... Figure 5 As shown, the specific steps are as follows:

[0197] (1) Initialization and merging of populations

[0198] The algorithm first generates an initial parent population, then produces a child population through simulated binary crossover and polynomial mutation, and merges the parent and child populations to form a mixed population of size 2N. Subsequently, a fast non-dominated sorting algorithm is used to stratify the mixed population, forming multiple non-dominated fronts.

[0199] (2) Reference point generation mechanism

[0200] In the M-dimensional target space, a structured reference point set is generated through uniform sampling. The number of reference points H is determined by the target dimension M and the number of segments p, and is calculated using the following formula:

[0201]

[0202] (3) Adaptive normalization processing

[0203] The objective function value is adaptively normalized using the intercepts of the linear hyperplane and the coordinate axes, as shown in the following formula:

[0204]

[0205] Where: f i(x) represents the target value of the i-th target; a is the minimum value of the i-th objective; i f is the intercept of the i-th target. i n (x) represents the standard value of the i-th objective of the n-th individual.

[0206] (4) Reference point association strategy

[0207] For each normalized solution f i n (x) connects the reference point to the origin on the linear hyperplane to define the reference line corresponding to each reference point, and calculates the perpendicular distance from the individual to each reference direction. Then, it is associated with the nearest reference point and marked as belonging to that reference point.

[0208] (5) Microhabitat conservation selection

[0209] Individuals are selected layer by layer from the non-dominated layers until the population capacity N is exceeded. When some layers need to be truncated, individuals with fewer associations with the reference point are selected first. The specific process includes:

[0210] a) Calculate the number of times the reference point is associated, ρ j Sort the reference points in ascending order by their number of associations and process them one by one.

[0211] b) For ρ j If the reference point is 0, select the solution closest to it; otherwise, randomly select the nearest solution.

[0212] c) Repeat until the population is full, ensuring that the population is evenly distributed in the target space.

[0213] (6) Iteration and Termination

[0214] Merge the offspring population with the parent population and proceed to the next generation optimization. Repeat steps (2) to (5) until the preset maximum number of iterations or the convergence condition is reached, and output the Pareto optimal solution set.

[0215] Unlike existing methods that often employ static or averaged travel assumptions, this scheme uses travel chain theory and historical data to fit the probability distributions of trip start and end times, mileage, and parking duration, constructing a charging demand model with individual differences and spatiotemporal dynamics. This model can more realistically reflect users' actual travel patterns, significantly improving the prediction accuracy of electric vehicle cluster charging load in both time and space dimensions, providing reliable input for subsequent scheduling. Furthermore, unlike traditional uniform time-of-use pricing or simple time-of-use pricing mechanisms, this scheme innovatively combines the real-time voltage levels of each node in the distribution network to construct a zoned dynamic pricing model, superimposed with peak-hour discharge incentive pricing. This mechanism not only guides users to avoid charging in areas with weak grid connections, alleviating local overload and voltage exceedance issues, but also actively incentivizes vehicles with V2G capabilities to supply power in reverse during critical periods, effectively suppressing peak-on-peak phenomena and enhancing the resilience of the distribution network.

[0216] Existing scheduling strategies often treat users as a homogeneous group, ignoring the differences in their travel rigidity and response flexibility. This solution divides users into rigid and flexible categories based on SOC and travel demand, and introduces a fuzzy inference model for flexible users to quantify their discharge intentions, making scheduling decisions more aligned with users' actual psychological and behavioral characteristics. Compared to a "one-size-fits-all" scheduling approach, this method significantly enhances user acceptance and response reliability. Furthermore, addressing the common neglect of the impact of V2G on battery aging in existing technologies, this solution incorporates battery degradation costs based on cyclic aging mechanisms into the decision-making model for flexible users, allowing them to consider long-term usage costs while evaluating discharge benefits. This mechanism avoids user resistance due to over-discharge, maintaining reasonable economic incentives while ensuring battery health and improving the sustainability of the V2G model. Moreover, this solution simultaneously optimizes three objectives: total user cost, distribution network load variance, and overall user satisfaction, generating a charging and discharging scheduling plan that balances technical feasibility and user experience through multi-objective solutions. This systematic and collaborative design effectively bridges the contradiction between grid regulation needs and individual user interests.

[0217] This method can achieve a synergistic improvement in power grid safety, user economic costs, and overall satisfaction.

[0218] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit the technical solutions. Those skilled in the art should understand that any modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the present invention should be covered within the scope of the claims of the present invention.

Claims

1. A method for real-time optimization of charging and discharging of an electric vehicle, characterized in that, The method comprises the following steps: S1, based on the trip chain theory and historical travel data, fitting the probability distribution of the start and end time, driving distance and parking time of the electric vehicle trip, and constructing a space-time characteristic model of the charging demand of the electric vehicle, which is used to represent the charging load distribution of the electric vehicle at different time and space nodes, and predict the charging behavior; S2, based on the real-time voltage level of each node of the distribution network, a zoned dynamic electricity price model is constructed, and a discharge incentive electricity price is superimposed in the peak electricity consumption period, thereby forming an electricity price collaborative guidance model; wherein the zoned dynamic electricity price model is used to guide the electric vehicle to reasonably distribute the charging load in space; the discharge incentive electricity price is used to encourage the electric vehicle with discharge capability to supply power to the distribution network in reverse during the peak period; the electricity price collaborative guidance model is used to provide time and space differentiated charging and discharging price signals, which reflect the power grid operation state and user incentive level in different regions and periods; S3, according to the current SOC of the electric vehicle and the required power of the next trip, the user is divided into rigid users and elastic users; wherein the rigid user is the current SOC less than the preset charging trigger value, or the remaining power less than the required power threshold of the next trip; For the elastic user, the discharge response willingness is quantified based on the fuzzy reasoning model to obtain a discharge willingness value; the discharge willingness value and the battery degradation cost are jointly included in the charging and discharging decision model of the elastic user, which is used to weigh between the discharge benefit and the battery loss to determine whether to participate in the discharge and the corresponding discharge power level; wherein the battery degradation cost is the cost obtained by economically converting the equivalent capacity attenuation of the electric vehicle caused by participating in the discharge according to the battery cycle aging model applicable to the vehicle-to-grid interaction scenario; S4, based on the space-time characteristic model of the charging demand in S1, the electricity price collaborative guidance model in S2, and the charging and discharging decision model in S3, a multi-objective optimization model is constructed, which takes the minimum total cost of the user, the minimum load variance of the distribution network, and the highest comprehensive satisfaction of the user as the target; S5, the multi-objective optimization model is solved to determine the charging and discharging power scheduling plan of the electric vehicle cluster, which is used to guide each electric vehicle to perform real-time charging and discharging operation.

2. The method for real-time optimization of charging and discharging of an electric vehicle according to claim 1, wherein: In S1, the space-time characteristic model of the charging demand is constructed based on the trip chain theory, comprising: Based on the historical travel data, the trip start and end time, driving distance and parking time of the electric vehicle are established to follow the normal distribution probability distribution model; Based on the driving distance, the SOC at the end of the trip is determined, and the end time of the trip is taken as the occurrence time of the charging demand; at the same time, the SOC, battery capacity, charging power and charging efficiency are used to establish a deterministic function relationship of the charging duration; Wherein, the probability distribution model and the deterministic function relationship are used to represent the charging load distribution of the electric vehicle in the time and space dimensions.

3. The method for real-time optimization of charging and discharging of an electric vehicle of claim 1, wherein: In S2, the zoned dynamic electricity price model is used to guide the electric vehicle to reasonably distribute the charging load in space by increasing the charging electricity price in the area where the voltage is lower than the preset low voltage value, and reducing the charging electricity price in the area where the voltage is higher than the preset high voltage value; The zoned dynamic electricity price model comprises: wherein, denotes the partition dynamic electricity price of the area where node k is located at time t; h denotes the basic electricity price adjustment coefficient; denotes the voltage per unit of the power distribution network node where area k is located at time t; denotes the electricity price at time t; c gd is the reference electricity price; P t th denotes the estimated theoretical average load at time t; denotes the expected average load; The discharge excitation voltage Is: ε t = 1 + tanh(gt|δ i,t,dis · P i,t,dis |); where ε i,t is the incentive factor; g t is the incentive coefficient at period t, representing the rate of change of the ratio of discharging power to incentive price; δ i,t,dis is the discharging decision variable, taking value 1 when the electric vehicle discharges at t, otherwise 0; P i,t,dis is the rated discharging power.

4. The method for real-time optimization of charging and discharging of an electric vehicle of claim 3, wherein: In S3, the fuzzy inference model takes the discharge incentive price, the current SOC and the parking time as input variables, and outputs a discharge response willingness value representing the degree of willingness of the user to participate in the discharge after fuzzy processing, fuzzy rule inference and defuzzy processing.

5. The method for real-time optimization of charging and discharging of an electric vehicle of claim 4, wherein: In S3, the battery cycle aging model is: wherein a, β, γ, σ, ∈ are predetermined cycle aging coefficients; is the charge and discharge rate at time t; Ah i,t is the ampere-hour throughput of the battery; K is the cumulative equivalent cycle number of the electric vehicle in the time period, which is used to quantify the cycle aging degree of the battery; represents the cycle aging loss per unit time of the i-th electric vehicle at time t; E Bat is the rated capacity of the battery; δ i,t,cha is the charging decision variable, which takes the value 1 when the electric vehicle is charging at time t, and 0 otherwise; P i,t,cha denotes the discharging power of the electric vehicle; P i,t,dis is the nominal discharging power; V is the nominal voltage of the battery, which is assumed to remain constant during the operation of the battery; Δt denotes the charging and discharging time; The battery degradation cost Is: where BP and SP are the purchase price and the scrap price of the battery, respectively; E n is the energy capacity at the end of the battery life.

6. The method for real-time optimization of charging and discharging of an electric vehicle of claim 5, wherein: In S3, the charging and discharging decision model of the elastic user comprises a discharging decision sub-model and a charging decision sub-model. The discharging decision sub-model is used to calculate the discharging benefit of the elastic user, and in combination with the discharge response willingness value generated by the fuzzy inference model, to determine whether the elastic user participates in the discharge. The charging decision sub-model is used to calculate the comprehensive charging willingness of the elastic user who does not participate in the discharge based on the mileage anxiety degree and the partition dynamic price influence factor, and to determine whether to perform the charging operation according to a preset threshold.

7. The method for real-time optimization of charging and discharging of an electric vehicle of claim 6, wherein: The discharging benefit is calculated by the following formula: In the formula, C income represents the discharge benefit; P i,t,dis is the rated discharge power; and Δt is the charge and discharge time. η dis is the discharge efficiency; When C income When C > 0 and the discharge response willingness value is greater than the preset willingness value, the discharge decision sub-model judges that the flexible user participates in the discharge.

8. The method for real-time optimization of charging and discharging of an electric vehicle of claim 7, wherein: The charging decision sub-model calculates the comprehensive charging willingness W of the elastic user who does not participate in discharging by the following formula i,t : W i,t = β i,t · R i,t + (1 - β i,t ) · η t ; where β i,t is the weight coefficient; R i,t is the range anxiety; η t is the partition dynamic electricity price influence factor; and are the maximum and minimum values of the partition dynamic electricity price, respectively; when the user's comprehensive charging willingness W i,t > 0.5, the user chooses to charge, otherwise not to charge; SOC Max is the maximum state of charge of the i-th electric vehicle; SOC i,t is the real-time SOC of the i-th electric vehicle at time t; D i is the driving distance of the electric vehicle to the next trip destination; α ev is the unit mileage energy consumption of the electric vehicle.

9. The method for real-time optimization of charging and discharging of an electric vehicle of claim 8, wherein: In S4, the objective function of the multi-objective optimization model is: min{F A (x i ),-F B (x i ),F C (x i )}; In the formula, F A (χ i ) is the charging and discharging economic cost of the i-th electric vehicle, F B (χ i ) is the user charging demand satisfaction of the i-th electric vehicle, F C (χ i ) is the power distribution network load fluctuation of the i-th electric vehicle; a n1 = min(SOC f,i,e / SOC f,i,exp ,1} ; α n2 = min{1 / (1+η i,charge ),1}; where N is the total number of electric vehicles; T is the scheduling time period; λ1, λ2 are weight coefficients; α n1 represents the user charging satisfaction; αn2 represents the battery loss satisfaction; SOC f,i,e represents the amount of electricity of the electric vehicle when the user arrives at the charging station; SOC f,i,exp represents the expected amount of electricity when the user is expected to leave; η i,charge represents the number of charge-discharge conversions of the i-th electric vehicle in the scheduling process.

10. The method for real-time optimization of charging and discharging of an electric vehicle of claim 1, wherein: In S4, the multi-objective optimization model satisfies the following constraint conditions: the upper and lower limits of the electric vehicle power, the upper and lower limits of the charging and discharging power, the maximum load capacity of the distribution network, the charging and discharging state exclusion constraint, and the constraint that the SOC at the departure time is not lower than the required power for the next trip; In S5, the NSGA-III multi-objective optimization algorithm is used to solve the multi-objective optimization model to obtain a Pareto optimal solution set, and the charging and discharging power scheduling plan of the electric vehicle cluster is determined from the set.

Images