Electric vehicle charging behavior statistical method based on time period analysis

Abstract

The invention relates to the technical field of electric vehicle charging, and discloses an electric vehicle charging behavior statistical method based on time period analysis, which comprises the following steps: collecting charging behavior data in different time periods through real-time feedback of a vehicle-mounted sensor, a charging pile information system and a user intelligent terminal; establishing a time-phased data set of the charging demand; according to charging demand change trends in different time periods, performing time period division on the charging demand data; calculating optimal charging coefficients of each charging pile in different time periods on the basis of the time period analysis result in combination with historical occupation information of each charging pile; the distribution scheme of the charging piles is optimized according to the optimal charging coefficients of the charging piles and the charging demand prediction result; after charging is started, whether the charging state of the electric vehicle is normal or not is judged by monitoring the battery state and the load power of the electric vehicle in real time and combining historical data. The method has the advantages of improving the prediction precision of the charging demand and the capturing capability of the time period change.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle charging, and specifically to a statistical method for electric vehicle charging behavior based on time period analysis. Background Art

[0002] With the continuous increase in the popularity of electric vehicles, the research on electric vehicle charging behavior has become particularly important. The statistical analysis of charging behavior not only provides important decision-making support for aspects such as the operation and management of electric vehicles, the layout optimization of charging infrastructure, and the load scheduling of power systems, but also provides key data references for urban traffic management, the development of the electric vehicle industry, and the promotion of green energy. However, the existing methods for analyzing electric vehicle charging behavior mainly focus on static and macroscopic statistics, lacking in-depth exploration of the time-periodic change characteristics of charging behavior. Most methods only rely on long-term average data, ignoring the volatility and sudden changes in charging demand and behavior at different time periods. In fact, there are significant differences in electric vehicle charging behavior between different time periods, and these differences are affected by the intertwined influence of various factors, such as users' travel habits, weather changes, traffic congestion, the geographical distribution of charging stations, electricity price fluctuations, and policy adjustments. Especially with the increase in social diversification and personalized needs, charging behavior is not limited to home charging, but also includes various scenarios such as public charging piles and fast charging stations. During daily peak periods, users often choose to charge during commuting hours, while during off-peak periods, they may conduct centralized charging. In addition, electricity price fluctuations also significantly affect users' charging decisions. Especially during periods of high electricity prices, many users may choose to delay charging. Moreover, the charging behavior patterns vary greatly in different regions, and there are obvious differences in charging demand, frequency, charging duration, etc. between the urban central area and the suburbs. Traditional static analysis methods are difficult to capture such spatio-temporal changes.

[0003] Existing methods have certain deficiencies in dynamic charging behavior modeling and capturing time-periodic changes, and cannot fully reflect the complexity and diversity of charging demand, especially unable to accurately predict the charging load characteristics under different time periods and different environmental conditions. Traditional analysis methods have not effectively considered the volatility, seasonality of charging demand over time, and the personalized association with user behavior, resulting in large errors in practical applications. For example, in power grid load scheduling, existing prediction models cannot provide accurate charging load predictions based on specific time-period characteristics, thus affecting the stability and efficiency of the power system. Therefore, it is necessary to design a statistical method for electric vehicle charging behavior based on time period analysis that can improve the prediction accuracy of charging demand and the ability to capture time-periodic changes. Summary of the Invention

[0004] (1) Technical Problems to be Solved

[0005] In view of the deficiencies of the prior art, the present invention provides a statistical method for electric vehicle charging behavior based on time period analysis, which has the advantages of improving the prediction accuracy of charging demand and the ability to capture time period changes, and solves the problems in the above-mentioned background art.

[0006] (II) Technical solution

[0007] To achieve the above-mentioned objectives of improving the prediction accuracy of charging demand and the ability to capture time period changes, the present invention provides the following technical solution: A statistical method for electric vehicle charging behavior based on time period analysis, including the following steps:

[0008] S1: Collect charging behavior data within different time periods. The data collection is carried out through the real-time feedback of in-vehicle sensors, charging pile information systems, and user intelligent terminals, and a time-period dataset of charging demand is established;

[0009] S2: According to the charging demand change trend within different time periods, divide the charging demand data into time periods, and use a time period analysis model to model the charging demand within each time period;

[0010] S3: Based on the time period analysis results, combined with the historical occupancy information of each charging pile, calculate the optimal charging coefficient of each charging pile in different time periods, and construct a charging demand prediction model;

[0011] S4: According to the optimal charging coefficient of the charging pile and the charging demand prediction result, optimize the charging pile allocation plan and provide charging pile recommendations for users;

[0012] S5: After charging starts, by real-time monitoring the battery state and load power of the electric vehicle, combined with historical data, judge whether the charging state of the electric vehicle is normal, and generate a warning signal according to the abnormal state.

[0013] Preferably, the time period analysis model includes:

[0014] Predict the dynamic changes of charging demand. The model uses machine learning algorithms to accurately predict the charging demand in different time periods. Using the time period division results, further identify the peak time periods and trough time periods of charging demand, and adjust the weights in the prediction model in real time.

[0015] Preferably, the calculation method of the optimal charging coefficient is:

[0016] Taking the charging pile as the origin, mark all the charging pile data within the area with a radius of r1, calculate the regional vacancy coefficient KZ, count the occupancy situation of the charging piles within each time period, calculate the occupancy coefficient JX, and calculate the optimal charging coefficient YC according to the historical occupancy information and time period division. The formula is:

[0017]

[0018] In the formula, g4 and g5 are coefficient factors, KZ is the regional vacancy coefficient, and JX is the occupancy coefficient.

[0019] Preferably, the calculation method of the regional vacancy coefficient KZ is as follows:

[0020] Count the number of idle charging piles and occupied charging piles in the area, which are L1 and L2 in sequence, and calculate the idle occupancy ratio Zc. The formula is:

[0021]

[0022] Calculate the regional vacancy coefficient KZ. The formula is:

[0023] KZ = L1 * Zc * μ

[0024] In the formula, μ is a coefficient factor.

[0025] Preferably, the calculation method of the occupancy coefficient JX is as follows:

[0026] Count the occupancy interval GT i, compare it with the preset interval threshold, and calculate the occupancy coefficient JX. The formula is:

[0027]

[0028] In the formula, C1 is the total number of occupancies, Zb1 is the ratio of the number of occupancies greater than the threshold, PZ is the total value of over-intervals, and g1, g2, and g3 are coefficient factors.

[0029] Preferably, the charging demand prediction model is:

[0030] Based on the analysis results of time periods, combined with the historical occupancy data of charging piles, use machine learning algorithms to predict the trend of charging demand, and continuously adjust the prediction accuracy of the model by comparing historical data with real-time data.

[0031] Preferably, the steps of charging status monitoring and anomaly analysis include:

[0032] Collect the battery data of electric vehicles at preset time intervals, including battery temperature, current, voltage, and load power;

[0033] Based on the collected battery data, judge the battery health status during the charging process. If the battery temperature and load power deviate from the set threshold, generate an anomaly signal;

[0034] Use the load migration analysis method, combined with the calculation of time difference, to judge whether the load change conforms to the normal charging process. If an anomaly is found, the charging connection will be automatically disconnected.

[0035] Preferably, the anomaly signals include:

[0036] The battery temperature is too high or too low;

[0037] The load power fluctuates abnormally;

[0038] Stagnation or reverse current during the charging process;

[0039] The state of health of the battery does not meet the preset standard.

[0040] (III) Beneficial effects

[0041] Compared with the prior art, the present invention provides a statistical method for electric vehicle charging behavior based on period analysis, having the following beneficial effects:

[0042] By collecting charging behavior data within different time periods, the data collection is carried out through real-time feedback from in-vehicle sensors, charging pile information systems, and user intelligent terminals, and a time-periodized data set of charging demands is established; according to the changing trend of charging demands within different time periods, the charging demand data is divided into time periods, and a period analysis model is used to model the charging demands within each time period; based on the period analysis results, combined with the historical occupancy information of each charging pile, the optimal charging coefficient of each charging pile in different time periods is calculated; according to the optimal charging coefficient of the charging pile and the charging demand prediction results, the allocation scheme of the charging piles is optimized, and charging pile recommendations are provided for users; after charging starts, by real-time monitoring the battery state and load power of the electric vehicle, combined with historical data, it is judged whether the charging state of the electric vehicle is normal, and a warning signal is generated according to the abnormal state. It solves the problem that traditional analysis methods fail to effectively consider the volatility, seasonality of charging demands over time, and the personalized association with user behavior, resulting in large errors in practical applications, and has the advantages of improving the prediction accuracy of charging demands and the ability to capture time-periodic changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a schematic diagram of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0045] The present invention provides a technical solution: a statistical method for electric vehicle charging behavior based on period analysis, including the following steps:

[0046] S1: Collect charging behavior data within different time periods. The data collection is carried out through the real-time feedback of in-vehicle sensors, charging pile information systems, and user intelligent terminals, and a time-period dataset of charging demand is established.

[0047] Specifically, the time-period analysis model includes predicting the dynamic changes in charging demand. The model uses machine learning algorithms to accurately predict the charging demand in different time periods. Using the time-period division results, it further identifies the peak and trough periods of charging demand and adjusts the weights in the prediction model in real time.

[0048] In this implementation plan, charging behavior data, including time, location, occupancy of charging piles, battery status, etc., is collected and cleaned. Through feature engineering, variables that affect charging demand, such as time periods, weather conditions, historical charging data, etc., are extracted. The machine learning model is trained using historical data, and the model finds the optimal prediction parameters by optimizing the loss function. During the training process, the model will learn the change patterns of charging demand in different time periods, capturing potential trends and seasonal fluctuations. During the charging demand prediction process, the model will predict the charging demand in the future for a period of time based on the input data at the current time point, such as time periods, weather, and vehicle owner behavior. Time-period division is to identify the regular and periodic changes in charging demand, helping the model better adapt to the demand fluctuations in different time periods. The time-period division results can reflect the peak and trough periods of charging demand. According to the time-period change trends of historical data, for example, the differences in charging demand between weekdays and non-weekdays, and the fluctuations in charging demand between day and night, the 24 hours of a day are divided into multiple time periods, such as morning rush hour, afternoon, and night. This can be achieved through time series analysis and clustering algorithms, such as K-means clustering. The weights of the machine learning model are automatically learned during the training process, but through the time-period division results, the weights can be dynamically adjusted in different time periods, making the model more adaptable to different time periods. For example: During peak periods, such as morning rush hour to work or Friday night, the model can increase the feature weights related to traffic flow, weather conditions, holidays, etc., thereby enhancing the influence of these factors on charging demand prediction. During trough periods, such as early morning or non-peak periods, the model can reduce the feature weights related to traffic flow and increase the weights of factors such as grid load and charging pile status, because during this time period, the charging demand is relatively stable and has a greater impact on the grid load. During the charging demand prediction process, the system will dynamically adjust the weights in the model based on real-time data and time-period division results. For example, if it is found that there are large fluctuations in the charging demand in a certain time period, such as holidays or special weather conditions, the system can adjust the weights of the prediction model to strengthen the response ability to these fluctuations and improve the prediction accuracy.

[0049] The specific method of weight adjustment can be achieved through the following several ways:

[0050] Dynamic weighting: According to the real-time time period division, increase or decrease the weight of features. For example, during peak hours, features such as time, weather, and traffic flow are weighted, giving them higher prediction weights.

[0051] Learning based on feedback: In practical applications, the system can adjust the parameters of the model in real time according to the error between the prediction result and the actual charging demand. Through the online learning mechanism, the system can continuously adjust the prediction ability and weight of the model through real-time monitoring and data feedback, making the model more flexible and accurate.

[0052] Time period adaptive weight: Embed a time period adaptive mechanism in the machine learning model. For example, the LSTM model can train different sub-models within each time period, or use a weighting mechanism to weight each model according to the time period, adjusting the model's response ability to different time periods.

[0053] The weight adjustment formula is:

[0054] W(t) = W0 + ΔW(t)

[0055] Where, W(t) is the weight at time period t, W0 is the initial weight obtained through training, and ΔW(t) is the adjustment of the weight based on time period division and real-time data.

[0056] Comprehensive collection of charging behavior data in different time periods, specifically including real-time feedback from in-vehicle sensors, charging pile information systems, and user intelligent terminals. In-vehicle sensors can capture information such as the operating status, remaining battery power, and current location of electric vehicles. The charging pile information system records parameters such as the occupancy status, charging duration, and charging power of each charging pile, while the user intelligent terminal provides the charging reservation information, charging demand, and behavior data of the vehicle owner. Through the collaborative work of these data sources, charging behavior data is obtained and updated in real time. On this basis, the collected data is divided into time periods using time period analysis methods, and the charging demands are classified and sorted according to different time periods, forming a time period-based charging demand dataset. This dataset can not only accurately reflect the fluctuations and trends of charging demands in different time periods but also take into account the external influencing factors of charging demands. In this way, the dynamic changes of charging demands can be efficiently captured, providing data support for charging demand prediction, load scheduling, and the layout optimization of charging piles. The adoption of multi-source data fusion and real-time update mechanisms ensures the comprehensiveness, accuracy, and timeliness of the data, providing a reliable decision-making basis for the optimized management of electric vehicle charging systems.

[0057] S2: According to the changing trends of charging demands in different time periods, divide the charging demand data into time periods, and use a time period analysis model to model the charging demands in each time period.

[0058] In this implementation, by collecting charging behavior data in different time periods, including real-time data feedback from in-vehicle sensors, charging pile information systems, and user intelligent terminals, and classifying the data based on timestamps, the charging demands in a day are divided into periods such as morning rush hours, off-peak periods, and evening rush hours. Subsequently, time period analysis models, such as the ARIMA model and LSTM neural network based on time series, are used to model the charging demands in each period, so as to capture the charging demand patterns, fluctuation rules in each period and their relationships with external factors, such as weather, electricity price, traffic conditions, etc. The time period analysis model can identify the changing trends of charging demands in different time periods, accurately predict the charging demand fluctuations in each period, and provide a scientific basis for charging pile scheduling, load optimization, and charging infrastructure planning according to these prediction results. Through precise time period division and modeling, the accuracy of charging demand prediction can be effectively improved, the load scheduling of the charging system can be optimized, the charging pressure during peak hours can be avoided from being too large, and the effective utilization of charging piles during off-peak periods can be ensured, thus improving the overall efficiency of the charging system and the user experience.

[0059] S3: Based on the time period analysis results, combined with the historical occupancy information of each charging pile, calculate the optimal charging coefficient of each charging pile in different time periods, and construct a charging demand prediction model.

[0060] Specifically, the charging demand prediction model includes, based on the time period analysis results, combined with the occupancy historical data of the charging pile, using machine learning algorithms to predict the trend of charging demands, and continuously adjusting the prediction accuracy of the model by comparing historical data with real-time data.

[0061] In this implementation, through historical charging behavior data and time period analysis models, a charging demand prediction model is constructed. The real-time data is compared with the historical data, and the prediction results are adjusted using the model feedback to dynamically update the weights and parameters in the model, thereby continuously optimizing the prediction accuracy. This process ensures that the prediction of charging demands can more accurately reflect the actual changes by continuously training the model. By comparing in real time and dynamically adjusting the model, the periodic changes of charging demands can be accurately captured, and the response ability and prediction accuracy of the model to charging demand changes can be improved. Finally, an efficient prediction based on historical data and real-time data is achieved, which not only improves the accuracy of charging demand prediction, but also adapts to the demand fluctuations in different time periods, and optimizes the resource allocation and scheduling strategies of charging piles.

[0062] Identify the peak and trough periods of charging demand through the time period analysis model, and systematically process the historical occupancy data of each charging pile, including the charging pile number, occupancy start and end times, occupancy duration, etc. Using the historical occupancy information of the charging pile, the regional vacancy coefficient (KZ) of each charging pile can be calculated, that is, the ratio of idle and occupied charging piles in the area during a specific time period, and the occupancy coefficient (JX) of the charging pile can be calculated, that is, based on the relationship between the occupancy interval of the charging pile and the preset interval threshold, reflecting the usage efficiency and idle time of the charging pile during this time period. The optimal charging coefficient (YC) is obtained through formula calculation. This coefficient comprehensively considers the regional vacancy situation, occupancy efficiency and its time period characteristics of the charging pile, and can quantify the charging priority of each charging pile during a specific time period. Dynamically evaluate the charging priority of the charging pile according to the historical occupancy information and the change of time period demand, and sort the charging piles according to the optimal charging coefficient to ensure that users can preferentially select charging piles during peak demand periods, while avoiding excessive congestion during peak hours and improving the utilization efficiency of charging resources.

[0063] S4: Optimize the allocation plan of the charging piles according to the optimal charging coefficient of the charging piles and the charging demand prediction results, and provide charging pile recommendations for users.

[0064] In this implementation plan, use the charging demand prediction model to accurately predict the charging demand in different time periods, and identify the peak and trough periods of charging demand. Combining the optimal charging coefficient (YC) of each charging pile, this coefficient considers factors such as the historical occupancy situation, regional vacancy coefficient, and occupancy efficiency of the charging pile in different time periods, and can quantify the charging priority of the charging pile during a specific time period. According to the charging demand and the optimal charging coefficient of the charging pile, the system optimizes the allocation plan of the charging pile, and preferentially guides users with higher charging demands to charging piles with higher optimal charging coefficients to balance the load of each charging pile and avoid excessive congestion of some charging piles during peak periods. Through the intelligent recommendation algorithm, according to the user's current location, charging demand and time period characteristics, provide the best charging pile selection for users. Adopt big data analysis and real-time load scheduling strategies, comprehensively consider the time-period charging demand and the optimal charging coefficient of the charging pile, and adjust the allocation strategy of the charging pile in real time to ensure the optimal allocation of charging pile resources in different time periods. Effectively avoid the over-concentrated use of charging pile resources, optimize the allocation of charging resources, improve the use efficiency of charging piles, improve the user experience, reduce waiting time, and enhance the overall operating efficiency of the charging system.

[0065] S5: After the charging starts, judge whether the charging status of the electric vehicle is normal by real-time monitoring the battery status and load power of the electric vehicle, combined with historical data, and generate a warning signal according to the abnormal status.

[0066] Specifically, the steps of charging status monitoring and anomaly analysis include collecting battery data of the electric vehicle at preset time intervals, including battery temperature, current, voltage, and load power;

[0067] Based on the collected battery data, judge the battery health status during the charging process. If the battery temperature and load power deviate from the set thresholds, generate an anomaly signal;

[0068] Using the load migration analysis method, combined with time difference calculation, judge whether the load change conforms to the normal charging process. If an anomaly is found, the charging connection will be automatically disconnected.

[0069] In this implementation plan, load migration refers to the change in battery load during the charging process. During the actual charging process, the load should change smoothly. Abnormal load changes, such as large fluctuations or sudden rises and falls, may indicate battery problems. By calculating the time difference, judge the reasonableness of the load change to ensure the normal progress of the charging process. The calculation steps of load migration include calculating the load power change between two time points and calculating the time difference of the load change. The formula is:

[0070] ΔP = P(t2) - P(t1)

[0071] ΔT = t2 - t1

[0072] In the formula, P(t1) and P(t2) are the load power values at time points t1 and t2 respectively, ΔP is the load change amount, and ΔT is the time difference between the two time points. If the change of ΔP relative to ΔT does not conform to the normal charging mode, it is judged as an abnormal load change.

[0073] Data such as the temperature, charging current, voltage, and load power of the battery are collected in real time through in-vehicle sensors, and the load power (Rt) of the battery is calculated using this data. At the same time, the system compares the current working state of the battery with historical data to analyze the change trend of the battery during the charging process and whether abnormal conditions occur. For example, if there are abnormal fluctuations in the load power (Rt) or the battery temperature (Et) exceeds the preset safety threshold, the system will immediately determine that there may be a problem with the charging state of the battery. In addition, by establishing a load migration model, the system can monitor the load migration situation in real time based on the relationship between the battery temperature and the load power, and evaluate the normality of the battery state through the load migration calculation formula. If abnormalities occur during the charging process of the battery, such as too high temperature, unstable current and voltage, etc., the system will generate a charging abnormality signal, issue a warning in time, and notify the vehicle owner or the control system. In terms of technical effects, this method can effectively detect abnormal situations during the charging process and ensure the safety of the charging process through real-time monitoring of the battery state and comprehensive analysis of historical data. At the same time, through the warning mechanism, potential charging faults or battery problems can be detected in advance, reducing the losses caused by charging faults in electric vehicles and improving the user experience and charging safety.

[0074] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0075] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made in these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A statistical method for electric vehicle charging behavior based on time period analysis, characterized in that, Including the following steps: S1: Collect charging behavior data within different time periods. The data collection is carried out through the real-time feedback of in-vehicle sensors, charging pile information systems, and user intelligent terminals, and a time-period dataset of charging demand is established. S2: According to the charging demand change trend within different time periods, divide the charging demand data into time periods, and use a time-period analysis model to model the charging demand within each time period. S3: Based on the time-period analysis results, combined with the historical occupancy information of each charging pile, calculate the optimal charging coefficient of each charging pile in different time periods, and construct a charging demand prediction model. S4: According to the optimal charging coefficient of the charging pile and the charging demand prediction results, optimize the charging pile allocation scheme and provide charging pile recommendations for users. S5: After charging starts, by real-time monitoring the battery state and load power of the electric vehicle, combined with historical data, judge whether the charging state of the electric vehicle is normal, and generate a warning signal according to the abnormal state.

2. The statistical method for electric vehicle charging behavior based on time period analysis according to claim 1, wherein The time-period analysis model includes: Predict the dynamic changes of charging demand. The model uses machine learning algorithms to accurately predict the charging demand in different time periods, and uses the time-period division results to further identify the peak and trough time periods of charging demand, and adjust the weights in the prediction model in real time.

3. The statistical method for electric vehicle charging behavior based on time period analysis according to claim 1, wherein The calculation method of the optimal charging coefficient is: Taking the charging pile as the origin, mark all the charging pile data within the area with a radius of r1, calculate the regional vacancy coefficient KZ, count the occupancy situation of the charging piles within each time period, calculate the occupancy coefficient JX, and calculate the optimal charging coefficient YC according to the historical occupancy information and time-period division. The formula is: In the formula, g4 and g5 are coefficient factors, KZ is the regional vacancy coefficient, and JX is the occupancy coefficient.

4. The statistical method for electric vehicle charging behavior based on time period analysis according to claim 3, characterized in that The calculation method of the regional vacancy coefficient KZ is: Count the number of idle charging piles and occupied charging piles in this area, which are L1 and L2 in sequence, and calculate the idle ratio Zc. The formula is: Calculate the regional vacancy coefficient KZ. The formula is: KZ = L1 * Zc * μ In the formula, μ is a coefficient factor.

5. The statistical method for electric vehicle charging behavior based on period analysis according to claim 3, wherein The calculation method of the occupancy coefficient JX is: Count the occupancy interval GTi, compare it with the preset interval threshold, and calculate the occupancy coefficient JX. The formula is: In the formula, C1 is the total number of occupancies, Zb1 is the proportion of the number of occupancies greater than the threshold, PZ is the total value of exceeding the interval, and g1, g2, and g3 are coefficient factors.

6. The statistical method for electric vehicle charging behavior based on time period analysis according to claim 1, characterized in that The charging demand prediction model includes: Based on the time-period analysis results, combined with the historical occupancy data of the charging pile, use machine learning algorithms to predict the trend of charging demand, and continuously adjust the prediction accuracy of the model by comparing historical data with real-time data.

7. The statistical method for electric vehicle charging behavior based on time period analysis according to claim 1, wherein, The steps of charging state monitoring and abnormal analysis include: Collect the battery data of the electric vehicle at preset time intervals, including battery temperature, current, voltage, and load power. Based on the collected battery data, judge the battery health state during charging. If the battery temperature and load power deviate from the set threshold, generate an abnormal signal. Use the load migration analysis method, combined with the time difference calculation, to judge whether the load change conforms to the normal charging process. If an abnormality is found, the charging connection will be automatically disconnected.

8. The statistical method for electric vehicle charging behavior based on time period analysis according to claim 7, characterized in that The abnormal signals include: The battery temperature is too high or too low; Abnormal fluctuations in load power; Stagnation or reverse current during charging; The battery health status does not meet the preset standards.

Images