User charging and battery swapping behavior mechanism identification method and system based on multi-source data fusion

Abstract

This invention discloses a method and system for identifying user charging and swapping behavior mechanisms based on multi-source data fusion. The method preprocesses heterogeneous multi-source data to construct a multi-dimensional feature vector set representing user charging and swapping behavior; it constructs a causal loop model including a charging / swapping time selection submodule and a charging / swapping location selection submodule; it establishes a cross-module coupling mechanism between the two submodules and corrects location flexibility through a negative constraint function to reflect the time-based rigidity limiting the spatial selection range; based on the causal loop model, it performs parameter sensitivity analysis to identify key guideable variables such as electricity price structure, facility distribution, and meteorological response, generating differentiated user behavior guidance strategies. This invention can accurately identify key guideable variables affecting user behavior, providing a scientific basis for formulating refined and flexible behavior guidance strategies, thereby improving the spatiotemporal adjustability of user behavior.

Description

Technical Field

[0001] This invention relates to the field of mechanism identification technology, and in particular to a method and system for identifying the mechanism of user charging and swapping behavior based on multi-source data fusion. Background Technology

[0002] With the rapid growth in the number of new energy vehicles in my country, user charging and swapping behaviors exhibit high concentration and significant differentiation in both time and space. Influenced by multiple factors such as vehicle status, charging infrastructure supply, traffic patterns, grid capacity, and weather conditions, charging and swapping demand displays complex coupling characteristics, posing operational risks such as load aggregation and equipment overload to local distribution networks. Faced with the ever-increasing charging and swapping load, traditional methods relying on expanding distribution network hardware involve huge investments, long development cycles, and are unlikely to remain effective in the long term under the trend of continuous load growth. Therefore, in the context of promoting vehicle-grid interaction, accurately identifying user behavior characteristics and transforming user load into a flexible, guideable, and adjustable resource has become a key requirement for ensuring distribution network stability, alleviating local supply-demand imbalances, and improving system operating efficiency.

[0003] However, existing research generally relies on a single data source, lacking effective integration of heterogeneous information from multiple sources such as vehicles, charging stations, roads, networks, and weather. This leads to insufficient understanding of the dynamic feedback mechanisms behind user behavior, particularly failing to accurately depict the coupling relationship between "time selection" and "spatial selection" in charging and swapping. This results in insufficient accuracy and interpretability of behavior identification, thus limiting the effectiveness of policy guidance and hindering the full exploitation of user-side flexibility, making it difficult to support the operational needs of scenarios with a higher proportion of new energy vehicles. Therefore, it is necessary to construct a novel modeling method that integrates multi-source data and can reveal the intrinsic logic of behavior.

[0004] With the rapid growth of electric vehicle ownership, users exhibit highly complex charging and swapping behaviors under different time, space, and energy demand conditions. These significant differences and uncertainties have become important factors affecting the stable operation of the power grid and the flexibility of load regulation. Therefore, in-depth analysis of the formation mechanism of charging and swapping behavior is of significant theoretical and engineering value for achieving orderly guidance on the user side, optimizing load management, and supporting the construction of new power systems.

[0005] Current research has made progress in characterizing user behavior. For example, Wang Xingyu's "Analysis of New Energy Vehicle Charging Behavior Based on Highway Service Area Charging Pile Data" uses Beta regression model to mine typical charging patterns and identify the main quantitative factors influencing user choices based on highway charging pile data; He Linru's "Research on Electric Vehicle Charging Behavior Characteristics Based on Clustering Algorithm for Different Building Types" uses K-means and DBSCAN clustering methods to extract typical load characteristics of residential and commercial scenarios from building-side data; and Yu Lei's "Optimization Method for Ordered Charging of Electric Vehicles Based on Deep Reinforcement Learning" uses deep reinforcement learning to achieve dynamic scheduling optimization of charging behavior based on historical charging data.

[0006] However, existing research generally relies on single-source charging data, failing to incorporate key factors such as road conditions, weather changes, parking conditions, and travel purpose into the analysis. Therefore, it is difficult to reveal the deep mechanisms underlying behavioral formation across dimensions. Traditional methods often focus on statistical induction or pattern recognition of behavioral representations, failing to depict the dynamic causal chains between multiple factors and struggling to reflect the feedback relationships between the vehicle-charging station-road-network-weather system. In fact, users' charging and swapping decisions are simultaneously influenced by multiple factors such as traffic flow, temperature, and parking convenience, and these correlations are often overlooked in studies driven by single-source data.

[0007] In the absence of multi-source information fusion, existing behavioral analysis methods are significantly inadequate in terms of explanatory accuracy, causal reliability, and guidance effectiveness. Summary of the Invention

[0008] To achieve a systematic and quantitative analysis of user charging and swapping behavior mechanisms and provide a scientific basis for charging and swapping network planning, power grid control strategies, and flexible load management, this invention proposes a user charging and swapping behavior mechanism identification method based on multi-source data fusion. This method can systematically reveal the interaction between behavioral driving factors, user preferences, and cost-benefit relationships, providing an interpretable, controllable, and verifiable scientific basis for precise guidance strategy formulation, operation optimization, and policy design, thereby fundamentally improving the orderly management level of charging and swapping behavior.

[0009] Another objective of this invention is to propose a user charging and swapping behavior mechanism identification system based on multi-source data fusion.

[0010] To achieve the above objectives, a first aspect of the present invention proposes a method for identifying user charging and swapping behavior mechanisms based on multi-source data fusion, comprising: After preprocessing the multi-source heterogeneous data, a set of multi-dimensional feature vectors representing users' charging and swapping behavior is constructed. The multidimensional feature vector set is input into the system dynamics modeling framework to construct a causal loop model that includes a charging / swapping time selection submodule and a charging / swapping location selection submodule. A cross-module coupling mechanism is established between the charging / swapping time selection submodule and the charging / swapping location selection submodule, and the location flexibility is corrected by a negative constraint function to reflect the limitation of time period rigidity on the spatial selection range. Based on the causal loop model, parameter sensitivity analysis is performed to identify key guideable variables such as electricity price structure, facility distribution, and meteorological response, and to generate differentiated user behavior guidance strategies.

[0011] In one embodiment of the present invention, after preprocessing multi-source heterogeneous data, a multi-dimensional feature vector set characterizing user charging and swapping behavior is constructed, including: Raw time-series data were collected from vehicles, charging and swapping facilities, transportation, power grids, and meteorological systems. The original time-series data is uniformly aligned on time scale, missing values ​​are imputed, and noise filtering is performed using exponential smoothing. Based on the preprocessed data, a multi-dimensional unified feature vector containing battery capacity, state of charge, traffic flow, time-of-use electricity price, and temperature is constructed for each user's charging and swapping behavior to form a multi-dimensional feature vector set.

[0012] In one embodiment of the present invention, the multidimensional feature vector set is input into a system dynamics modeling framework to construct a causal loop model including a charging / swapping time selection submodule and a charging / swapping location selection submodule, comprising: Based on the multidimensional feature vector set, a charging and swapping time selection submodule is constructed. By defining charging and swapping demand, adjustable power capacity and time period flexibility state variables, three types of feedback loop mechanisms are quantified: grid price-demand suppression, battery status + weather-preference driving, and travel characteristics + weather-guided constraint. A charging and swapping location selection submodule is constructed. By defining state variables of facility accessibility, price utility, and location flexibility, three types of feedback loop mechanisms are quantified: pile-accessibility-location flexibility, network-price-location selection, and weather-site attractiveness-accessibility correction. The charging / swapping time selection submodule and the charging / swapping location selection submodule are integrated with system dynamics to form an overall causal loop model structure that includes multi-loop causal chains and the interaction relationship of state variables.

[0013] In one embodiment of the present invention, a cross-module coupling mechanism is established between the charging / swapping time selection submodule and the charging / swapping location selection submodule, and a negative constraint function is used to correct the location flexibility to reflect the limitation of time period rigidity on the spatial selection range, including: Based on the time period flexibility variable output by the charging and swapping time selection submodule and the location flexibility variable output by the charging and swapping location selection submodule, the negative correlation coupling relationship between the two is determined to characterize the constraint effect of time period rigidity on the spatial selection range. Construct a spatiotemporal coupling negative constraint function, substitute the time period flexibility variable as the independent variable, and calculate the correction coefficient for location flexibility; The original position flexibility variables are weighted and corrected using the correction coefficients, and the coupled and corrected position flexibility variables are output to complete cross-module linkage modeling.

[0014] In one embodiment of the present invention, parameter sensitivity analysis is performed based on the causal loop model to identify key guideable variables such as electricity price structure, facility distribution, and meteorological response, and to generate differentiated user behavior guidance strategies, including: Based on the complete set of parameters of the causal loop model, parameter perturbation schemes and sensitivity evaluation indicators are designed to quantify the degree of influence of each parameter on behavioral flexibility. The parameter sensitivity analysis calculation was performed. By systematically adjusting the electricity price elasticity coefficient, facility density weight, and meteorological sensitivity coefficient, the response changes of flexibility and location flexibility in different time periods were evaluated, and the sensitivity ranking results were obtained by ranking the degree of influence of the parameters. Based on the sensitivity ranking results, the key inductive variables are identified as electricity price structure, public pile coverage density, and meteorological response factor. Differentiated user behavior guidance strategies are generated, including the adjustment range of time-of-use electricity prices, the direction of facility spatial layout optimization, and the meteorological early warning linkage mechanism.

[0015] To achieve the above objectives, a second aspect of the present invention proposes a user charging and swapping behavior mechanism identification system based on multi-source data fusion, comprising: The feature vector construction module is used to construct a set of multi-dimensional feature vectors that characterize user charging and swapping behavior after preprocessing multi-source heterogeneous data. The causal loop model construction module is used to input the multidimensional feature vector set into the system dynamics modeling framework to construct a causal loop model that includes a charging / swapping time selection submodule and a charging / swapping location selection submodule. The coupling mechanism construction module is used to establish a cross-module coupling mechanism between the charging and swapping time selection submodule and the charging and swapping location selection submodule, and to correct the location flexibility through a negative constraint function to reflect the limitation of time period rigidity on the spatial selection range. The sensitivity analysis module is used to perform parameter sensitivity analysis based on the causal loop model, identify key guideable variables such as electricity price structure, facility distribution and meteorological response, and generate differentiated user behavior guidance strategies.

[0016] This invention presents a method and system for identifying user charging and swapping behavior mechanisms based on multi-source data fusion. By integrating heterogeneous data from vehicles, charging piles, roads, networks, and meteorology, it constructs a multi-loop causal relationship model, enabling systematic analysis and quantitative characterization of user charging and swapping time and location selection behaviors. This invention can accurately identify key guideable variables influencing user behavior, providing a scientific basis for developing refined and flexible behavior guidance strategies. This improves the spatiotemporal adjustability of user behavior, alleviates grid pressure caused by disorderly charging, promotes optimized allocation of energy resources and rational network layout, and provides technical support for building a safe, economical, and efficient electric vehicle charging mode.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

[0018] The technical effects of this invention are as follows: 1. Improve the accuracy and dynamic response capability of user behavior mechanism identification. The multi-source data fusion system constructed in this invention breaks through the limitations of traditional reliance on single charging data, realizes the joint processing of multi-dimensional information such as vehicle, charging pile, road, network and weather, effectively eliminates the data silo problem, and makes the identification of user charging and swapping behavior mechanism more comprehensive, precise and dynamic, significantly improving the model's fitting degree and explanatory power to real behavior.

[0019] 2. Achieve quantitative characterization of multi-factor interactions and enhance simulation and prediction capabilities. By employing system dynamics modeling and causal loop analysis, user behavior is decomposed into a dual decision-making process in time and space. A structured model with multiple feedback loops is constructed, which can quantify the interaction paths between different factors, making the behavioral evolution trend simulable and predictable, which has significant advantages over traditional static analysis methods.

[0020] 3. Identify key guideable variables to improve the effectiveness of precise intervention and resource utilization efficiency. Through parameter sensitivity analysis and causal chain quantification, this invention can identify key guideable variables that affect user behavior and clarify their direction and intensity of influence, providing a direct basis for designing differentiated behavior guidance strategies, thereby improving the utilization efficiency of charging and swapping resources and reducing system congestion.

[0021] 4. Achieve cross-module collaborative modeling of time-space coupled behavior, avoiding the one-sidedness of traditional methods. This invention is the first to incorporate the time period selection and location selection of user charging and swapping behavior into a unified system dynamics framework, revealing the bidirectional coupling mechanism between the two. This overcomes the bias caused by the separate modeling of time and space factors in existing studies, making the obtained guidance strategy more consistent with real decision-making logic.

[0022] 5. The method is highly versatile and can be extended to various types of transportation-energy coupling behaviors. The multi-source data fusion method, causal loop structure, and sensitivity analysis mechanism of this invention all have good scalability and can be applied to various cross-domain research scenarios of transportation and energy, such as electric vehicle travel behavior, road energy supply network planning, and transportation-power grid coordinated control, demonstrating high versatility and application value. Attached Figure Description

[0023] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart illustrating a method for identifying user charging and swapping behavior mechanisms based on multi-source data fusion, provided in an embodiment of the present invention; Figure 2 Causal loop diagram of the charging / swapping time selection submodule and the location selection submodule provided in the embodiments of the present invention; Figure 3 Flow diagrams for the charging / swapping time selection submodule and the location selection submodule provided in this embodiment of the invention; Figure 4 This is a flowchart illustrating the user charging and swapping behavior mechanism identification method based on multi-source data fusion provided in this embodiment of the invention. Figure 5 This is a structural diagram of a user charging and swapping behavior mechanism identification system based on multi-source data fusion, provided in an embodiment of the present invention. Detailed Implementation

[0024] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0025] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0026] The following describes, with reference to the accompanying drawings, a method and system for identifying user charging and swapping behavior mechanisms based on multi-source data fusion, according to an embodiment of the present invention.

[0027] This embodiment provides a method for identifying user charging and swapping behavior mechanisms based on multi-source data fusion. For example... Figure 1 As shown, it includes: S1. After preprocessing the multi-source heterogeneous data, construct a set of multi-dimensional feature vectors representing the user's charging and swapping behavior; S2, input the set of multidimensional feature vectors into the system dynamics modeling framework to construct a causal loop model that includes a charging / swapping time selection submodule and a charging / swapping location selection submodule; S3. Establish a cross-module coupling mechanism between the charging / swapping time selection submodule and the charging / swapping location selection submodule, and correct the location flexibility through a negative constraint function to reflect the limitation of time period rigidity on the spatial selection range. S4. Based on the causal loop model, perform parameter sensitivity analysis to identify key guideable variables such as electricity price structure, facility distribution, and meteorological response, and generate differentiated user behavior guidance strategies.

[0028] Understandably, to address the problems in existing technologies such as insufficient analysis of factors influencing charging and swapping behavior, difficulty in quantifying the coupling relationships between variables, and lack of systematic and guideable factor identification, this invention proposes a method for identifying the mechanism of user charging and swapping behavior based on multi-source data fusion. This method utilizes system dynamics modeling to construct a multi-loop causal circuit diagram (as attached). Figure 2 As shown in the figure, this invention performs structured analysis on the potentially complex causal chains in multi-source heterogeneous data from vehicles, charging piles, roads, networks, and meteorology, enabling a quantitative characterization of the influence mechanisms of users' charging and swapping time and location choices. Simultaneously, this invention combines causal network analysis to automatically identify key, controllable variables with regulatory value; the logical connections between various influencing factors are shown in the appendix. Figure 3 As shown. Through the above modeling and analysis process, this invention can scientifically reveal the deep-seated mechanistic characteristics of user behavior, providing a verifiable and quantifiable theoretical basis for optimizing energy resource allocation, designing price incentive mechanisms, and formulating behavior guidance strategies. The specific process of the method described in this invention is attached. Figure 4 As shown, the specific steps are as follows.

[0029] Preferably, to accurately identify the mechanisms of user charging and swapping behavior, this invention collects multi-source data from vehicles, charging and swapping facilities, transportation, power grids, and meteorological systems, specifically including: (1) Vehicle data: Battery capacity Battery health status Current state of charge Unit energy consumption ; (2) Charging pile data: charging and swapping power Facility type ; (3) Road data: traffic flow Parking duration Travel mileage ; (4) Network data: Time-of-use electricity pricing Incentive prices Service fee Area capacity ; (5) Meteorological data: temperature Rainfall .

[0030] To ensure consistency and usability of heterogeneous data from multiple sources, including vehicles, infrastructure, roads, networks, and meteorology, during the modeling process, this invention performs uniform preprocessing operations on the collected raw data, including time scale alignment, missing value completion, and noise filtering. Given the large volume, high update frequency, and significant volatility of the multi-source data, this invention employs exponential smoothing to smooth the original time series data, effectively reducing random noise while preserving the main trend characteristics of the data. The specific formula is as follows:

[0031] in: The original observation value at time t; Data after exponential smoothing; : Smoothing coefficient (0 < α < 1); : The smoothed value at the previous time step.

[0032] After completing the preprocessing of multi-source data, for each user's first... The unified feature vector representation of the second-charge / battery swapping behavior is as follows:

[0033] Each feature corresponds to a multi-dimensional factor, including vehicles, batteries, facilities, traffic, power grids, and weather. Organizing all user behavior samples according to this feature structure yields a complete sample set. This set serves as the core input data foundation for subsequent system dynamics model construction and charging / swapping behavior mechanism identification.

[0034] Preferably, to achieve a systematic analysis of the user's charging and swapping behavior mechanism, this invention constructs an overall model structure based on system dynamics. This model consists of a "charging / swapping time selection submodule" and a "charging / swapping location selection submodule," corresponding to the two core decision dimensions of user behavior, respectively. By characterizing the causal chains, feedback paths, and interaction effects among multiple factors, and further analyzing the decision coupling relationship of "time selection → location selection," this model can fully reproduce the user's behavior generation mechanism, providing theoretical support for the optimal allocation of energy replenishment resources and the design of behavior guidance strategies.

[0035] Preferably, in the charging / swapping time selection submodule, this invention focuses on analyzing the inherent logic of users selecting specific charging / swapping time periods. This submodule integrates multi-source data such as vehicle operating status, charging / swapping facility conditions, travel characteristics, grid prices and capacity constraints, and meteorological environment. By constructing a multi-loop, strongly coupled causal feedback system, it reveals the influence path and dynamic response law of each factor on users' time selection behavior.

[0036] Preferably, in the charging / swapping time selection submodule, the charging / swapping demand and adjustable power model consists of three core elements: user charging / swapping demand, adjustable power, and time period flexibility. By quantifying these three types of indicators, the rigid demand and schedulable space of users in time period selection can be systematically characterized, and the calculation formula is shown below.

[0037] (1) User charging and battery swapping demand: User charging and battery swapping demand refers to the amount of electrical energy that an electric vehicle must replenish at a specific time or during a specific travel cycle to meet range requirements, battery safety thresholds, and travel plans. This demand depends on factors such as the vehicle's current remaining SOC, expected travel mileage, vehicle energy consumption level, and user charging preferences. It is the most core and rigid indicator of user energy replenishment behavior.

[0038]

[0039] In the formula: Correction factor, taking into account charging and discharging efficiency loss; Battery capacity; Target SOC; Current SOC.

[0040] (2) Adjustable power (reflecting pricing guidance space):

[0041] In the formula: Price elasticity; : Correction factor for the ratio of intraday price to benchmark price; The ratio of the intraday price to the benchmark price; Excitation intensity correction coefficient; Incentive intensity.

[0042] (3) Definition of Time-of-Use Flexibility: The time-of-use flexibility of charging and swapping is determined by both the user's energy demand and the adjustable capacity. The demand for charging and swapping exerts a negative constraint on time-of-use flexibility, reflecting the rigidity of the user's energy consumption; the adjustable capacity provides a positive impetus, reflecting the adjustment space brought about by price and incentives. The two are combined in a weighted manner to form the indicator of "time-of-use flexibility," which is used to quantify the user's potential response capability in load guidance and charging timing adjustment. The specific formula is as follows:

[0043] In the formula: The flexibility coefficient of charging facilities during this period; Adjustable power coefficient; Adjustable power consumption; User charging and battery swapping demand coefficient; User demand for charging and battery swapping.

[0044] Preferably, in the charging / swapping time selection submodule, the model mainly consists of three types of core feedback loops: the first is the "grid price - adjustable capacity - charging / swapping demand" loop, used to describe the dynamic impact of the price mechanism on the user's dispatchable space and energy replenishment demand; the second is the "vehicle battery status + weather conditions - charging preference - charging / swapping demand" loop, used to characterize the impact of the vehicle's remaining battery capacity, health status, and environmental factors on the user's willingness to replenish energy; and the third is the "travel characteristics + weather conditions - travel demand - charging / swapping demand" loop, used to reflect the driving role of user travel patterns and weather environment on the evolution of energy replenishment demand. The mathematical expressions of the mechanism and key parameters of the above three feedback loops are shown below.

[0045] Preferably, the grid (price) - adjustable capacity - charging / swapping demand loop: This loop characterizes the dynamic regulation effect of the electricity pricing mechanism on users' energy replenishment demand. Grid-side factors such as charging / swapping service fees, time-of-use pricing, and incentive pricing, through price elasticity coefficients, comprehensively affect users' adjustable capacity, creating a suppressive effect on the original demand. A negative feedback relationship exists between adjustable capacity and charging / swapping demand, forming a chain of action: "price → adjustable capacity → demand (negative feedback)," thereby achieving flexible guidance of user behavior. Based on this, users' actual energy replenishment demand can be expressed as:

[0046] In the formula: :time The original charging and swapping demand, that is, the theoretical energy replenishment demand when it is not affected by price. Adjustable electricity consumption calculated based on price factors represents the amount of electricity a user can adjust at any given time due to changes in electricity prices. The negative feedback adjustment coefficient (price elasticity adjustment strength) is used to reflect the degree of impact of price changes on adjustable electricity. : The actual energy replenishment demand of users after the price regulation effect.

[0047] Preferably, the feedback loop of vehicle (battery state) + weather → charging preference → charging / swapping demand reveals the intrinsic mechanism by which vehicle battery state and weather conditions influence users' charging preferences, thereby shaping charging / swapping demand. Battery capacity and state of health (SOH) reduce users' demand for higher initial SOC, while range anxiety prompts users to recharge earlier, thus increasing their preference for initial SOC. Simultaneously, weather factors (such as ambient temperature) indirectly increase users' recharge demand by affecting vehicle energy consumption levels. Under these combined effects, this loop forms a composite feedback path of "battery state + weather → charging preference / remaining capacity → charging / swapping demand": initial SOC preference positively drives demand, while the current remaining SOC negatively inhibits it.

[0048] SOC Preference Model:

[0049] In the formula: : The user's expected initial SOC level (psychological preference value); : Basic preference level (usually corresponding to the typical initial SOC formed by daily vehicle travel habits); : Normalized value of battery health status (the lower the SOH, the more likely the user is to charge the battery earlier, so it has a negative impact). : The influence coefficient on preference (negative coefficient); : Range Anxiety Index (the higher the value, the more inclined one is to maintain a higher SOC). : Mileage anxiety impact coefficient (positive coefficient); Energy consumption per unit after meteorological influence (the greater the temperature difference, the higher the energy consumption, thus increasing SOC preference); : The influence coefficient of energy consumption level on preferences (positive coefficient).

[0050] Model of the impact of temperature on energy consumption:

[0051] In the formula: : Normalization factor for unit energy consumption after taking temperature into account. : Current ambient temperature. Optimal energy consumption temperature (typically around 20–25°C for electric vehicles). : Absolute value of temperature difference; the greater the temperature difference, the higher the energy consumption. : Sensitivity coefficient of temperature to energy consumption.

[0052] Revised charging / swapping demand formula:

[0053] In the formula: The final adjusted charging and swapping requirements take into account battery status, energy consumption, and preferences. Demand correction factor, used to adjust the scale and sensitivity of the calculation results. : Nominal battery capacity. : The user's target initial SOC preference value. : The vehicle's current remaining SOC value.

[0054] Preferably, the loop is structured as follows: Road (travel characteristics) + Weather - Travel Demand - Charging / Swapping Demand. This loop reveals the synergistic constraint mechanism of travel behavior, weather disturbances, and grid capacity on charging / swapping demand. Specifically, it includes two dominant paths: First, traffic flow and parking duration inhibit travel, while travel mileage positively drives travel demand; increased travel demand reduces users' time-based guidance and ultimately affects charging / swapping demand. Second, grid capacity, as an external supply constraint, directly compresses charging / swapping power demand by limiting the actual available charging power. Simultaneously, weather factors (such as temperature and rainfall), as external disturbances, indirectly affect the system by altering travel demand. These paths collectively constitute a composite feedback loop of "travel behavior + weather → travel demand → guidance / grid capacity → charging / swapping demand," reflecting the joint external constraints imposed on user charging / swapping behavior by the road, weather, and grid sides.

[0055] Travel demand model:

[0056] In the formula: The intensity of travel demand indicates the user's motivation for traveling on that day. : Basic travel demand constant term. Traffic flow (the higher the value, the more congested the traffic). Traffic flow's inhibitory effect on travel demand. Parking duration reflects the probability that a user will not travel. The inhibition coefficient of parking duration on travel demand. : Planned or actual travel distance. : The positive impact coefficient of travel mileage on travel demand. Meteorological influencing factors (obtained by normalization of temperature, rainfall, etc.). : The impact coefficient of weather on travel demand.

[0057] Behavioral guidance model:

[0058] In the formula: User time-related guidance (the lower the value, the more difficult it is to control); The coefficient by which travel demand diminishes the effectiveness of guidance. Intensity of travel demand.

[0059] Grid capacity constraints:

[0060] In the formula: Time t represents the actual charging power that can be provided to the vehicle. : Rated maximum charging power of the charging pile. Total available capacity of the distribution area (grid power supply capacity). The number of vehicles charging in parallel during the same period.

[0061] Preferably, the charging / swapping location selection submodule consists of a loop of charging pile (facility type) – accessibility – location flexibility, a loop of network (price) – location flexibility, and a loop of weather – station attractiveness – facility accessibility. This submodule aims to reveal the decision-making mechanism of users in choosing charging / swapping locations in the spatial dimension. Its core lies in quantifying the comprehensive impact of facility supply characteristics and spatial accessibility on user choice behavior. The model comprehensively considers key variables such as facility type, private charging pile availability, charging / swapping coverage density, electricity price level, and weather conditions, and on this basis, forms the "charging / swapping location selection flexibility" index. This index is jointly determined by facility accessibility and location flexibility, and can comprehensively reflect the user's ability to adjust charging / swapping locations in different spatial scenarios, providing a quantitative basis for optimizing facility layout and formulating differentiated guidance strategies.

[0062] Preferably, a loop of "pile (facility type) – accessibility – location flexibility" is used to characterize the driving mechanism of the charging and swapping facility supply structure on the user's location selection flexibility. Specifically, facility types are divided into private piles and public piles. The higher the prevalence of private piles, the lower the user's dependence on public facilities, thus weakening the overall network's public accessibility; while the higher the coverage density of public facilities, the more significantly it can enhance the convenience of charging and swapping within the area, improving the user's accessibility level. In addition, the queuing rate of public piles will reduce the actual availability of facilities, creating a negative impact. The facility accessibility obtained by combining the above factors will positively affect "location selection flexibility," forming a logical chain of "facility type → supply structure → accessibility → location flexibility," revealing the fundamental driving role of infrastructure layout on users' spatial selection ability. The facility accessibility model is expressed as follows:

[0063] In the formula: :Location The facility accessibility metric is used to quantify how easily users within a region can reach available facilities. : Model constant term, used to reflect the basic accessibility level. :Location Public stake coverage density (number of public stakes / area). : Influence coefficient of public pile coverage density (positive influence). Private parking lot penetration rate (the percentage of users with private parking lots). : The impact coefficient of the private pile popularization rate (negative impact). Queuing rate or busyness of public parking spaces (indicating congestion). Queuing rate impact coefficient (negative impact).

[0064] Network (Price) – Location Flexibility is used to characterize the driving mechanism of how infrastructure supply structure affects the flexibility of user location choices. First, infrastructure types are divided into private charging stations and public facilities. The prevalence of private charging stations directly reduces users' reliance on accessibility to public facilities, exhibiting a negative impact; while the coverage density of public facilities directly improves the overall network accessibility, exhibiting a positive impact. Ultimately, infrastructure accessibility positively determines the location flexibility of charging and swapping stations, forming a transmission path of "infrastructure type → (private / public charging station) supply → accessibility → location flexibility," revealing the fundamental constraints and drivers of infrastructure layout on the flexibility of user spatial choices.

[0065] Building upon this, price utility is introduced as a significant exogenous factor influencing location selection, reflecting the impact of service fee differences between different stations on user preferences. By integrating facility accessibility, price utility, and user dependence on public utility points, a quantitative model of location selection flexibility can be constructed. The price utility model is defined as follows:

[0066] In the formula: :Location Price utility is used to measure the impact of service prices on user location preferences. Price sensitivity coefficient: This indicates the degree to which users respond to differences in charging and battery swapping service fees. The regional average service fee serves as a benchmark price level. :Location The actual service fee level.

[0067] The formula for calculating location selection flexibility is as follows:

[0068] In the formula: :Location The flexibility of charging and swapping location selection is used to measure the user's ability to adjust charging and swapping locations in a spatial dimension. : Facility accessibility weight, representing the degree to which accessibility contributes to location flexibility. :Location Accessibility indicators for facilities. Price utility weight represents the degree to which price factors affect spatial choice flexibility. : The weighting coefficient of the degree of dependence on common piles, representing the degree to which dependence constrains flexibility. The degree of user dependence on public stubs (the higher the dependence, the lower the flexibility).

[0069] The meteorological-site attractiveness-facility accessibility loop reveals how meteorological conditions indirectly affect the accessibility of charging and swapping facilities and ultimately influence location selection flexibility by influencing users' protection needs and site attributes. Specifically, temperatures deviating from the comfort range and increased rainfall raise users' needs for sun and rain protection; these needs, along with the actual protection conditions of the site, jointly determine the site's attractiveness. When protection needs increase but site protection conditions are insufficient, its attractiveness decreases significantly.

[0070] Subsequently, the attractiveness of charging stations negatively impacts perceived accessibility by influencing users' subjective willingness to choose: the lower the attractiveness, the less willing users are to choose that location, leading to a decrease in their perceived accessibility. This ultimately forms a holistic causal chain of "weather → protection needs → charging station attractiveness → accessibility → location flexibility," reflecting the indirect constraint mechanism of external weather disturbances on charging and swapping location selection behavior.

[0071] User protection requirements:

[0072] In the formula: The level of protection required by the user. Sensitivity coefficient of protective requirements to temperature deviations (too high or too low). The deviation of the temperature from the comfortable temperature zone (e.g.) ). Sensitivity coefficient of rainfall to protection requirements. Rainfall amount index.

[0073] Station attractiveness:

[0074] In the formula: :Location The attractiveness of the station (the larger the station, the more attractive it is to users). : The positive contribution coefficient of site protection conditions to attractiveness. :Location Actual protective conditions (such as sunshades, awnings, etc.). : The negative impact coefficient of user protection needs on attractiveness. User protection requirements.

[0075] Attraction modifies accessibility:

[0076] In the formula: :Location Actual (perceived) facility accessibility. :Location Basic accessibility without considering the attractiveness of the station. : Attraction sensitivity coefficient, used to characterize the strength of the decrease in attraction affecting accessibility. Station attractiveness. : Reflects the degree of lack of attractiveness (the lower the attractiveness → the larger the value).

[0077] Preferably, in the model structure of this invention, the user's time period selection and location selection have a cross-module coupling relationship: the flexibility of charging and swapping time periods negatively affects the flexibility of location. When users lack adjustment space in the time dimension (e.g., they must charge at a specific time), the range of stations they can choose will also shrink, leading to a decrease in location selection flexibility. This coupling relationship reflects the behavioral linkage of users in both time and space dimensions, embodying the system association and mutual constraint mechanism between the two sub-modules.

[0078]

[0079] In the formula: The flexibility of charging and swapping locations after time-period coupling correction; Original value of positional flexibility without considering coupling effects; : Flexibility of charging and swapping time periods (0-1); : The rigidity of the time period (the higher the rigidity, the less room there is for time selection); : Spatiotemporal coupling sensitivity coefficient, used to quantify the constraint strength of time period rigidity on location selection.

[0080] Preferably, this invention is based on a causal loop model to compare and analyze the driving mechanism and regulation effect of user charging and swapping time period selection behavior in "no guidance strategy scenario" and "with guidance strategy scenario", and to clarify the role path of the guidance mechanism in time period flexibility.

[0081] Preferably, the scenario without guidance strategy is characterized by systems with strong rigidity and weak flexibility.

[0082] (1) Rigid demand driven by the vehicle's spontaneous state. Under unguided conditions, the vehicle's battery capacity, SOH and other physical parameters spontaneously determine the user's initial SOC for charging. The user's high sensitivity to battery degradation further strengthens the preference for charging in the "high SOC safe range", resulting in a significant rigidity in demand. At the same time, extreme temperatures increase vehicle energy consumption, causing the user's target SOC to shift upward, resulting in early charging and overcharging, and demand will concentrate during convenient times such as evening. In addition, range anxiety continues to dominate user behavior in the absence of technical and cognitive intervention, making them tend to charge earlier in the day, thus lacking the ability to flexibly adjust according to the time of day.

[0083] (2) Travel patterns lead to time rigidity and peak overlap. Users' travel behavior forms a stable commuting rhythm under the influence of factors such as traffic flow, mileage, and parking time, making morning and evening rush hours unavoidable periods of load overlap. In this context, energy replenishment behavior is passively concentrated in specific peak time windows, forming a typical "behavioral solidification" pattern, making it difficult for users to break away from fixed energy replenishment paths during fixed time periods.

[0084] (3) Lack of market mechanism and rigid grid capacity strengthen behavioral entrenching. Due to the lack of effective price regulation, the negative feedback link between price and demand is weak, and users are not sensitive to changes in electricity prices, remaining in their habitual charging periods. At the same time, the physical constraints of transformer capacity lead to passive voltage drop during peak hours. Although this reduces the user experience, the lack of clear alternative incentives prevents users from voluntarily shifting to off-peak hours, thus further locking in their behavioral structure.

[0085] (4) Meteorological factors cause nonlinear fluctuations in demand over time. High and low temperatures significantly increase power consumption, leading to a concentrated increase in users' energy replenishment demand, while adverse weather conditions such as rain and strong winds suppress travel activities, causing a sudden drop in demand. These meteorological disturbances cause charging and swapping behavior to exhibit a nonlinear fluctuation characteristic of "normal concentration - extreme drop", further weakening the stability of the system in the time dimension.

[0086] Preferably, there are guidance strategy scenarios: negative feedback reinforcement and behavioral flexibility enhancement.

[0087] (1) Technological and cognitive interventions enhance time-of-use flexibility. Through technologies such as battery health transparency, charge and discharge loss prediction, and high-precision range estimation, users can more accurately judge their energy replenishment needs and reduce their dependence on high SOC. Intelligent temperature control technology smooths out energy consumption fluctuations caused by extreme temperatures, and weather warnings combined with energy-saving driving and early energy replenishment reminders reduce concentrated charging. At the same time, intelligent navigation based on the road condition-range fusion model provides dynamic energy replenishment suggestions and reservation services, improving users' confidence in traveling under low SOC conditions and fundamentally improving the ability to adjust time of day.

[0088] (2) Travel intervention promotes demand dispersion. By means of off-peak energy replenishment incentives, traffic signal optimization and commuting organization adjustment, the amount of travel during peak hours is effectively reduced, giving users more opportunities to charge and swap batteries during off-peak hours, and prompting energy replenishment demand to be distributed more evenly over time.

[0089] (3) Price adjustment and capacity flexibility work together to build a negative feedback loop in the system. By implementing differentiated time-of-use pricing, peak electricity prices are increased by 30%–50%, and off-peak electricity prices are decreased by 20%–40%, strengthening the negative feedback loop of "price change-demand response" and guiding users to move their adjustable electricity to off-peak periods. At the same time, combined with dynamic capacity expansion, orderly charging, and flexible load control strategies, peak power is controlled and off-peak power resources are more fully available, thereby significantly improving users' willingness to migrate between time periods.

[0090] (4) Meteorological early warning linkage mechanism to smooth demand. By providing advance energy replenishment suggestions based on meteorological forecasts, users are guided to complete the necessary energy replenishment before extreme weather arrives and reduce unnecessary high energy consumption during severe weather, so that the overall energy replenishment demand is smoother in time series and the stability and controllability of the system are improved.

[0091] Preferably, the presence or absence of a guidance mechanism leads to different patterns in how users choose charging and swapping locations. In scenarios without guidance, charging and swapping facilities primarily rely on spontaneous market supply to form their layout, and users' location choices are limited by the conditions of private charging pile construction and the original coverage level of public facilities, exhibiting the following characteristics.

[0092] (1) Limited choices due to the spontaneous deployment of facilities. Since the coverage density of public charging and swapping facilities is spontaneously regulated by the market, there is generally insufficient supply in high-frequency travel areas (such as business districts and main commuting routes), while there is facility redundancy in low-frequency areas. This results in low overall accessibility, which significantly restricts users' location choices and makes it difficult to effectively improve location flexibility.

[0093] (2) Path dependence restricts location. Users have stable path preferences in their daily travel, and their charging and swapping decisions are often limited to the area around their fixed commute or regular routes. Path dependence strengthens travel inertia, making it difficult for users to expand to non-habitual areas, thereby reducing the spatiotemporal elasticity of location selection.

[0094] (3) Price restrictions lead to behavioral solidification. In the absence of service fee regulation, price competition among facilities tends to become disorderly, with low-priced facilities prone to user concentration and high-priced facilities experiencing underutilization. Price becomes the sole factor influencing user choices, leading to uneven facility utilization and further solidifying location selection behavior.

[0095] (4) Behavioral fluctuations caused by meteorological factors. Some open-air stations lack necessary protective structures (such as sunshade and rain protection facilities), and their availability decreases under extreme weather conditions such as high temperature and heavy rain. This causes users to reduce their use under adverse weather conditions, which has an unstable impact on the flexibility of location selection.

[0096] Preferably, in scenarios with guidance strategies, the establishment of facility supply regulation and price guidance mechanisms can create positive incentive feedback, thereby effectively improving the flexibility of users' charging and swapping location selection. Specifically, this includes: (1) Optimization of choices under planning intervention. By regulating the type and quantity of facilities through policy measures, such as restricting the approval of new facilities in areas with saturated public facilities and providing construction subsidies in areas with shortages, the limitation of the original coverage of public facilities on accessibility can be weakened, so as to form a more balanced spatial distribution of facilities. Furthermore, by implementing targeted coverage densification based on travel big data and charging and swapping demand heat maps, the facility density in high-frequency areas can be optimized, thereby enhancing users' flexibility in spatial choices.

[0097] (2) Route guidance under travel intervention. By integrating charging and swapping information into the navigation system, pushing the "charging and swapping + travel" joint optimized route, and recommending alternative stations with detour costs lower than queuing costs, users' fixed path dependence can be reduced, and some demand can be diverted to idle facility areas to improve the overall controllability of location selection.

[0098] (3) Price-guided behavior optimization. By setting a reasonable range for station service fees (such as a cost markup of 10%–20%), extreme low-price competition can be suppressed, making different facilities more attractive in terms of price. This can effectively balance the utilization level of stations and encourage users to choose locations more evenly.

[0099] (4) Stable behavior under facility upgrades. By adding adaptive structures such as sunshades and rain shelters to open-air stations, the availability of facilities under extreme weather conditions can be improved, reducing the impact of weather disturbances on location selection and keeping user behavior relatively stable in time and space.

[0100] In summary, the user charging and swapping behavior mechanism identification method based on multi-source data fusion of the present invention can have the following beneficial effects: This invention constructs a fusion system for multi-dimensional heterogeneous data, including vehicle, charging pile, road, network, and meteorological data. Through unified feature vector construction and data cleaning, alignment, and deep fusion technologies, it achieves a comprehensive and fine-grained characterization of user charging and swapping behavior. Compared to existing solutions that rely solely on a single data source such as a vehicle or charging pile, this invention overcomes the limitations of data silos, significantly improving the accuracy of behavioral feature recognition and providing a high-quality data foundation for dynamic behavior modeling and subsequent intervention.

[0101] This invention proposes to decompose user charging and swapping behavior into two core decision dimensions: time period selection and location selection. It also constructs a system dynamics model incorporating multiple factors such as battery status, travel patterns, price signals, and weather conditions, forming multiple positive / negative feedback loops to quantify the dynamic evolution and mutual reinforcement mechanisms of behavior. This modeling method can capture the cumulative characteristics and coupling effects of behavior over time, representing a key innovation that distinguishes it from existing static statistical models.

[0102] This invention reveals for the first time the objective law that "time-of-use flexibility and location-of-use flexibility are negatively correlated" in user charging and swapping. By using cross-module coupling functions and a linked feedback structure, it integrates the two decision dimensions into the same dynamic system for unified solution, achieving a quantifiable description of the mutual constraints and mutual driving relationships between different dimensional behaviors. This mechanism breaks through the limitations of traditional research that separates the two types of behaviors, providing a more realistic and systematic model foundation for accurately predicting user behavior paths.

[0103] This invention uses methods such as sensitivity analysis and feedback link strength calculation to systematically identify "guideable variables" that can be explicitly controlled by factors such as policy, technology, or price, and constructs a strategy generation chain from variable intervention → behavioral change → system performance improvement. Based on this method, differentiated intervention measures can be formulated for different behavioral drivers, achieving precise guidance on users' time and location choices, improving the utilization efficiency and balance of charging and swapping facilities, and forming a decision-making tool that can be directly applied to actual operation and policy formulation.

[0104] To implement the methods of the above embodiments, the present invention also provides a user charging and swapping behavior mechanism identification system 10 based on multi-source data fusion, such as... Figure 5 As shown, it includes: The feature vector construction module 100 is used to construct a set of multi-dimensional feature vectors representing user charging and swapping behavior after preprocessing multi-source heterogeneous data. The causal loop model construction module 200 is used to input the multidimensional feature vector set into the system dynamics modeling framework to construct a causal loop model that includes a charging / swapping time selection submodule and a charging / swapping location selection submodule. The coupling mechanism construction module 300 is used to establish a cross-module coupling mechanism between the charging and swapping time selection submodule and the charging and swapping location selection submodule, and to correct the location flexibility through a negative constraint function to reflect the limitation of time period rigidity on the spatial selection range. The sensitivity analysis module 400 is used to perform parameter sensitivity analysis based on the causal loop model, identify key guideable variables such as electricity price structure, facility distribution and meteorological response, and generate differentiated user behavior guidance strategies.

[0105] The user charging and swapping behavior mechanism identification system based on multi-source data fusion in this invention can accurately identify key guideable variables that affect user behavior, providing a scientific basis for formulating refined and flexible behavior guidance strategies. This enhances the spatiotemporal adjustability of user behavior, alleviates grid pressure caused by disorderly charging, promotes the optimal allocation of energy replenishment resources and the rational layout of the network, and provides technical support for building a safe, economical and efficient electric vehicle charging mode.

[0106] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0107] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0108] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A method for identifying user charging and swapping behavior mechanisms based on multi-source data fusion, characterized in that, include: After preprocessing the multi-source heterogeneous data, a set of multi-dimensional feature vectors representing users' charging and swapping behavior is constructed. The multidimensional feature vector set is input into the system dynamics modeling framework to construct a causal loop model that includes a charging / swapping time selection submodule and a charging / swapping location selection submodule. A cross-module coupling mechanism is established between the charging / swapping time selection submodule and the charging / swapping location selection submodule, and the location flexibility is corrected by a negative constraint function to reflect the limitation of time period rigidity on the spatial selection range. Based on the causal loop model, parameter sensitivity analysis is performed to identify key guideable variables such as electricity price structure, facility distribution, and meteorological response, and to generate differentiated user behavior guidance strategies.

2. The method as described in claim 1, characterized in that, After preprocessing the multi-source heterogeneous data, a multi-dimensional feature vector set representing user charging and swapping behavior is constructed, including: Raw time-series data were collected from vehicles, charging and swapping facilities, transportation, power grids, and meteorological systems. The original time-series data is uniformly aligned on time scale, missing values ​​are imputed, and noise filtering is performed using exponential smoothing. Based on the preprocessed data, a multi-dimensional unified feature vector containing battery capacity, state of charge, traffic flow, time-of-use electricity price, and temperature is constructed for each user's charging and swapping behavior to form a multi-dimensional feature vector set.

3. The method as described in claim 1, characterized in that, The multidimensional feature vector set is input into the system dynamics modeling framework to construct a causal loop model including a charging / swapping time selection submodule and a charging / swapping location selection submodule, including: Based on the multidimensional feature vector set, a charging and swapping time selection submodule is constructed. By defining charging and swapping demand, adjustable power capacity and time period flexibility state variables, three types of feedback loop mechanisms are quantified: grid price-demand suppression, battery status + weather-preference driving, and travel characteristics + weather-guided constraint. A charging and swapping location selection submodule is constructed. By defining state variables of facility accessibility, price utility, and location flexibility, three types of feedback loop mechanisms are quantified: pile-accessibility-location flexibility, network-price-location selection, and weather-site attractiveness-accessibility correction. The charging / swapping time selection submodule and the charging / swapping location selection submodule are integrated with system dynamics to form an overall causal loop model structure that includes multi-loop causal chains and the interaction relationship of state variables.

4. The method as described in claim 1, characterized in that, A cross-module coupling mechanism is established between the charging / swapping time selection submodule and the charging / swapping location selection submodule. A negative constraint function is used to correct the location flexibility, reflecting the time-period rigidity's limitation on the spatial selection range, including: Based on the time period flexibility variable output by the charging and swapping time selection submodule and the location flexibility variable output by the charging and swapping location selection submodule, the negative correlation coupling relationship between the two is determined to characterize the constraint effect of time period rigidity on the spatial selection range. Construct a spatiotemporal coupling negative constraint function, substitute the time period flexibility variable as the independent variable, and calculate the correction coefficient for location flexibility; The original position flexibility variables are weighted and corrected using the correction coefficients, and the coupled and corrected position flexibility variables are output to complete cross-module linkage modeling.

5. The method as described in claim 1, characterized in that, Based on the aforementioned causal loop model, parameter sensitivity analysis is performed to identify key guideable variables such as electricity price structure, facility distribution, and meteorological response, generating differentiated user behavior guidance strategies, including: Based on the complete set of parameters of the causal loop model, parameter perturbation schemes and sensitivity evaluation indicators are designed to quantify the degree of influence of each parameter on behavioral flexibility. The parameter sensitivity analysis calculation was performed. By systematically adjusting the electricity price elasticity coefficient, facility density weight, and meteorological sensitivity coefficient, the response changes of flexibility and location flexibility in different time periods were evaluated, and the sensitivity ranking results were obtained by ranking the degree of influence of the parameters. Based on the sensitivity ranking results, the key inductive variables are identified as electricity price structure, public pile coverage density, and meteorological response factor. Differentiated user behavior guidance strategies are generated, including the adjustment range of time-of-use electricity prices, the direction of facility spatial layout optimization, and the meteorological early warning linkage mechanism.

6. The method as described in claim 2, characterized in that, The multi-source heterogeneous data includes at least: Vehicle data: Battery capacity Battery health status Current state of charge Unit energy consumption ; Charging and swapping power data Facility type ; Road data: Traffic flow Parking duration Travel mileage ; Network data: Time-of-use electricity pricing Incentive prices Service fee Area capacity ; Meteorological data: temperature Rainfall .

7. The method as described in claim 6, characterized in that, The original time series is smoothed using the exponential smoothing method, as shown in the following formula: in: The original observation value at time t; Data after exponential smoothing; Smoothing coefficient (0 < α < 1); : The smoothed value at the previous time step; After completing the preprocessing of multi-source data, for each user's first... The unified feature vector representation of the second-charge / battery swapping behavior is as follows: The features correspond to various factors including vehicles, batteries, facilities, traffic, power grids, and weather. All user behavior samples are organized according to this feature structure to obtain a complete sample set. .

8. The method as described in claim 7, characterized in that, In the charging / swapping time selection submodule, the charging / swapping demand and adjustable capacity model consists of three core elements: user charging / swapping demand, adjustable capacity, and time slot flexibility. These three indicators quantify the user's rigid demand and schedulable space in time slot selection, and the calculation formulas are shown below: User charging and battery swapping demand: This demand depends on factors such as the vehicle's current remaining SOC, expected travel range, vehicle energy consumption level, and user charging preferences. In the formula: Correction factor, taking into account charging and discharging efficiency loss; Battery capacity; Target SOC; Current SOC; Adjustable power: In the formula: Price elasticity; : Correction factor for the ratio of intraday price to benchmark price; The ratio of the intraday price to the benchmark price; Excitation intensity correction coefficient; Incentive intensity; Time-of-use flexibility definition: The flexibility of charging and swapping time periods is determined by both the user's power demand and the adjustable power capacity; these two factors are combined in a weighted manner to quantify the user's potential responsiveness in load guidance and charging timing adjustments, as shown in the following formula: In the formula: The flexibility coefficient of charging facilities during this period; Adjustable power coefficient; Adjustable power consumption; User charging and battery swapping demand coefficient; User demand for charging and battery swapping.

9. The method as described in claim 8, characterized in that, In the charging / swapping time selection submodule, the model consists of three core feedback loops: grid price—adjustable capacity—charging / swapping demand loop, vehicle battery status + weather conditions—charging preference—charging / swapping demand loop, and travel characteristics + weather conditions—travel demand—charging / swapping demand loop; the mathematical expression is shown below: Grid price - adjustable capacity - charging / swapping demand loop: The user's actual energy replenishment demand is expressed as: In the formula: :time The original charging and swapping demand, that is, the theoretical energy replenishment demand when it is not affected by price. Adjustable electricity consumption calculated based on price factors represents the amount of electricity a user can adjust at any given time due to changes in electricity prices. : Negative feedback adjustment coefficient, used to reflect the degree of impact of price changes on adjustable electricity; : The actual energy replenishment needs of users after price regulation; Vehicle battery status + weather conditions — charging preference — charging / swapping demand loop: SOC Preference Model: In the formula: User's expected initial SOC level for charging; Basic preference level; : Normalized value of battery health status; : The influence coefficient on preferences; Mileage anxiety index; : Range anxiety impact factor; Energy consumption per unit after weather influence; The coefficient of influence of energy consumption level on preferences; Model of the impact of temperature on energy consumption: In the formula: : Normalization factor for unit energy consumption after taking temperature into account; Current ambient temperature; Optimal energy consumption temperature; : Absolute value of temperature difference; the larger the temperature difference, the higher the energy consumption. Temperature sensitivity coefficient to energy consumption; Revised charging / swapping demand formula: In the formula: The final adjusted charging and swapping requirements take into account battery status, energy consumption, and preferences. Demand correction factor, used to adjust the scale and sensitivity of calculation results; Nominal battery capacity; : The user's initial target SOC preference value; : The vehicle's current remaining SOC value.

10. A user charging and swapping behavior mechanism identification system based on multi-source data fusion, characterized in that, include: The feature vector construction module is used to construct a set of multi-dimensional feature vectors that characterize user charging and swapping behavior after preprocessing multi-source heterogeneous data. The causal loop model construction module is used to input the multidimensional feature vector set into the system dynamics modeling framework to construct a causal loop model that includes a charging / swapping time selection submodule and a charging / swapping location selection submodule. The coupling mechanism construction module is used to establish a cross-module coupling mechanism between the charging and swapping time selection submodule and the charging and swapping location selection submodule, and to correct the location flexibility through a negative constraint function to reflect the limitation of time period rigidity on the spatial selection range. The sensitivity analysis module is used to perform parameter sensitivity analysis based on the causal loop model, identify key guideable variables such as electricity price structure, facility distribution and meteorological response, and generate differentiated user behavior guidance strategies.

Images