Two-wheeled vehicle charging early warning evaluation method and system based on charging pile AC side

Abstract

The invention discloses a two-wheeled vehicle charging early warning evaluation method and system based on the AC side of a charging pile, and the method comprises the following steps: S1, collecting the electrical parameter data of the AC side in the charging process of a two-wheeled vehicle through a charging device, and binding the electrical parameter data with the unique identity of the two-wheeled vehicle; s2, extracting charging curve morphological characteristics representing the health state of the battery; s3, establishing an individualized longitudinal dynamic baseline of the two-wheeled vehicle, and realizing charging abnormity judgment through standardized deviation degree calculation in combination with the transverse reference baselines of the batteries of the same type; and S4, generating and outputting corresponding early warning information. According to the invention, the electrical parameters of the AC side are collected and the unique identifier of the vehicle is bound, so that accurate positioning of the vehicle is realized without depending on BMS or DC side data, and the defect of no positioning in macroscopic monitoring is overcome; health-related curve characteristics are extracted, an individualized longitudinal baseline is established and is combined with the same type of transverse baselines, and the problem of low evaluation precision is solved by standardizing deviation degree to discriminate anomalies and eliminate individual difference interference.

Description

Technical Field

[0001] This application relates to the field of charging safety and health management, specifically to a method and system for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile. Background Technology

[0002] Electric bicycles, also known as electric two-wheelers, pose a significant challenge to public safety through charging safety and battery health management. As the existing market for electric two-wheelers continues to expand, safety issues arising during charging are becoming increasingly prominent, necessitating reliable monitoring and evaluation technologies.

[0003] Currently, the industry has developed two main technical approaches for charging monitoring and battery health assessment of electric two-wheelers, but both have significant drawbacks in fully adapting to the application scenarios of electric two-wheelers. The first technical approach originates from the diagnostic approach for new energy vehicles. It analyzes the fine DC-side waveforms of the interaction between the charging pile and the vehicle, uses waveform decomposition technology to remove interference, and then compares the waveforms with a pre-set universal battery reference model library to achieve assessment. However, many lead-acid batteries in electric two-wheelers are not equipped with a battery management system (BMS), making it impossible to obtain battery charging status data. Furthermore, the battery types and capacities of different brands and models of electric two-wheelers exhibit extreme heterogeneity, making the establishment of a universal model library impractical. This approach is difficult to apply to the individualized and precise management of electric two-wheelers.

[0004] The second technical approach involves macroscopic monitoring from the electricity consumption side. It collects a mixed total load signal at the main power inlet of the parking shed and indirectly identifies charging behavior using a non-intrusive load decomposition algorithm. While this method improves coverage, it has inherent limitations: the low signal-to-noise ratio in the total load signal restricts monitoring accuracy; the algorithm is complex and cannot pinpoint specific charging vehicles, making it difficult to diagnose root causes of anomalies and resulting in poor practical operability.

[0005] In summary, existing technologies are caught in a polarization between high-precision precision models and low-precision macroscopic estimations. They lack the data acquisition hardware foundation and engineering implementation architecture that match the massive, heterogeneous, and low-cost characteristics of electric two-wheelers. They cannot simultaneously meet the practical needs of universality, accuracy, and feasibility, and it is difficult to achieve effective charging anomaly warning and battery health status assessment for the massive number of existing electric two-wheelers. Summary of the Invention

[0006] In view of this, the present invention proposes a two-wheeled vehicle charging early warning assessment method based on the AC side of the charging pile, which aims to solve the above problems.

[0007] Includes the following steps: S1: Collect electrical parameter data on the AC side during the charging process of the two-wheeled vehicle through the charging equipment, and bind the electrical parameter data with the unique identifier of the two-wheeled vehicle; S2: Process the bound electrical parameter data and extract the charging curve morphology features that characterize the battery health status; S3: Based on the charging curve morphology characteristics, establish an individualized longitudinal dynamic baseline for the two-wheeled vehicle, and combine it with the lateral reference baseline of the same type of battery to achieve charging anomaly detection through standardized deviation calculation; S4: Generate and output the corresponding early warning information based on the anomaly detection results.

[0008] Furthermore, in step S1, the electrical parameter data includes voltage, current, power, power, and charging time, and is acquired through a separate metering acquisition mode. The unique identifier is strongly bound to the charging session through NFC card swiping or QR code scanning charging process.

[0009] Furthermore, step S1 also includes obtaining the core battery parameters of the two-wheeled vehicle. The core battery parameters are stored in association with the unique identifier, and the core battery parameters include battery voltage, capacity, and type information.

[0010] Furthermore, in step S2, the charging curve morphology features include the duration of the constant current phase and the slope of the current decrease curve in the constant voltage phase. The duration of the constant current phase is calculated by combining the sliding window variance detection method with the state machine to identify the constant current phase. The slope of the current decrease curve during the constant voltage stage is obtained by fitting the current value during the constant voltage stage using the least squares method after performing a logarithmic transformation.

[0011] Furthermore, the implementation process of the sliding window variance detection method includes: performing a moving average filter on the acquired current time series; constructing a sliding window with a preset number of data points and calculating the variance of the power data within each sliding window; setting a variance threshold, and determining the transition from the constant current stage to the constant voltage stage when the window variance exceeds the variance threshold. The start time of the constant current stage is the time when the current first reaches a first preset proportion of the current rating, and the end time of the constant current stage is the time when the variance threshold is exceeded.

[0012] Furthermore, the calculation process of the current decline curve slope in the constant voltage stage includes: taking the natural logarithm of the current value collected in the constant voltage stage to obtain the transformed data point (Xi, Yi), where Xi is the time t after the start of the constant voltage stage, and Yi is the natural logarithm of the current at the corresponding time ln(I); fitting the straight line ln(I)=a+b*t using the least squares method, where a is the intercept, b is the slope, and b is the slope of the current decline curve in the constant voltage stage; The start time of the constant voltage stage is the end time of the constant current stage, and the end time of the constant voltage stage is when the current drops to the second preset proportion of the current rating for the first time during the constant voltage stage.

[0013] Furthermore, in step S3, the process of establishing the individualized longitudinal dynamic baseline includes: setting an initial learning period, during which charging data is collected a preset number of times; based on the morphological characteristics of the historical charging curves during the learning period, calculating the mean slope μ and standard deviation σ of the current drop curve in the constant voltage stage, and forming the individualized longitudinal dynamic baseline.

[0014] Furthermore, in step S3, the standardized deviation calculation includes longitudinal deviation Z1 and lateral deviation Z2. The formula for calculating the longitudinal deviation Z1 is Z1=(μ0-μ) / σ, where μ0 is the slope of the current decrease curve during the constant voltage stage of the current charging; the lateral reference baseline is the typical characteristic value range of the corresponding battery model in the knowledge base, including the typical mean μ_ref and the typical standard deviation σ_ref; the formula for calculating the lateral deviation Z2 is Z2=(μ0-μ_ref) / σ_ref; when |Z1|>M1 and |Z2|>M2, it is determined to be a charging abnormality, where M1 is the longitudinal deviation ratio parameter and M2 is the lateral deviation ratio parameter.

[0015] Furthermore, in step S4, the type of warning information is determined by the duration of the constant current phase, the degree of deviation between |Z1| and M1, and the degree of deviation between |Z2| and M2. The types of warning information include yellow warning, orange warning, and red warning. The warning information is pushed to users and maintenance personnel through preset communication channels, including mini-programs, SMS, and telephone.

[0016] Accordingly, this invention also proposes a two-wheeled vehicle charging early warning and assessment system based on the AC side of a charging pile, for implementing the above method, including: Data acquisition and binding module: used to collect electrical parameter data on the AC side during the charging process of the two-wheeled vehicle through the charging equipment, and bind the electrical parameter data with the unique identifier of the two-wheeled vehicle; Feature extraction module: used to process the bound electrical parameter data and extract the charging curve morphology features that characterize the battery health status; Baseline establishment and anomaly detection module: Based on the charging curve morphology characteristics, it establishes an individualized longitudinal dynamic baseline for the two-wheeled vehicle and combines it with the lateral reference baseline of the same type of battery to detect charging anomalies through standardized deviation calculation. Early warning output module: Used to generate and output corresponding early warning information based on the anomaly detection results.

[0017] This invention addresses the issue of inability to cover heterogeneous existing vehicles by collecting AC-side electrical parameters and binding them to a unique vehicle identifier, eliminating the need for BMS or DC-side data. It also achieves precise vehicle positioning, overcoming the shortcomings of macro-monitoring which lacks location tracking. Furthermore, it extracts health-related curve features, establishes individualized longitudinal baselines, and combines them with similar lateral baselines. By standardizing deviations, it identifies anomalies, eliminating individual differences and resolving the problem of low assessment accuracy. It boasts strong universality, covering various types of two-wheeled vehicles; individual baselines reduce misjudgments and ensure high accuracy; it is highly implementable; and its process is compatible with existing charging equipment, effectively achieving charging warnings and health assessments. Attached Figure Description

[0018] Figure 1 This is a flowchart of a two-wheeled vehicle charging early warning assessment method based on the AC side of a charging pile according to the present invention; Figure 2 This is a three-stage charging power curve of a 48V lead-acid battery; Figure 3 This is a comparison chart of the charging power curves of a healthy battery and an aged battery; Figure 4 This is a diagram illustrating the composition of a two-wheeled vehicle charging early warning and assessment system based on the AC side of a charging pile, according to the present invention. Figure 5 This is an architecture diagram of a two-wheeled vehicle charging early warning and evaluation system based on the AC side of a charging pile according to the present invention. Detailed Implementation

[0019] Electric bicycles, also known as electric two-wheelers, pose a significant challenge to public safety through charging safety and battery health management. As the existing market for electric two-wheelers continues to expand, safety issues arising during charging are becoming increasingly prominent, necessitating reliable monitoring and evaluation technologies.

[0020] Currently, the industry has developed two main technical approaches for charging monitoring and battery health assessment of electric two-wheelers, but both have significant drawbacks in fully adapting to the application scenarios of electric two-wheelers. The first technical approach originates from the diagnostic approach for new energy vehicles. It analyzes the fine DC-side waveforms of the interaction between the charging pile and the vehicle, uses waveform decomposition technology to remove interference, and then compares the waveforms with a pre-set universal battery reference model library to achieve assessment. However, many lead-acid batteries in electric two-wheelers are not equipped with a battery management system (BMS), making it impossible to obtain battery charging status data. Furthermore, the battery types and capacities of different brands and models of electric two-wheelers exhibit extreme heterogeneity, making the establishment of a universal model library impractical. This approach is difficult to apply to the individualized and precise management of electric two-wheelers.

[0021] The second technical approach involves macroscopic monitoring from the electricity consumption side. It collects a mixed total load signal at the main power inlet of the parking shed and indirectly identifies charging behavior using a non-intrusive load decomposition algorithm. While this method improves coverage, it has inherent limitations: the low signal-to-noise ratio in the total load signal restricts monitoring accuracy; the algorithm is complex and cannot pinpoint specific charging vehicles, making it difficult to diagnose root causes of anomalies and resulting in poor practical operability.

[0022] In summary, existing technologies are caught in a polarization between high-precision models and low-precision macroscopic estimations. The core reason is the lack of data acquisition hardware infrastructure and engineering implementation architecture that match the massive, heterogeneous, and low-cost characteristics of electric two-wheelers. This makes it impossible to simultaneously meet the practical needs of universality, accuracy, and feasibility, and to achieve effective charging anomaly warning and battery health status assessment for the massive number of existing electric two-wheelers.

[0023] In view of this, the present invention proposes a two-wheeled vehicle charging early warning assessment method based on the AC side of the charging pile, which aims to solve the above problems.

[0024] The AC side, or alternating current side, in the electric two-wheeler charging scenario of this invention based on charging piles, specifically refers to the side where the charging pile connects to the external 220V residential AC power grid, and the transmission link related to the charging pile providing AC power to the electric two-wheeler charger. This side transmits and processes AC power, and the electrical parameters that can be collected (such as voltage, current, power, etc.) are all AC electrical signals, without relying on the vehicle battery management system (BMS) or special data from the DC output of the charging pile.

[0025] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the various embodiments of this application will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been presented in the various embodiments of this application to enable the reader to better understand this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various changes and modifications based on the following embodiments.

[0026] In the following description, certain specific details are set forth for the purpose of illustrating various disclosed embodiments in order to provide a thorough understanding of the various disclosed embodiments. However, those skilled in the art will recognize that embodiments may be practiced without one or more of these specific details. In other instances, well-known apparatuses, structures, and techniques associated with this application may not have been shown or described in detail to avoid unnecessarily obscuring the description of the embodiments.

[0027] Unless the context requires otherwise, throughout the specification and claims, the word “comprising” and its variations, such as “including” and “having”, shall be understood to have an open, inclusive meaning, that is, to be interpreted as “including, but not limited to”.

[0028] The embodiments of this application will be described in detail below with reference to the accompanying drawings to provide a clearer understanding of the purpose, features, and advantages of this application. It should be understood that the embodiments shown in the drawings are not intended to limit the scope of this application, but are merely for illustrating the essential spirit of the technical solution of this application.

[0029] Throughout this specification, references to "an embodiment" or "an embodiment" indicate that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. Therefore, the appearance of "in an embodiment" or "an embodiment" in various places throughout the specification does not necessarily refer to the same embodiment. Furthermore, a particular feature, structure, or characteristic may be combined in any manner in one or more embodiments.

[0030] like Figure 1 As shown, the method includes the following steps: S1: Collect electrical parameter data on the AC side during the charging process of the two-wheeled vehicle through the charging equipment, and bind the electrical parameter data with the unique identifier of the two-wheeled vehicle; S2: Process the bound electrical parameter data and extract the charging curve morphology features that characterize the battery health status; S3: Based on the charging curve morphology characteristics, establish an individualized longitudinal dynamic baseline for the two-wheeled vehicle, and combine it with the lateral reference baseline of the same type of battery to achieve charging anomaly detection through standardized deviation calculation; S4: Generate and output the corresponding early warning information based on the anomaly detection results.

[0031] Vehicle identification is a prerequisite for personalized longitudinal dynamic baseline analysis. This invention provides two charging methods: NFC card charging and WeChat mini-program QR code charging. For vehicle identification, card charging identifies which vehicle is being charged. WeChat mini-program QR code charging requires vehicle registration; by default, a user has one electric two-wheeler, and payment is made immediately after scanning the code. If a user has multiple vehicles, they must select which vehicle to charge, and the user is prompted. Incorrect selection triggers an alarm or even terminates charging.

[0032] Therefore, in step S1, the electrical parameter data of the AC side during the charging process of the two-wheeled vehicle is collected by the charging equipment, and the electrical parameter data is bound with the unique identification of the two-wheeled vehicle to achieve traceability to the specific vehicle.

[0033] The electrical parameter data includes voltage, current, power, battery level, and charging time. To ensure the purity and traceability of the data, a separate metering collection mode is adopted, which avoids interference from mixed charging data from different vehicles. A unique identifier is strongly bound to the charging session via NFC card swiping or QR code scanning. When a user charges by NFC card swiping, each NFC card corresponds one-to-one with a specific two-wheeled vehicle; the card swiping action triggers the charging session and simultaneously completes the identity binding. When a user charges by QR code scanning, they must first scan the code to associate it with the unique identifier of the two-wheeled vehicle under their account. After confirmation, the charging session is initiated, thereby achieving precise binding of electrical parameter data to the vehicle.

[0034] Meanwhile, to facilitate subsequent lateral reference baseline retrieval and more accurate feature analysis, step S1 also includes obtaining the core battery parameters of the two-wheeled vehicle. These core battery parameters include battery voltage, capacity, and type information. The acquisition process is completed by guiding the user to input or select the corresponding parameters on the interactive interface before charging or when using the system for the first time, such as "48V20Ah lead-acid battery" or "48V30Ah ternary lithium battery". After input, the core battery parameters are associated with the unique identifier of the two-wheeled vehicle and stored in the system database for quick retrieval and use in subsequent steps.

[0035] like Figure 2 As shown in the charging process of a 48V lead-acid battery, the charging of electric vehicles is divided into three stages: constant current (CC), constant voltage (CV), and float charging (trickle charging). How to identify these three stages, especially the characteristics of constant current (CC) and constant voltage (CV) on the AC side (220V), is the key to analyzing battery health (SOH).

[0036] The characteristics of the constant current stage are: the current value is stable at a relatively high level with very little fluctuation.

[0037] The characteristic of the constant voltage stage is that the current decreases approximately exponentially over time.

[0038] Furthermore, healthy batteries and aged batteries exhibit significant differences in their power curve performance. Figure 3 As shown: For a healthy battery, in the constant current phase: the charger charges at the maximum permissible current. During this phase, the current remains at a stable level, and thus the power also remains relatively stable. In the constant voltage phase: the charger maintains a constant voltage, and the current exhibits a smooth, continuous, and relatively steep decreasing curve, eventually dropping smoothly to the trickle charging current. In aging batteries, the constant current phase is significantly shortened. The current begins to decrease rapidly, making it impossible to sustain high-current charging for extended periods. In the constant voltage phase, the current decreases very slowly, the curve becomes flat, and it may even remain at a relatively high current value for a long time without decreasing.

[0039] Therefore, in step S2, the bound electrical parameter data is processed to extract the charging curve morphology features that characterize the battery health status, and the original electrical parameter data is transformed into an effective indicator that can reflect the internal health status of the battery, thus solving the problem in the prior art that it is difficult to accurately assess battery health through AC side data.

[0040] The charging curve morphology features include the duration of the constant current phase and the slope of the current drop curve in the constant voltage phase. These two features reflect battery performance from different stages of the charging process. The duration of the constant current phase is related to the battery's ability to accept high current charging, while the slope of the current drop curve in the constant voltage phase is related to the battery's capacity and internal impedance health.

[0041] The duration of the constant current phase is calculated by combining the sliding window variance detection method with the state machine to identify the constant current phase.

[0042] The specific implementation process is as follows: First, the collected current time series is filtered by moving average to eliminate the interference of small spikes and noise on the data; then, a sliding window is formed by a preset number of data points, and the window is slid across the power data and the variance of the power data in each sliding window is calculated; then, a variance threshold is set. Since the power is stable due to the stable current in the constant current stage, the window variance is small. When the window variance exceeds the variance threshold, it indicates that the current begins to change continuously, and it is determined that the stage has changed from the constant current stage to the constant voltage stage.

[0043] The start time of the constant current stage is the time when the current first reaches the first preset proportion of the current rating. This first preset proportion is set according to common battery charging characteristics to ensure the accuracy of the start time identification. The end time of the constant current stage is the time when the variance threshold is exceeded. The duration of the constant current stage can be obtained by the difference between the start time and the end time.

[0044] The slope of the current decrease curve during the constant voltage stage is obtained by performing a logarithmic transformation on the current value during the constant voltage stage and then fitting it using the least squares method.

[0045] The start time of the constant voltage stage is the end time of the constant current stage. The end time of the constant voltage stage is when the current drops to the second preset ratio of the current rating for the first time during the constant voltage stage. This second preset ratio is set according to the battery trickle charging start conditions.

[0046] The specific calculation process is as follows: the natural logarithm of the current value collected during the constant voltage stage is taken to transform the exponentially decreasing current change into an approximately linear change, resulting in the transformed data points (Xi, Yi), where Xi is the time t after the start of the constant voltage stage, and Yi is the natural logarithm of the current at the corresponding time, ln(I); then, the least squares method is used to fit the straight line ln(I) = a + b*t, where a is the intercept and b is the slope. This slope b is the slope of the current decrease curve during the constant voltage stage. The magnitude of the slope directly reflects the rate of decrease of the battery current, and thus reflects the health status of the battery.

[0047] Step S3 establishes an individualized longitudinal dynamic baseline for the two-wheeled vehicle based on the morphological characteristics of the charging curve, and combines it with the lateral reference baseline of the same type of battery. The charging anomaly is determined by standardizing the deviation calculation. This step aims to solve the problem that it is difficult to overcome individual differences of vehicles and accurately determine anomalies in the existing technology.

[0048] The process of establishing an individualized longitudinal dynamic baseline includes a preset initial learning period during which charging data is collected a preset number of times. This preset number of times is set according to the battery charging characteristics and data statistics requirements to ensure that sufficient historical data can be collected to reflect the normal charging state of the vehicle. Based on the morphological characteristics of the historical charging curves during the learning period, the mean μ and standard deviation σ of the current drop curve in the constant voltage stage are calculated. The mean μ represents the average level of the slope when the vehicle is charging normally, and the standard deviation σ reflects the fluctuation range of the slope under normal conditions. Together, they constitute the individualized longitudinal dynamic baseline, which can accurately match the charging characteristics of a specific vehicle and avoid the deviations caused by the general model.

[0049] The standardized deviation calculation includes longitudinal deviation Z1 and lateral deviation Z2. Longitudinal deviation Z1 is used to determine the degree of deviation between the current charging characteristics and the vehicle's historical normal state. Its calculation formula is Z1=(μ0-μ) / σ, where μ0 is the slope of the current decrease curve in the constant voltage stage of the current charging. The lateral reference baseline is the typical characteristic value range of the corresponding battery model in the knowledge base. This knowledge base is built and continuously updated based on massive charging data of the same type of battery, including the typical mean μ_ref and the typical standard deviation σ_ref. Lateral deviation Z2 is used to determine the degree of deviation between the current charging characteristics and the normal state of a healthy battery of the same type. Its calculation formula is Z2=(μ0-μ_ref) / σ_ref.

[0050] When |Z1|>M1 and |Z2|>M2, it is determined to be a charging abnormality. M1 is the longitudinal deviation ratio parameter and M2 is the lateral deviation ratio parameter. M1 and M2 are set based on a large amount of experimental data and safety warning requirements to ensure that the false alarm rate is reduced while effectively identifying abnormalities.

[0051] Furthermore, step S4 generates and outputs corresponding early warning information based on the anomaly identification results. This step transforms the anomaly identification results into practically applicable early warning signals, ensuring that users and maintenance personnel can promptly be aware of and handle charging anomalies.

[0052] The type of warning information is determined by the degree of deviation between |Z1| and M1, and the degree of deviation between |Z2| and M2. Different degrees of anomaly correspond to different types of warning information, including yellow warning, orange warning and red warning.

[0053] In one specific embodiment, the following classification logic and rule set are defined: The first step is to divide the charging session into eight different states, each corresponding to a different stage or situation in the charging process, as defined below: STANDBY state: This is the idle state of the charging device. At this time, the charging device has not established a charging connection with any two-wheeled vehicle, nor has it started any charging-related operations.

[0054] AUTHENTICATED status: This indicates that NFC card swipe authentication or WeChat QR code scanning has been completed, the charging device has confirmed the user's identity and charging permissions, and is in a state of preparing to start charging.

[0055] CC_CHARGING state: This is the constant current charging stage. At this time, the charging equipment charges the two-wheeled vehicle battery with a constant current, which is the main stage for the battery to quickly replenish its power.

[0056] CV_CHARGING state: This is the constant voltage charging stage state. It enters this state after the constant current charging stage ends. The charging device maintains a constant voltage output, and the battery current gradually decreases until it approaches the trickle charging current.

[0057] TOP_OFF ​​state: This is the trickle charge stage. At this time, the charging current drops to a low level to supplement the battery with a small current, ensuring that the battery is fully charged and avoiding overcharging.

[0058] COMPLETED status: This indicates that the charging process has ended normally, the battery is fully charged, the charging device automatically stops charging output, and the charging session ends normally.

[0059] INTERRUPTED status: This indicates an interrupted charging state, usually caused by physical connection problems, such as poor contact of the charging plug or circuit failure, which causes the charging process to stop unexpectedly.

[0060] ANOMALY_HOLD state: This is the abnormal holding state. When a suspected abnormality is detected but the type of abnormality cannot be determined or the development trend of the abnormality needs to be further observed, the charging session enters this state in order to continuously collect data for subsequent judgment.

[0061] The second step is to classify and identify charging anomalies, including the following five categories: (a) Identification of suspected vehicle or battery replacement Triggering Timing: At the start of the AUTHENTICATED state, or at the initial stage of CC_CHARGING. Judgment Characteristics: A significant difference between the maximum charging power and the historical baseline, exceeding 60W; and an absolute value of the lateral deviation (compared to a healthy battery of the corresponding model extracted from the knowledge base) greater than 2, or an absolute value of the lateral deviation greater than 3 during the learning period. Edge Decision: Inferring a suspected vehicle or battery replacement. Warning Action: Generate a yellow warning, report to the platform and record it, without immediately stopping charging; send a message to the user via WeChat mini-program to confirm whether a replacement has occurred. If the user confirms a replacement, update the baseline for that user ID with the new data; if the power exceeds the charging pile's power, immediately stop charging and report to the platform, and send a message to the user via WeChat mini-program.

[0062] (ii) Identification of suspected battery aging Triggering Timing: CV_CHARGING state. Judgment Characteristics: During the constant voltage phase, the absolute value of the current decrease slope is significantly less than the baseline, and the absolute value of the vertical deviation is greater than 3; and the absolute value of the lateral deviation (compared to a healthy battery of the corresponding model extracted from the knowledge base) is greater than 2, or the absolute value of the lateral deviation is greater than 3 during the learning period. Edge Decision: The inability to reduce the current and the low energy input during the constant voltage phase are core indicators of capacity decay, suggesting suspected battery aging. Warning Action: Generate an orange warning, report to the platform, and indicate suspected battery aging.

[0063] (III) Identification of suspected loose connection points Triggering Timing: CC_CHARGING or CV_CHARGING state. Key Characteristics: A sudden increase in the short-term standard deviation of real-time power; continuous, irregular spikes and drops in current, followed by recovery. Edge Decision: Unlike the continuous deterioration of battery faults, this phenomenon manifests as intermittent fluctuations, suggesting a possible loose connection point. Warning Action: Generate a red warning, the highest priority safety alert; immediately suspend charging, report to the platform, and notify the user via WeChat and SMS of the connection point abnormality and potential overheating risk.

[0064] (iv) Determining if the plug is disconnected Triggering Timing: Any state. Judgment Characteristics: Charging power suddenly drops to near zero, charging power consumption is less than 10W (this is used as the criterion because the charger itself consumes power). Edge Decision: Physical connection is broken, inferred to be a plug detachment. Warning Action: Terminate the session, push a message to the user indicating charging is finished, without generating a warning; if the user did not actively unplug the plug, the plug detachment can be addressed on-site before recharging.

[0065] (v) Judgment based on low charging power and suspected user's low previous power consumption and high initial SOC. Triggering timing: At the start of the session. Identification characteristics: Shortly after entering the CC_CHARGING state (e.g., within tens of seconds), it transitions to CV_CHARGING, but the current drop curve slope and final charging energy during the CV_CHARGING phase conform to the historical baseline. Edge decision: This is a normal phenomenon, inferred to be due to low power consumption and a high initial SOC in the previous user session. Warning action: Normal charging, but this type of session data should not be used to update relevant baselines for the constant current phase (e.g., the duration of the constant current phase) to avoid polluting the baseline.

[0066] Warning information is pushed to users and maintenance personnel through preset communication channels, including mini-programs, SMS, and telephone, to ensure timely delivery of warning information. For yellow warnings, priority is given to pushing via mini-programs; for orange warnings, a combination of mini-programs and SMS is used; and for red warnings, multiple pushes are made via mini-programs, SMS, and telephone to maximize the chances that relevant personnel receive the warning information in a timely manner and take appropriate measures.

[0067] Furthermore, to implement the above method, this invention also proposes a two-wheeled vehicle charging early warning and assessment system based on the AC side of a charging pile, such as... Figure 4 As shown, it includes a data acquisition and binding module, a feature extraction module, a baseline establishment and anomaly detection module, and an early warning output module. These modules work together to realize the functions of the above-mentioned two-wheeled vehicle charging early warning assessment method based on the AC side of the charging pile.

[0068] The data acquisition and binding module is used to collect electrical parameter data on the AC side during the charging process of the two-wheeled vehicle through the charging equipment, and bind the electrical parameter data with the unique identification of the two-wheeled vehicle. This module is equipped with a data acquisition unit and an identification binding unit. The data acquisition unit corresponds to the separate metering acquisition mode in step S1, realizing the acquisition of voltage, current, power, power and charging time. The identification binding unit corresponds to the NFC card swiping or QR code binding process in step S1, and also has the function of acquiring and storing battery core parameters, associating and storing battery core parameters with the unique identification.

[0069] The feature extraction module is used to process the bound electrical parameter data and extract the charging curve morphology features that characterize the battery's health status. This module includes a data processing unit, a constant current stage feature extraction unit, and a constant voltage stage feature extraction unit. The data processing unit implements moving average filtering of the current time series. The constant current stage feature extraction unit completes the identification and duration calculation of the constant current stage by combining the sliding window variance detection method with a state machine. The constant voltage stage feature extraction unit completes the slope calculation of the current decrease curve in the constant voltage stage by using logarithmic transformation and least squares fitting, corresponding to the function of step S2.

[0070] The baseline establishment and anomaly detection module is used to establish an individualized longitudinal dynamic baseline for the two-wheeled vehicle based on the charging curve morphology characteristics, and to detect charging anomalies by combining the lateral reference baseline of the same type of battery through standardized deviation calculation. This module includes a baseline establishment unit and an anomaly detection unit. The baseline establishment unit presets an initial learning period, calculates the mean μ and standard deviation σ of historical features, and establishes an individualized longitudinal dynamic baseline. The anomaly detection unit calls the lateral reference baseline in the knowledge base, calculates the longitudinal deviation Z1 and lateral deviation Z2, and completes the anomaly detection by combining the duration of the constant current stage, which corresponds to the function of step S3.

[0071] The early warning output module is used to generate and output corresponding early warning information based on the anomaly identification results. This module includes an early warning type determination unit and an information push unit. The early warning type determination unit determines whether it is a yellow, orange, or red early warning based on the duration of the constant current stage and the degree of deviation. The information push unit pushes the early warning information to users and maintenance personnel through communication channels such as mini-programs, SMS, and telephone, which corresponds to the function in step S4.

[0072] In one embodiment, an architectural approach to the system of the present invention is also provided, such as... Figure 5 As shown, it includes an edge layer, a platform layer, and an application layer, as detailed below: 1. Edge Layer - Deployed on charging stations / gateways Function: Responsible for real-time data collection, preliminary processing, and emergency response.

[0073] Data acquisition: The charging pile hardware continuously collects raw data such as voltage, current, power, and ambient temperature on the AC side.

[0074] Feature extraction engine: A lightweight algorithm is built into the terminal to calculate the core features of this charging session (such as average power and effective charging time) in real time and generate feature vectors. This greatly reduces the amount of data that needs to be uploaded.

[0075] Local real-time judgment engine: The system includes the latest baseline model (μ,σ) of the vehicle, downloaded from the cloud, where μ is the mean and σ is the standard deviation.

[0076] Real-time Z-Score calculation is performed on the charging characteristics.

[0077] Local, real-time decisions are made only for urgent anomalies (such as extremely high Z-Score, drastic power fluctuations, power over-limit, etc., which pose immediate safety risks) and operations such as power-off are performed.

[0078] The remaining ordinary data and feature vectors are encrypted and then uploaded to the cloud.

[0079] Low latency and high reliability. It can perform the most critical security protection functions even in the event of a network outage.

[0080] 2. Platform Layer - Cloud-based Intelligent Center Function: It is the brain of the system, responsible for processing, storing, modeling, and making core decisions on massive amounts of data.

[0081] Data access: Receive uplink data from all edge terminals concurrently via message queues (such as Kafka).

[0082] Integrated stream and batch processing engine: Uses frameworks such as Spark to perform real-time stream processing and batch processing of data.

[0083] Big Data Warehouse: Stores all raw, cleaned, and historical feature data, as well as baseline model parameters (μ, σ) for all vehicles.

[0084] Analysis Service Center (Core): Individual longitudinal baseline model: dynamically updates the unique baseline (mean μ and standard deviation σ) for each vehicle.

[0085] Anomaly pattern classifier: performs root cause analysis on behaviors that are identified as anomalous (e.g., classified as "capacity decay" or "connection failure").

[0086] Alarm and ticket engine: Generates different levels of alert information (mini-program push, SMS) based on the output of AI services.

[0087] Centralized and intelligent. Enables complex model calculations, ensuring continuous iteration and optimization of the algorithm model.

[0088] 3. Application Layer - Business Services Function: Transform the processing results into usable services and present them to end users and maintenance personnel.

[0089] Real-time Alarm Center: Disseminates alert information to users and maintenance personnel through multiple channels (WeChat mini-program, SMS, telephone).

[0090] Battery Health Archive: A complete lifecycle health record is established for each vehicle, recording its historical charging behavior and health change trends, which can be queried by users.

[0091] The solutions provided by the embodiments of this application have been described in detail above. The various embodiments in the specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make several improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of the claims of this application.

[0092] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0093] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly by hardware, a software module executed by a processor, or a combination of both. The software module can be located in random access memory (RAM), main memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art.

Claims

1. A method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile, characterized in that, Includes the following steps: S1: Collect electrical parameter data on the AC side during the charging process of the two-wheeled vehicle through the charging equipment, and bind the electrical parameter data with the unique identifier of the two-wheeled vehicle; S2: Process the bound electrical parameter data and extract the charging curve morphology features that characterize the battery health status; S3: Based on the charging curve morphology characteristics, establish an individualized longitudinal dynamic baseline for the two-wheeled vehicle, and combine it with the lateral reference baseline of the same type of battery to achieve charging anomaly detection through standardized deviation calculation; S4: Generate and output the corresponding early warning information based on the anomaly detection results.

2. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 1, characterized in that, In step S1, the electrical parameter data includes voltage, current, power, power, and charging time, and is acquired through a separate metering acquisition mode. The unique identifier is strongly bound to the charging session through NFC card swiping or QR code scanning charging process.

3. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 1, characterized in that, Step S1 also includes obtaining the core battery parameters of the two-wheeled vehicle. The core battery parameters are stored in association with the unique identifier. The core battery parameters include battery voltage, capacity and type information.

4. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 1, characterized in that, In step S2, the charging curve morphology features include the duration of the constant current phase and the slope of the current decrease curve in the constant voltage phase. The duration of the constant current phase is calculated by combining the sliding window variance detection method with the state machine to identify the constant current phase. The slope of the current decrease curve during the constant voltage stage is obtained by fitting the current value during the constant voltage stage using the least squares method after performing a logarithmic transformation.

5. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 4, characterized in that, The implementation process of the sliding window variance detection method includes: performing a moving average filter on the acquired current time series; constructing a sliding window with a preset number of data points and calculating the variance of the power data in each sliding window; setting a variance threshold, and determining the transition from the constant current stage to the constant voltage stage when the window variance exceeds the variance threshold. The start time of the constant current stage is the time when the current first reaches a first preset proportion of the current rating, and the end time of the constant current stage is the time when the variance threshold is exceeded.

6. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 4, characterized in that, The calculation process of the current drop curve slope in the constant voltage stage includes: taking the natural logarithm of the current value collected in the constant voltage stage to obtain the transformed data point (Xi, Yi), where Xi is the time t after the start of the constant voltage stage, and Yi is the natural logarithm of the current at the corresponding time ln(I); fitting the straight line ln(I)=a+b*t by the least squares method, where a is the intercept, b is the slope, and b is the slope of the current drop curve in the constant voltage stage. The start time of the constant voltage stage is the end time of the constant current stage, and the end time of the constant voltage stage is when the current drops to the second preset proportion of the current rating for the first time during the constant voltage stage.

7. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 1, characterized in that, In step S3, the process of establishing the individualized longitudinal dynamic baseline includes: setting an initial learning period, during which charging data is collected a preset number of times; based on the morphological characteristics of the historical charging curves during the learning period, calculating the mean slope μ and standard deviation σ of the current drop curve in the constant voltage stage, and forming the individualized longitudinal dynamic baseline.

8. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 7, characterized in that, In step S3, the standardized deviation calculation includes longitudinal deviation Z1 and lateral deviation Z2. The formula for calculating the longitudinal deviation Z1 is Z1=(μ0-μ) / σ, where μ0 is the slope of the current decrease curve during the constant voltage stage of the current charging. The lateral reference baseline is the typical characteristic value range of the corresponding battery model in the knowledge base, including the typical mean μ_ref and the typical standard deviation σ_ref. The formula for calculating the lateral deviation Z2 is Z2=(μ0-μ_ref) / σ_ref. When |Z1|>M1 and |Z2|>M2, it is determined to be a charging abnormality. M1 is the longitudinal deviation ratio parameter and M2 is the lateral deviation ratio parameter.

9. The method for early warning assessment of two-wheeled vehicle charging based on the AC side of a charging pile according to claim 8, characterized in that, In step S4, the type of warning information is determined by the duration of the constant current phase, the degree of deviation between |Z1| and M1, and the degree of deviation between |Z2| and M2. The types of warning information include yellow warning, orange warning, and red warning. The warning information is pushed to users and maintenance personnel through preset communication channels, including mini-programs, SMS, and telephone.

10. A two-wheeled vehicle charging early warning and assessment system based on the AC side of a charging pile, characterized in that, include: Data acquisition and binding module: used to collect electrical parameter data on the AC side during the charging process of the two-wheeled vehicle through the charging equipment, and bind the electrical parameter data with the unique identifier of the two-wheeled vehicle; Feature extraction module: used to process the bound electrical parameter data and extract the charging curve morphology features that characterize the battery health status; Baseline establishment and anomaly detection module: Based on the charging curve morphology characteristics, it establishes an individualized longitudinal dynamic baseline for the two-wheeled vehicle and combines it with the lateral reference baseline of the same type of battery to detect charging anomalies through standardized deviation calculation. Early warning output module: Used to generate and output corresponding early warning information based on the anomaly detection results.

Images