Vehicle battery health state evaluation method and system based on charging pile

Abstract

The invention provides a vehicle battery health state assessment method and system based on a charging pile, and the method comprises the steps: collecting the electrical parameters in the connection process of the charging pile and a vehicle, marking a key event, carrying out the decomposition processing of a charging current waveform, obtaining a basic charging current component and a pulsation component caused by an equalization circuit, generating a basic charging current curve and a cell consistency index; current response characteristics in different charge state intervals are calculated based on the basic charging current curve, correction is carried out in combination with the power distribution state of the charging pile, and current response parameters of the battery are obtained; and combining the current response parameter and the cell consistency index with a historical charging record to form a time sequence feature, and performing comparative analysis on the time sequence feature and a preset reference mode to obtain a current health state index of the battery. According to the invention, the health state of the battery can be accurately evaluated in a multi-vehicle parallel charging environment, the consistency problem of the battery cells can be identified, and the evaluation service of the health state of the battery is provided for the electric vehicle.

Description

Technical Field

[0001] The present invention relates to the technical field of battery evaluation, and in particular to a vehicle battery health status evaluation method and system based on a charging pile. Background Art

[0002] With the rapid adoption of electric vehicles, assessing the vehicle's battery state of health (SOH) has become increasingly important, directly impacting the vehicle's range, safety, and affordability. Currently, vehicle battery SOH assessment primarily relies on internal monitoring within the vehicle's battery management system (BMS) or offline testing using specialized equipment. However, these methods either rely on internal vehicle data and lack comprehensive historical usage data, or require the vehicle to be sent to a specialized testing facility for evaluation, increasing user costs.

[0003] Existing charging pile-based battery health assessment methods are beginning to attract attention, but they face numerous technical challenges. In particular, in environments with multiple vehicles charging in parallel, issues such as uneven power distribution caused by shared charging pile resources, interference from onboard balancing circuits on external measurement data, and the complex relationship between state-of-charge (SOC) changes and current response during charging make accurate battery health assessment difficult. These issues lead to existing assessment methods suffering from insufficient accuracy and limited applicability in practical applications. Summary of the Invention

[0004] The main purpose of this invention is to solve the technical problem that the existing battery health status assessment method based on charging piles is insufficient in assessment accuracy due to interference from the working of the equalization circuit and uneven power distribution in a multi-vehicle parallel charging environment; The present invention provides a vehicle battery health status assessment method based on a charging pile, the vehicle battery health status assessment method based on a charging pile comprising: Collect electrical parameters during the connection between the charging pile and the vehicle, and mark key events in the charging process to obtain a structured data stream; Decomposing the charging current waveform according to the structured data stream, and generating a basic charging current curve and a cell consistency index based on the decomposed basic charging current component and the pulsating component caused by the balancing circuit; Calculate the current response characteristics in different state-of-charge intervals based on the basic charging current curve, and perform corrections based on the power distribution state of the charging pile to obtain the current response parameters of the battery; According to the current response parameters and cell consistency indicators, combined with historical charging records and preset reference modes, a comparison analysis is performed to obtain the current health status index of the battery.

[0005] The present invention also provides a vehicle battery health status assessment system based on a charging pile, the vehicle battery health status assessment system based on a charging pile comprising: The data acquisition module is used to collect electrical parameters during the connection between the charging pile and the vehicle, and mark key events in the charging process to obtain a structured data stream; A waveform decomposition module is used to decompose the charging current waveform according to the structured data stream, and generate a basic charging current curve and a cell consistency index based on the decomposed basic charging current component and the pulsating component caused by the balancing circuit; A response characteristic module is used to calculate the current response characteristics in different state of charge intervals based on the basic charging current curve, and perform corrections in combination with the power distribution state of the charging pile to obtain the current response parameters of the battery; The health assessment module is used to form a time series feature based on the current response parameter and the cell consistency index in combination with historical charging records, and compare and analyze the time series feature with a preset reference pattern to obtain the current health status index of the battery.

[0006] The above-mentioned vehicle battery health status assessment method and system based on charging piles collects electrical parameters during the connection process between the charging pile and the vehicle and marks key events, decomposes and processes the charging current waveform, obtains the basic charging current component and the pulsating component caused by the balancing circuit, and generates a basic charging current curve and a cell consistency index; calculates the current response characteristics in different state of charge intervals based on the basic charging current curve, and corrects them in combination with the power distribution state of the charging pile to obtain the current response parameters of the battery; combines the current response parameters and the cell consistency index with the historical charging records to form a time series feature, and obtains the current battery health status index by comparing and analyzing with a preset reference pattern. The present invention can accurately assess the battery health status in a multi-vehicle parallel charging environment, and can identify cell consistency problems, providing battery health status assessment services for electric vehicles.

[0007] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or understood by practicing the present invention. The purposes and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0008] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 Schematic diagram of a first embodiment of a vehicle battery health status assessment method based on a charging pile in an embodiment of the present invention; Figure 2 Schematic diagram of a second embodiment of a vehicle battery health status assessment method based on a charging pile in an embodiment of the present invention; Figure 3 Schematic diagram of an embodiment of a vehicle battery health status assessment system based on a charging pile in an embodiment of the present invention. DETAILED DESCRIPTION

[0010] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0011] The terms "including," "having," and any variations thereof, as used in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device comprising a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to the process, method, product, or device.

[0012] To facilitate understanding of this embodiment, a vehicle battery health status assessment method based on a charging pile disclosed in an embodiment of the present invention is first described in detail. Figure 1 As shown, this method includes the following steps: 101. Collect electrical parameters during the connection between the charging pile and the vehicle, and mark key events in the charging process to obtain a structured data stream; In this embodiment, the data acquisition system uses a multi-sensor array integrated into the charging pile to monitor key electrical parameters of the charging process in real time. This sensor array includes a high-precision current sensor, a voltage sampling module, and a temperature monitoring unit, which are responsible for collecting core data such as charging current, port voltage, connection impedance, and ambient temperature.

[0013] The sensor network uses a layered sampling strategy for data acquisition. Under normal charging conditions, the system continuously collects various electrical parameters at a base frequency of 1Hz, ensuring a complete record of the overall charging process trends. When specific charging events are detected, the sampling frequency automatically increases to a high-frequency mode of 20-100Hz to capture critical transient characteristic changes.

[0014] Key event identification is achieved through charging state machine monitoring. The system continuously tracks changes in the battery's state of charge (SOC). When the SOC passes through preset critical points (such as 30%, 50%, and 80% charge stage demarcation points), high-frequency sampling is automatically triggered. Charging mode transitions are also important monitoring targets, including the transition from constant current to constant voltage and the activation of the onboard balancing circuit.

[0015] The data tagging mechanism provides a time base and event context for subsequent analysis. Each sampled data point is assigned a precise timestamp and corresponding event label, such as "CC to CV mode," "Balance start," and "SOC_80% reached." These tags not only identify the moment the event occurred but also include the charging environment parameters at the time.

[0016] To improve data processing efficiency, the system organizes collected raw data and event markers according to a unified data model, generating a structured data stream output. This data stream is organized as a time series, with each time segment containing a complete electrical parameter snapshot and corresponding status markers, facilitating subsequent signal processing and feature extraction.

[0017] Through this event-driven adaptive sampling strategy, the system effectively controls data storage requirements while ensuring the complete capture of key information. It is particularly suitable for charging station application scenarios that require long-term operation monitoring.

[0018] 102. Decompose the charging current waveform according to the structured data stream, and generate a basic charging current curve and cell consistency index based on the separated basic charging current component and the pulsating component caused by the balancing circuit; In this embodiment, equalization status markers are read from the structured data stream to identify key time points such as "equalization start" and "equalization end" to determine the complete operating cycle of the equalization circuit. For intermittent operation, the system combines adjacent equalization segments with a time interval of less than 60 seconds into a continuous operating window to ensure that the complete equalization behavior characteristics are captured.

[0019] Within the defined balancing time window, the system performs wavelet transform analysis on the charging current data, employing a 5-level decomposition using the db4 wavelet basis function. The system then extracts the energy distribution of the high-frequency components corresponding to wavelet coefficients at levels 2-4. The system then groups the data by SOC interval, analyzing the energy concentration within each frequency band and identifying the dominant operating frequency band for the balancing circuit. By reconstructing the wavelet coefficients of the dominant frequency band, the system extracts key characteristic parameters of the balancing waveform, including frequency calculated from the time interval between adjacent peaks, amplitude derived from the difference between peaks and valleys, and duration determined by the duration of the continuous pulsation period.

[0020] Based on the extracted waveform features, the system establishes a mapping relationship between SOC and equalization characteristics. By observing how characteristic parameters change with SOC, the system uses different fitting methods: a linear function is used to describe the frequency characteristic's decreasing trend with increasing SOC; a parabolic function is used to fit the amplitude characteristic, which reaches its peak at a moderate SOC; and a linear function is used to describe the duration's relatively gradual change. These fitting functions form a set of equalization circuit parameters, enabling the prediction of the equalization circuit's operating characteristics based on the current SOC.

[0021] The signal decomposition process utilizes the empirical mode decomposition algorithm, using the equalization circuit parameters as the basis for separation. The system identifies components belonging to the equalization pulsation by calculating the dominant frequency of each mode within the multiple intrinsic mode functions generated by the standard EMD algorithm and matching them with the expected frequencies in the equalization circuit parameters. Modes with matching frequencies and similar waveform characteristics are classified as the equalization pulsation components, while the remaining high-frequency components are classified as ambient noise. The low-frequency smooth components constitute the fundamental charging current components.

[0022] The system performs a detailed spectral characteristic analysis on the separated pulsation components. Short-time Fourier transform is used to obtain the time-frequency energy distribution, extract the maximum value, average value and standard deviation of the amplitude change, and calculate the coefficient of variation to quantify the degree of amplitude fluctuation. The duration feature is calculated by Hilbert transform to calculate the instantaneous amplitude envelope, set a dynamic threshold to identify the boundary of the pulsation event, count the interquartile range and median of the duration, and calculate the discrete index. The battery cell consistency index is obtained by weighted fusion of the amplitude variation coefficient and the duration discrete index, and the weight is adjusted according to the battery type. The basic charging current component is reconstructed into a pure charging current curve through wavelet threshold denoising.

[0023] 103. Calculate the current response characteristics in different state-of-charge intervals based on the basic charging current curve, and perform corrections based on the charging pile power distribution state to obtain the battery current response parameters; In this embodiment, the current response characteristics in different state-of-charge intervals are calculated based on the basic charging current curve, and corrected in combination with the power distribution state of the charging pile to obtain the current response parameters of the battery, including: analyzing the basic charging current curve to identify the transition point from the constant current stage to the constant voltage stage during the charging process; dividing the charging process into a constant current charging stage and a constant voltage charging stage based on the transition point, and dividing each stage into a plurality of state-of-charge intervals according to a predetermined state-of-charge interval; for each state-of-charge interval, calculating the ratio of the current change rate to the state-of-charge change rate, and using the ratio as the current response characteristic value of the corresponding state-of-charge interval. The current response characteristic values ​​of all charge state intervals are combined in the order of charge state to obtain the initial current response parameters; the real-time power distribution data of the charging pile is collected, and whether the charging pile is in the power limitation mode in each charge state interval is determined based on the real-time power distribution data; for the charge state interval in the power limitation mode, the power correction coefficient is calculated based on the total load parameters and distribution strategy of the charging pile at that time, and the power correction coefficient is used to correct the initial current response parameters in the charge state interval of the power limitation mode; the corrected initial current response parameters of all charge state intervals are arranged in the order of charge state to obtain the current response parameters of the battery.

[0024] Specifically, the transition point identification uses a multi-layer detection algorithm. The system first uses a sliding window technique to calculate the local slope change of the basic charging current curve, with the window width set to 60 seconds. The linear regression slope of the current change is calculated within each window. When the slope of three consecutive windows changes from near zero to a continuously negative value, the system identifies the starting point of the transition as the transition point from constant current to constant voltage. To improve detection accuracy, the system also verifies the difference in current smoothness before and after the transition point, ensuring that the current fluctuation before the transition is less than 5% and that the current shows a clear downward trend after the transition.

[0025] Based on the identified transition points, the system divides the charging process into two main phases. The constant-current phase uses equal SOC intervals, typically creating multiple intervals at 5% intervals. The constant-voltage phase takes into account the nonlinearity of the charging rate and adopts a variable interval strategy: 2.5% intervals are used in the early stages when SOC changes more rapidly, and 10% intervals are used in the later stages. The system obtains SOC data from the vehicle's BMS communication or the Coulomb integration method to ensure the accuracy of the interval division.

[0026] The current response characteristic value is calculated for each state-of-charge interval. The system first preprocesses the raw data within the interval, including outlier detection and data smoothing. Outlier detection uses the Z-score method, setting a threshold of 3.0 to identify outliers. Data smoothing uses a Savitzky-Golay filter with a window size of 10% of the number of data points in the interval. After preprocessing, the system calculates the current change rate (dI / dt) and SOC change rate (dSOC / dt) for each sampling point in the interval, using the central difference method to improve calculation accuracy. When the SOC change is too small, the system automatically expands the calculation window until the minimum change requirement is met. Finally, the ratio of dI / dt to dSOC / dt is calculated as the current response characteristic value for the interval, and the median method is used to process the ratio of multiple sampling points to improve the robustness of the result.

[0027] The power allocation status is determined based on multiple real-time monitoring indicators. The system collects data such as the total output power of the charging station, the power allocated to each interface, and the number of connected vehicles, with a sampling interval of 10 seconds. Power limit determination is based on three levels of indicators: total load utilization exceeding 90%, the ratio of actual vehicle power to requested power below 85%, and the continuous fluctuation rate of power allocation exceeding 15%. If any two of these indicators are met simultaneously and for a duration of more than 60 seconds, the corresponding state of charge interval is marked as power-limited mode.

[0028] Furthermore, for the state of charge interval in the power limiting mode, the power correction coefficient is calculated based on the total load parameters and allocation strategy of the charging pile at that time, and the initial current response parameters in the state of charge interval in the power limiting mode are corrected using the power correction coefficient, including: collecting the total load parameters of the charging pile and the charging demand parameters of all currently connected vehicles, and calculating the theoretical current value that the vehicle should obtain under ideal non-power limiting conditions based on the total load parameters and charging demand parameters, combined with the power allocation strategy of the charging pile; calculating the ratio of the charging current actually obtained by the vehicle to the theoretical current value to obtain the power correction coefficient; for each sampling point in the state of charge interval in the power limiting mode, dividing the current response characteristic value at the corresponding moment by the power correction coefficient to obtain a corrected current response characteristic value; replacing the corresponding value in the initial current response parameter with the corrected current response characteristic value to obtain a corrected current response parameter.

[0029] Specifically, the calculation of the theoretical current value requires comprehensive consideration of the charging pile's allocation strategy and vehicle needs. The system first identifies the power allocation mode currently used by the charging pile, which mainly includes three modes: average allocation, priority allocation, and dynamic adjustment. For the average allocation mode, the system divides the total rated power of the charging pile by the number of currently connected vehicles to obtain the theoretical allocated power for each vehicle. For priority allocation, the system calculates power allocation based on the order of vehicle access or preset priority. For the dynamic adjustment mode, the system considers the SOC status and remaining charging time of each vehicle for weighted allocation. After the theoretical allocated power is determined, the system compares it with the maximum power requested by the vehicle's BMS and takes the smaller value as the vehicle's ideal power. Finally, based on the vehicle's current charging voltage, the theoretical current value is calculated by dividing the power by the voltage.

[0030] The power correction factor is calculated by comparing the actual current to the theoretical current. The system extracts the actual current value for the corresponding time period from the basic charging current curve and divides it by the calculated theoretical current value to obtain the original correction factor. To account for instantaneous fluctuations during the charging process, the system applies an exponentially weighted moving average to the original coefficient, with a smoothing factor set to 0.3 to reflect power changes while avoiding excessive fluctuations. If the correction factor is less than 0.1, the system deems the power limit too severe and replaces it with the coefficient value from the adjacent time period or a historical typical value to avoid overcorrection.

[0031] The current response eigenvalue correction process is performed point by point. For each sampling point within the power limit mode interval, the system searches for the power correction coefficient at the corresponding moment and uses linear interpolation to handle cases where time points are not completely aligned. The correction calculation uses a division operation, dividing the original eigenvalue by the correction coefficient to obtain the eigenvalue after the power limit effect is eliminated. When the original eigenvalue approaches zero, the system uses additive correction to avoid numerical instability. After the correction is completed, the system verifies the rationality of the corrected eigenvalue to ensure that its change trend conforms to the physical characteristics of the battery.

[0032] During the final integration of the current response parameters, the system re-enters the corrected eigenvalues ​​into the corresponding SOC intervals, while maintaining the original eigenvalues ​​in the non-power-limited intervals. To ensure the continuity of the parameter vector, the system employs a smooth transition at the boundary between the power-limited and non-power-limited intervals to avoid sudden changes in eigenvalues. The integrated current response parameters form a complete eigenvector that objectively reflects the battery's true response characteristics at each state of charge, providing accurate baseline data for subsequent health assessments.

[0033] 104. Based on the current response parameters and cell consistency indicators, combined with historical charging records, time series characteristics are formed, and the time series characteristics are compared and analyzed with the preset reference mode to obtain the current health status index of the battery.

[0034] In this embodiment, the current response parameters and cell consistency indicators are combined with historical charging records to form a time series feature, and the time series features are compared and analyzed with preset reference patterns to obtain the current health status index of the battery, including: forming a time series feature according to the current response parameters and cell consistency indicators in combination with historical charging records; standardizing the time series features through a temperature correction algorithm according to the ambient temperature of the charging process, and mapping data collected under different temperature conditions to a unified reference temperature condition; retrieving a reference pattern set corresponding to the battery model of the vehicle from a preset reference battery database, and using a weighted dynamic time warping algorithm to calculate the similarity score between the normalized time series features and each pattern in the reference pattern set; determining the most matching reference pattern according to the similarity score, and calculating the current health status index of the battery in combination with the battery usage time and the number of charge and discharge cycles.

[0035] Specifically, the process of constructing time series features begins with historical data retrieval. The system uses the vehicle's unique identifier (charging card ID or VIN code) to retrieve the target vehicle's charging records from the charging station database over the past six months, typically capturing data from 15-25 valid charging processes. During the retrieval process, the system screens for records with charging times exceeding 30 minutes and SOC increments greater than 20% to ensure data validity and integrity. For each historical record, the system extracts the corresponding current response parameters and battery cell consistency indicators, while also recording key environmental parameters during charging, including ambient temperature, charging power level, SOC status at the start of charging, and other key information.

[0036] The quality verification of historical data utilizes a multi-dimensional inspection mechanism. The system first checks the integrity of the current response parameter vector to ensure that the characteristic values ​​for each SOC interval exist and are reasonable. For records with missing values, the system interpolates the values ​​based on the previous and next records, using cubic spline interpolation to maintain data smoothness. Verification of cell consistency indicators focuses on the rationality of their numerical ranges and changing trends, eliminating clearly abnormal numerical points. Quality-verified historical data is combined with the latest analysis results in chronological order to form a complete sequence containing time-evolution information.

[0037] Time series features are organized using a multi-layered structure. The first layer is the time dimension, which arranges charging records in chronological order. The second layer is the feature type, which includes two categories: current response parameters and cell consistency indicators. The third layer is the specific numerical data. The current response parameters are stored as vectors of characteristic values ​​for each SOC interval, and the cell consistency indicators are recorded as scalars. To facilitate subsequent processing, the system adds metadata tags to each time point, including contextual information such as charging date, ambient temperature, and charging mode.

[0038] The implementation of the temperature correction algorithm is based on the temperature sensitivity of the battery's electrochemical characteristics. The system selects 25°C as the standard reference temperature and performs temperature compensation on all data collected under non-25°C conditions. The construction of the temperature correction function takes into account the impact of different temperatures on the battery's internal resistance, ionic conductivity, and electrochemical reaction rate. For the current response parameters, the system adopts a segmented compensation strategy: in the low temperature range (5-15°C), the compensation coefficient is large, reflecting the significant inhibition of low temperature on charge acceptance; in the medium temperature range (15-35°C), the compensation coefficient is close to 1, indicating that the temperature effect is relatively small; in the high temperature range (35-45°C), the compensation coefficient takes into account the negative impact of high temperature on battery performance.

[0039] Temperature correction of battery cell consistency indicators is more complex because temperature changes can affect the degree of performance differences between battery cells. The system establishes a temperature-consistency correction model, which is built based on a large amount of experimental data and describes the influence of temperature changes on the amplification or reduction of battery cell differences. Under low temperature conditions, the performance differences between battery cells are usually amplified, so the consistency indicators need to be adjusted downward; under high temperature conditions, the performance of battery cells tends to be homogenized, and the consistency indicators need to be adjusted upward. The time series characteristics after temperature correction eliminate the impact of ambient temperature changes on data comparability, ensuring a consistent benchmark for data collected at different times.

[0040] The reference battery database retrieval process is based on a precise match between the vehicle's battery model and chemistry. The system first identifies the battery's detailed specifications, including capacity, chemistry (e.g., lithium iron phosphate, ternary lithium), manufacturer, and production batch, through the vehicle's BMS communication protocol or the vehicle model information recorded by the charging station. Based on this information, the system queries the corresponding battery model library from a pre-built reference database. The reference database contains a large number of standard characteristic patterns for batteries in different health states. Each battery model typically includes 10-15 reference patterns at different health levels, with health levels decreasing from 100% to 60%, with each level representing 5% health.

[0041] The construction of the reference pattern set is based on statistical methods. For each health level, the database stores characteristic data from multiple actual battery samples. Statistical analysis is then used to identify the typical characteristic patterns for that health level. These pattern data include standard current response parameter vectors and expected values ​​for cell consistency metrics. The range and standard deviation of these characteristics are also recorded, providing an uncertainty measure for similarity calculations.

[0042] The implementation of the weighted dynamic time warping algorithm takes into account the differing importance of current response parameters and cell consistency metrics. The algorithm first calculates the DTW distance for each of the two feature types. For the current response parameters, the DTW algorithm processes a multidimensional vector sequence, constructing a cost matrix by calculating the Euclidean distance between the eigenvalues ​​of each SOC interval, and then searching for the optimal alignment path. This algorithm can handle nonlinear time stretching in time series and adapt to differences in the rate of SOC change during different charging processes.

[0043] For cell consistency metrics, DTW processes scalar time series and constructs a cost matrix by calculating the absolute difference between metric values ​​at corresponding time points. Given the relatively slow changes in consistency metrics, a larger time window constraint is set in the algorithm to allow for greater time alignment deviation.

[0044] The weighting is based on the degree of influence of the two characteristics on the battery health status. Generally, the weight of the current response parameter is set to 0.7, and the weight of the cell consistency index is set to 0.3, but this can be adjusted according to the battery type and usage mode. For lithium iron phosphate batteries, since their cell consistency is relatively stable, the weight of the current response parameter can be increased to 0.8. For ternary lithium batteries, the changes in cell consistency are more sensitive, and the two weights can be set to 0.6 and 0.4.

[0045] The similarity score calculation converts the DTW distance into a similarity score on a scale of 0-100. The conversion formula uses an exponential decay function to ensure that the smaller the distance, the higher the similarity. The system calculates the similarity score between the target time series and each pattern in the reference pattern set, forming a score vector.

[0046] The most matching reference pattern is determined using a multi-candidate fusion strategy. The system selects the top three reference patterns with the highest similarity scores and calculates their weighted average healthiness as a preliminary estimate. The weights are proportional to the similarity scores, ensuring that the most similar pattern has the greatest influence. This multi-pattern fusion approach reduces the bias of single-pattern matching and improves the stability and reliability of the evaluation results.

[0047] Battery history calibration takes into account the impact of time on battery aging. The system collects the battery's cumulative usage time (calculated from initial activation) and the number of charge and discharge cycles (recorded by the BMS or through charging history statistics). Based on battery aging theory, the system establishes a time calibration model that describes the theoretical decay of battery health over time and cycle count. The calibration model compares the initial health estimate with the theoretical decay curve, calculates the degree of deviation, and adjusts the final health index accordingly.

[0048] The health index output ranges from 0 to 100, with higher values ​​indicating better battery health. The system also outputs a confidence score, calculated based on the similarity distribution during the matching process and the completeness of historical data. When the confidence score is low, the system recommends additional observation data or more detailed testing and analysis.

[0049] In this embodiment, by collecting electrical parameters during the connection process between the charging pile and the vehicle and marking key events, the charging current waveform is decomposed and processed to obtain the basic charging current component and the pulsating component caused by the balancing circuit, and a basic charging current curve and a cell consistency index are generated; based on the basic charging current curve, the current response characteristics in different state of charge intervals are calculated, and correction is performed in combination with the power distribution state of the charging pile to obtain the current response parameters of the battery; the current response parameters and the cell consistency index are combined with historical charging records to form a time series feature, and the current health status index of the battery is obtained by comparison and analysis with a preset reference mode. The present invention can accurately assess the battery health status in a multi-vehicle parallel charging environment, and can identify cell consistency problems, providing battery health status assessment services for electric vehicles.

[0050] See also Figure 2 Another embodiment of the vehicle battery health status assessment method based on the charging pile in the embodiment of the present application includes: 201. Collect electrical parameters during the connection between the charging pile and the vehicle, and mark key events in the charging process to obtain a structured data stream; In this embodiment, step 201 is similar to step 101 in the first embodiment and will not be described again.

[0051] 202. Determine a time window for operation of the equalization circuit according to the equalization state mark in the structured data stream, extract waveform features of the equalization circuit according to the time window, and generate equalization circuit parameters; In this embodiment, the method of determining the time window for the operation of the balancing circuit according to the balancing state mark in the structured data stream, and extracting the waveform characteristics of the balancing circuit according to the time window to generate the balancing circuit parameters includes: identifying the starting time point and the ending time point of the operation of the balancing circuit according to the balancing state mark in the structured data stream, and determining the time window for the operation of the balancing circuit; within the time window, performing a wavelet transform on the charging current waveform measured by the charging pile to extract the energy distribution characteristics of the high-frequency component; identifying the waveform characteristics caused by the balancing circuit under different charge states according to the energy distribution characteristics, including the frequency, amplitude and duration of the waveform; performing a correlation analysis on the characteristic waveform and the corresponding charge state, establishing a mapping relationship between the charge state and the characteristic waveform parameters of the balancing circuit, and using the mapping relationship to construct the balancing circuit parameters.

[0052] Specifically, the identification process for balancing status markers uses state machine logic to process event sequences in structured data streams. The system scans the marker fields in the data stream and identifies timestamp information for marker types such as "balancing start," "balancing in progress," and "balancing end." For the "balancing start" marker, the system records its timestamp as the start time of the balancing operation; the corresponding "balancing end" marker timestamp is used as the end time of the operation. In actual applications, the BMS of some vehicle models may use an intermittent balancing strategy, resulting in multiple start and end markers within a short period of time.

[0053] To handle intermittent balancing modes, the system implements a time continuity algorithm. When the time interval between consecutive "Balance End" and "Balance Start" markers is less than a preset threshold (typically 60 seconds), the system merges these intermittent time periods into a single continuous balancing window. During this merging process, the system retains the timestamp of the first "Balance Start" marker as the overall start point and the timestamp of the last "Balance End" marker as the overall end point.

[0054] In cases where the marker is missing or abnormal, the system uses current waveform characteristics to assist in judgment. By analyzing the high-frequency components of the charging current, the system can identify the characteristic current fluctuations generated by the balancing circuit. If a significant periodic pulsation in the 2-10Hz frequency band is detected and lasts for more than 30 seconds, the system marks this period as a suspected balancing window and verifies its validity through subsequent analysis.

[0055] Once the time windows are determined, the system constructs a balancing time mapping table, recording each window's start and end times, duration, and corresponding state-of-charge range. This mapping table provides a time index for subsequent waveform feature extraction, ensuring the accuracy and completeness of the analysis process.

[0056] Wavelet transform processing uses a multi-resolution analysis method to extract the frequency domain characteristics of the equalizing circuit. The system uses the db4 wavelet from the Daubechies wavelet family as the basis function. This wavelet has excellent time-frequency localization properties and is suitable for analyzing non-stationary equalizing current signals. The wavelet transform uses a five-level decomposition to decompose the original signal into wavelet coefficients at different frequency scales.

[0057] During the wavelet decomposition process, the system focuses on wavelet coefficients at levels 2-4. The frequency range (approximately 1-16 Hz) corresponding to these coefficients perfectly covers the operating frequency band of typical onboard balancing circuits. For each decomposition level, the system calculates the energy density of the corresponding wavelet coefficients, converts them to the frequency domain through a Fourier transform, and analyzes the energy distribution characteristics of each frequency band.

[0058] Extracting energy distribution features involves calculating parameters across multiple dimensions. First, the frequency band energy contribution. The system calculates the ratio of each frequency band's energy to the total energy, identifying the primary frequency bands where energy is concentrated. Second, the variance and skewness of the energy distribution reflect the concentration and asymmetry of the energy distribution. Third, time-varying characteristic analysis uses a sliding window to calculate the time-varying pattern of the energy distribution, capturing the dynamic characteristics of the equalizer circuit's operating state.

[0059] The waveform feature recognition process uses the results of energy distribution analysis to perform in-depth parameter extraction. The system first groups the balancing operating time windows by current SOC value, typically dividing the 0-100% SOC range into 10-20 intervals, each covering a 5-10% SOC range. Within each SOC interval, the system selects the two to three frequency bands with the highest energy contribution for focused analysis.

[0060] Frequency signatures are extracted through wavelet coefficient reconstruction. The system selects the wavelet coefficients corresponding to the target frequency band and reconstructs the time-domain signal for that band through an inverse wavelet transform. The reconstructed signal reflects the operating characteristics of the equalizer circuit at a specific frequency. By analyzing the autocorrelation function of the reconstructed signal, the system calculates the dominant period of the signal and takes its inverse to determine the operating frequency of the equalizer circuit. In cases where multiple significant periods exist, the system records the dominant two to three frequency components and their relative strengths.

[0061] The calculation of amplitude characteristics is based on envelope analysis of the reconstructed signal. The system uses the Hilbert transform method to calculate the instantaneous amplitude of the reconstructed signal, forming an envelope curve. Several amplitude parameters are extracted from the envelope curve: the maximum amplitude reflects the maximum modulation intensity of the equalizer circuit; the average amplitude indicates typical operating intensity; the amplitude standard deviation reflects the stability of amplitude variations; and the peak-to-peak value indicates the amplitude variation range. Together, these parameters describe the amplitude characteristics of the equalizer circuit under specific SOC conditions.

[0062] Duration features are extracted through pulsation event detection. The system sets a dynamic threshold on the envelope curve, typically 70-80% of the envelope mean. A continuous period exceeding the threshold is identified as a complete, balanced pulsation event. For each detected event, the system records its duration and calculates the duration distribution of all events within the entire SOC interval. Distribution parameters include mean duration, median duration, standard deviation of duration, and interquartile range.

[0063] The correlation analysis process establishes a quantitative relationship between SOC and waveform characteristic parameters. The system collects frequency, amplitude, and duration characteristic data for each SOC interval and uses regression analysis to fit the functional relationship between these parameters and SOC changes. Regarding the frequency characteristics, the system observes that the equilibrium frequency of most vehicle models is higher at low SOC and gradually decreases as SOC increases. Therefore, a negative exponential function or a linear decreasing function is used for fitting.

[0064] The SOC correlation of the amplitude feature exhibits a more complex pattern. The system found that the equilibrium intensity peaks in the mid-SOC range (typically 40-70%) and decreases at both ends. This characteristic is best described by a quadratic or Gaussian function. The least squares method is used to fit the function parameters by minimizing the squared error between the predicted and observed values.

[0065] The relationship between duration and SOC is relatively flat, often exhibiting a slow-changing trend. The system uses either a linear function or a constant function for fitting, with the choice based on a goodness-of-fit comparison. For cases where the fit is poor, the system uses a piecewise function approach, dividing the SOC range into two or three subranges and fitting a functional relationship within each subrange.

[0066] The mapping relationship is validated using cross-validation. The system divides the observed data into a training set and a validation set. The training set is used to fit the mapping function, and the validation set is used to evaluate the fitting accuracy. Validation metrics include root mean square error, correlation coefficient, and coefficient of determination. If the validation accuracy does not meet the requirements, the system adjusts the fitting function type or increases the function complexity.

[0067] The construction of balancing circuit parameters integrates all mapping relationships into a unified parameter model. This parameter model is organized in a hierarchical structure: the top layer identifies the battery type, the middle layer indexes the SOC range, and the bottom layer contains specific characteristic parameters and function coefficients. The model also includes metadata, such as the temperature range of data collection, charging power level, and sample size, to provide a basis for evaluating parameter applicability.

[0068] The parameter model has adaptive updating capabilities. When the system obtains new observation data, it can automatically adjust the parameters of the mapping function to improve the accuracy and adaptability of the model. The update process uses recursive least squares, while maintaining the weight of historical data, appropriately increasing the influence of new data to achieve online optimization of the model.

[0069] 203. Using the balancing circuit parameters as constraints, decompose the charging current waveform measured by the charging pile using an empirical mode decomposition algorithm to obtain a basic charging current component, a pulsating component caused by the balancing circuit, and an environmental noise component; In this embodiment, the spectral analysis of the pulsating component, the calculation of the variation characteristics of the amplitude and duration of the pulsating component, and the calculation of the cell consistency index characterizing the consistency of the battery cells in the battery pack based on the variation characteristics include: performing a fast Fourier transform on the pulsating component to obtain the spectral distribution of the pulsating component, and identifying the characteristic frequency interval of the balancing circuit in the spectral distribution; calculating the energy density in the characteristic frequency interval, and analyzing the trend of the energy density changing with the state of charge to obtain frequency domain characteristic parameters; performing envelope analysis on the pulsating component in the time domain to extract the amplitude variation curve and duration distribution of the pulsating waveform; calculating the amplitude variation coefficient based on the degree of fluctuation of the amplitude variation curve, and calculating the duration dispersion index based on the concentration of the duration distribution; combining the frequency domain characteristic parameters, the amplitude variation coefficient and the duration dispersion index according to preset weights to obtain the cell consistency index, wherein the weights are adjusted according to the vehicle battery type and charging mode.

[0070] Specifically, the constrained empirical mode decomposition (EMD) process converts equalizer circuit parameters into guiding conditions for the EMD algorithm. The standard EMD algorithm iteratively decomposes a composite signal into multiple intrinsic mode functions (IMFs), each representing an oscillation mode within a specific frequency range. To improve decomposition accuracy, the system incorporates the frequency information in the equalizer circuit parameters as constraints for modal identification.

[0071] The EMD decomposition process first preprocesses the raw charging current waveform, including removing the DC component and low-frequency trend terms. The system then performs a standard EMD iteration process: identifying all local extreme points of the signal, constructing the upper and lower envelopes using cubic spline interpolation, calculating the mean of these two envelopes, and subtracting the mean from the original signal to obtain a candidate IMF. This process is repeated until the candidate IMF meets the definition of an IMF: the number of local extreme points is equal to or less than one difference from the number of zero crossings, and the mean of the upper and lower envelopes is close to zero.

[0072] After obtaining all IMFs from the EMD decomposition, the system performs modal classification using the equalizer circuit parameters. For each IMF, the system calculates its dominant frequency and obtains its spectral characteristics through Fourier or Hilbert transforms. The dominant frequency of the IMF is then compared with the expected frequency under the corresponding SOC state, as specified in the equalizer circuit parameters. The matching criteria include frequency overlap and energy similarity.

[0073] Frequency overlap is determined by calculating the ratio of the intersection of the IMF frequency range and the expected frequency range of the equalization circuit. When the overlap exceeds 70%, the IMF is considered to contain equalization pulsation components. Energy similarity is calculated by comparing the power spectral density distribution of the IMF with the energy distribution pattern predicted by the equalization circuit parameters. The system uses the Pearson correlation coefficient to quantify the similarity between the two distributions. IMFs with correlation coefficients greater than 0.6 are preliminarily classified as equalization-related components.

[0074] For IMFs that meet both frequency overlap and energy similarity requirements, the system further verifies the consistency of their time-domain characteristics. The system calculates the IMF's envelope characteristics, including parameters such as average amplitude, amplitude variation range, and pulsation duration, and compares them with the corresponding values ​​in the equalization circuit parameters. If the difference in time-domain characteristics is less than 20%, the IMF is confirmed to be a balanced pulsating component.

[0075] During the component classification process, the system linearly superimposes all identified equalization-related IMFs to form a comprehensive equalization pulsation component. The remaining IMFs are further classified based on their frequency characteristics: high-frequency random components (frequencies greater than 20 Hz and lacking significant periodicity) are classified as ambient noise components; low-frequency smooth components (frequencies less than 0.1 Hz) are classified as part of the basic charging current component. The system also retains the residual term from the EMD decomposition, which typically contains slowly varying trends in the signal, and combines it with the basic charging current component.

[0076] Fast Fourier transform analysis extracts detailed frequency-domain features from the separated pulsating components. The system employs a Hanning window function to reduce spectral leakage. The FFT analysis uses a 0.1Hz resolution and a 0-50Hz frequency range to capture all possible operating frequencies of the equalizer circuit.

[0077] The identification of characteristic frequency intervals is based on power spectral density analysis. The system calculates the power spectral density function of the pulsating component and identifies frequency intervals where the power density is significantly higher than the noise floor. This identification process uses a threshold detection method, with the threshold set at three times the mean noise power. Frequency ranges that continuously exceed the threshold are marked as characteristic frequency intervals. In the case of multiple discrete intervals, the system records the boundary frequencies and center frequencies of all intervals.

[0078] Energy density is calculated by integrating the power spectral density within a characteristic frequency interval. The system calculates the total energy and average energy density for each characteristic frequency interval, while also recording the peak position and half-power bandwidth of the power distribution within that interval. These parameters reflect the energy distribution characteristics of the equalization circuit at different frequencies.

[0079] State-of-charge trend analysis is performed using time-frequency analysis. The system segments the pulsating component into SOC intervals and calculates the energy density parameters corresponding to each SOC interval. By observing the energy density variation pattern with SOC, the system extracts trend characteristics, including the rate of change, direction of change, and stability of the energy density. Trend analysis uses linear regression and nonparametric trend testing to quantify the significance and direction of energy density changes.

[0080] The comprehensive frequency domain characteristic parameters include indicators from multiple dimensions. The main frequency stability is quantified by calculating the coefficient of variation of the main frequency across various SOC intervals; the bandwidth is calculated by averaging the width of the characteristic frequency intervals; energy concentration is assessed by calculating the energy percentage within a ±1Hz range around the main frequency; and frequency drift is determined by fitting a trend line of the main frequency versus SOC and calculating the slope. Together, these parameters constitute the frequency domain characteristic parameter set, reflecting the frequency domain stability and regularity of the equalization circuit's operating characteristics.

[0081] Envelope analysis uses the Hilbert transform method to calculate the instantaneous amplitude of the pulsating component. The system first performs a Hilbert transform on the pulsating component to obtain an analytical signal. The modulus of the analytical signal is then calculated as a sequence of instantaneous amplitudes. To remove high-frequency noise from the envelope, the system applies a low-pass filter to the instantaneous amplitude sequence, with a cutoff frequency set to 1 / 5 of the main pulsating frequency.

[0082] The amplitude variation curve is extracted by performing peak detection on the filtered envelope sequence. The system uses an adaptive threshold peak detection algorithm, with the threshold set at the envelope mean plus 0.5 times the standard deviation. The detected peak points constitute the key nodes of the amplitude variation curve, and the system records the amplitude value and occurrence time of each peak.

[0083] Duration distribution statistics are based on the detection of pulsating event boundaries. The system sets start and end thresholds on the envelope curve, typically 30% and 20% of the peak value, respectively. From the peak point, the system searches forward for the first point below the start threshold as the event start boundary, and backward for the first point below the end threshold as the event end boundary. The duration of each event is the difference between the end and start times.

[0084] The coefficient of variation (CV) of amplitude is calculated based on all detected peak amplitudes. The system calculates the arithmetic mean and sample standard deviation of the peak amplitudes. The CV is the ratio of the standard deviation to the mean. This metric reflects the relative fluctuation in balancing strength. A larger value indicates greater balancing strength instability, indirectly reflecting significant performance differences between cells.

[0085] The duration dispersion index is calculated using the interquartile range method. The system sorts all duration data and calculates the first quartile, Q1, and the third quartile, Q3, with the interquartile range being Q3 - Q1. The duration dispersion index is defined as the ratio of the interquartile range to the median. This indicator is robust to outliers and can consistently reflect the degree of dispersion in the duration distribution.

[0086] The cell consistency index is calculated using a weighted linear combination method. The system first normalizes the frequency domain characteristic parameters, amplitude variation coefficient, and duration dispersion index to a standard range of 0-100. This normalization process uses a maximum and minimum value normalization method to ensure comparability of parameters of different dimensions.

[0087] Weight settings are configured differently based on battery type and charging mode. For lithium iron phosphate batteries, due to their relatively stable cell characteristics, the weight of the frequency domain characteristic parameters is set to 0.5, the amplitude coefficient of variation is 0.3, and the duration dispersion index is 0.2. For ternary lithium batteries, due to the more sensitive changes in cell consistency, the weight distribution is adjusted to 0.4, 0.4, and 0.2. In fast charging mode, due to the higher charging current, the weight of amplitude-related indicators will be appropriately increased.

[0088] The cell consistency index is calculated as the sum of the products of each normalized parameter and its corresponding weight. The index ranges from 0 to 100, with higher values ​​indicating greater cell-to-cell variability and a stronger need for balancing. The system also calculates the index's confidence interval and performs error propagation analysis based on the uncertainty of the input parameters, providing a basis for assessing the index's reliability.

[0089] 204. Performing spectrum analysis on the pulsation component, calculating variation characteristics of the amplitude and duration of the pulsation component, and calculating a cell consistency index representing the consistency of cells in the battery pack based on the variation characteristics; In this embodiment, the spectral analysis of the pulsation component uses short-time Fourier transform technology to convert the time-domain pulsation signal into a time-frequency domain representation. The system sets the analysis window length to 2 seconds and the overlap ratio to 50% to ensure accurate capture of the time-frequency characteristics of the pulsation signal. Through STFT analysis, the system obtains the spectral distribution of the pulsation component at different times, focusing on the energy variation pattern in the 2-10Hz frequency band.

[0090] The calculation of amplitude variation characteristics is based on envelope detection technology. The system applies a Hilbert transform to the pulsating component, calculating its instantaneous amplitude to form an envelope curve. A sliding window method is used on this envelope curve to identify local peaks, with the window size set to 1.5 times the estimated pulsation period. The system then calculates the amplitude distribution of all peaks and calculates the mean, standard deviation, and coefficient of variation of the peak amplitudes. The coefficient of variation reflects the relative degree of amplitude fluctuation.

[0091] Duration variation characteristics are captured by detecting the boundaries of pulsating events. The system sets dual thresholds on the envelope curve: an upper threshold of 80% of the envelope mean and a lower threshold of 40% of the envelope mean. The onset of a pulsating event is defined as the moment the envelope value rises from below the lower threshold and crosses the upper threshold, and the end of a pulsating event is defined as the moment the envelope value falls from the upper threshold and falls below the lower threshold. The system records the duration of each pulsating event and calculates its distribution characteristics, including mean duration, standard deviation, and interquartile range.

[0092] The calculation of the cell consistency index comprehensively considers the variability of amplitude and duration. The system normalizes the coefficient of variation of amplitude and the interquartile range of duration separately, then weights them together according to preset weights. For lithium iron phosphate batteries, the amplitude feature weight is set to 0.6, and the duration feature weight is set to 0.4; for ternary lithium batteries, both weights are set to 0.5. The cell consistency index value ranges from 0 to 100, with higher values ​​indicating greater differences between cells and a stronger need for balancing.

[0093] 205. Reconstruct and filter the basic charging current component to generate a basic charging current curve; In this embodiment, the reconstruction process of the basic charging current components first integrates the relevant modal components obtained through EMD decomposition. The system linearly superimposes the IMF components classified as the basic charging current with the residual terms to form a preliminary reconstructed signal. During the reconstruction process, the system verifies the phase consistency of each component to ensure that the superimposed signal retains its original time domain characteristics.

[0094] The filtering process uses a wavelet threshold denoising method to remove residual noise from the reconstructed signal. The system uses the sym8 wavelet as the basis function and performs a six-level wavelet decomposition of the reconstructed signal. At each decomposition level, the system calculates a noise estimate of the wavelet coefficients and shrinks them using a soft thresholding method. The threshold value is adaptively determined based on the Bayesian risk criterion, removing noise while preserving important signal features.

[0095] The filtered signal also needs to be corrected for boundary effects. Because wavelet transforms can produce false oscillations at signal boundaries, the system uses a symmetric extension method to address boundary issues. Data is extended by a wavelet filter length at each end of the signal, and after filtering, the original signal length is truncated.

[0096] During the generation of the basic charging current curve, the system also verifies the curve's physical plausibility. By checking the monotonicity and continuity of the charging current, the reconstructed curve ensures that it conforms to the physical laws of battery charging. During the constant-current charging phase, the current should remain relatively stable; during the constant-voltage charging phase, the current should exhibit a monotonically decreasing trend. If unreasonable fluctuations are detected, the system uses cubic spline interpolation for local smoothing, ultimately generating a basic charging current curve that accurately reflects the battery's intrinsic charging characteristics.

[0097] 206. Calculate the current response characteristics in different state-of-charge intervals based on the basic charging current curve, and perform corrections based on the power distribution state of the charging pile to obtain the current response parameters of the battery; 207. According to the current response parameter and the cell consistency index, a comparison analysis is performed in combination with historical charging records and a preset reference mode to obtain a current health status index of the battery.

[0098] In this embodiment, steps 206-207 are similar to steps 103-104 in the first embodiment and are not described again here.

[0099] In this embodiment, by collecting electrical parameters during the connection process between the charging pile and the vehicle and marking key events, the charging current waveform is decomposed and processed to obtain the basic charging current component and the pulsating component caused by the balancing circuit, and a basic charging current curve and a cell consistency index are generated; based on the basic charging current curve, the current response characteristics in different state of charge intervals are calculated, and correction is performed in combination with the power distribution state of the charging pile to obtain the current response parameters of the battery; the current response parameters and the cell consistency index are combined with historical charging records to form a time series feature, and the current health status index of the battery is obtained by comparison and analysis with a preset reference mode. The present invention can accurately assess the battery health status in a multi-vehicle parallel charging environment, and can identify cell consistency problems, providing battery health status assessment services for electric vehicles.

[0100] The above describes the vehicle battery health status assessment method based on the charging pile in the embodiment of the present invention. The following describes the vehicle battery health status assessment system based on the charging pile in the embodiment of the present invention. Figure 3 In one embodiment of the present invention, a vehicle battery health status assessment system based on a charging pile includes: The data acquisition module 301 is used to collect electrical parameters during the connection between the charging pile and the vehicle, and mark key events in the charging process to obtain a structured data stream; A waveform decomposition module 302 is configured to decompose the charging current waveform according to the structured data stream, and generate a basic charging current curve and a cell consistency index based on the separated basic charging current component and the pulsating component caused by the balancing circuit; The response characteristic module 303 is used to calculate the current response characteristics in different state of charge intervals based on the basic charging current curve, and perform corrections based on the power distribution state of the charging pile to obtain the current response parameters of the battery; The health assessment module 304 is used to form a time series feature based on the current response parameter and the cell consistency index in combination with historical charging records, and compare and analyze the time series feature with a preset reference pattern to obtain the current health status index of the battery.

[0101] In an embodiment of the present invention, the vehicle battery health status assessment system based on the charging pile runs the above-mentioned vehicle battery health status assessment method based on the charging pile. The vehicle battery health status assessment system based on the charging pile collects electrical parameters during the connection process between the charging pile and the vehicle and marks key events, decomposes and processes the charging current waveform, obtains the basic charging current component and the pulsating component caused by the balancing circuit, and generates a basic charging current curve and a cell consistency index; calculates the current response characteristics in different state of charge intervals based on the basic charging current curve, and corrects them in combination with the power distribution state of the charging pile to obtain the current response parameters of the battery; combines the current response parameters and the cell consistency index with the historical charging records to form a time series feature, and obtains the battery current health status index by comparing and analyzing with a preset reference mode. The present invention can accurately assess the battery health status in a multi-vehicle parallel charging environment, and can identify cell consistency problems, providing battery health status assessment services for electric vehicles.

[0102] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0103] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes various media that can store program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0104] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A vehicle battery health status assessment method based on a charging pile, characterized in that: The vehicle battery health status assessment method includes: Collect electrical parameters during the connection between the charging pile and the vehicle, and mark key events in the charging process to obtain a structured data stream; Decomposing the charging current waveform according to the structured data stream, and generating a basic charging current curve and a cell consistency index based on the decomposed basic charging current component and the pulsating component caused by the balancing circuit; Calculate the current response characteristics in different state-of-charge intervals based on the basic charging current curve, and perform corrections based on the power distribution state of the charging pile to obtain the current response parameters of the battery; According to the current response parameters and cell consistency indicators, combined with historical charging records and preset reference modes, a comparison analysis is performed to obtain the current health status index of the battery.

2. The vehicle battery health status assessment method according to claim 1, characterized in that: Decomposing the charging current waveform according to the structured data stream, and generating a basic charging current curve and a cell consistency index based on the decomposed basic charging current component and the pulsating component caused by the balancing circuit include: Determining a time window for operation of the equalization circuit according to the equalization state mark in the structured data stream, and extracting waveform features of the equalization circuit according to the time window to generate equalization circuit parameters; Using the equalization circuit parameters as constraints, the charging current waveform measured by the charging pile is decomposed using the empirical mode decomposition algorithm to obtain the basic charging current component, the pulsating component caused by the equalization circuit, and the environmental noise component; Performing a spectrum analysis on the pulsation component, calculating variation characteristics of the amplitude and duration of the pulsation component, and calculating a cell consistency index characterizing the consistency of cells in the battery pack based on the variation characteristics; The basic charging current component is reconstructed and filtered to generate a basic charging current curve.

3. The vehicle battery health status assessment method according to claim 2, characterized in that: The step of determining a time window for operation of the equalization circuit according to the equalization state mark in the structured data stream, extracting waveform features of the equalization circuit according to the time window, and generating equalization circuit parameters includes: Identifying the start time point and the end time point of the operation of the equalization circuit according to the equalization state mark in the structured data stream, and determining the time window of the operation of the equalization circuit; Performing wavelet transform on the charging current waveform measured by the charging pile within the time window to extract the energy distribution characteristics of the high-frequency component; identifying waveform characteristics caused by the balancing circuit under different states of charge based on the energy distribution characteristics, including the frequency, amplitude, and duration of the waveform; A correlation analysis is performed on the characteristic waveform and the corresponding state of charge, a mapping relationship between the state of charge and the characteristic waveform parameters of the balancing circuit is established, and the balancing circuit parameters are constructed using the mapping relationship.

4. The vehicle battery health status assessment method according to claim 2, characterized in that: The performing of spectrum analysis on the pulsating component, calculating the variation characteristics of the amplitude and duration of the pulsating component, and calculating the cell consistency index characterizing the consistency of the cells in the battery pack according to the variation characteristics includes: Performing a fast Fourier transform on the pulsating component to obtain a frequency spectrum distribution of the pulsating component, and identifying a characteristic frequency interval of the equalizing circuit in the frequency spectrum distribution; Calculating the energy density within the characteristic frequency interval and analyzing the trend of the energy density changing with the state of charge to obtain frequency domain characteristic parameters; Performing envelope analysis on the pulsation component in the time domain to extract the amplitude variation curve and duration distribution of the pulsation waveform; Calculating the amplitude variation coefficient according to the fluctuation degree of the amplitude change curve, and calculating the duration dispersion index according to the concentration degree of the duration distribution; The frequency domain characteristic parameters, the amplitude variation coefficient and the duration dispersion index are combined according to preset weights to obtain a cell consistency index, wherein the weights are adjusted according to the vehicle battery type and charging mode.

5. The vehicle battery health status assessment method according to claim 1, characterized in that: The current response characteristics in different state of charge intervals are calculated based on the basic charging current curve, and the current response parameters of the battery are obtained by correcting the current response characteristics in combination with the power distribution state of the charging pile. Analyzing the basic charging current curve to identify the transition point from the constant current phase to the constant voltage phase during the charging process; Based on the conversion point, the charging process is divided into a constant current charging stage and a constant voltage charging stage, and each stage is divided into a plurality of state of charge intervals according to a predetermined state of charge interval; For each state of charge interval, calculating the ratio of the current change rate to the state of charge change rate, using the ratio as the current response characteristic value of the corresponding state of charge interval, and combining the current response characteristic values ​​of all state of charge intervals in order of state of charge to obtain an initial current response parameter; Collecting real-time power distribution data of the charging pile, and determining whether the charging pile is in power limiting mode within each state of charge interval based on the real-time power distribution data; For the state-of-charge interval in power-limited mode, a power correction coefficient is calculated based on the total load parameters and allocation strategy of the charging pile at that time, and the initial current response parameters in the state-of-charge interval in power-limited mode are corrected using the power correction coefficient; The corrected initial current response parameters of all state-of-charge intervals are arranged in order of the state of charge to obtain the current response parameters of the battery.

6. The vehicle battery health status assessment method according to claim 5, characterized in that: The calculating of the power correction coefficient according to the total load parameter and the allocation strategy of the charging pile in the state of charge interval in the power limiting mode, and using the power correction coefficient to correct the initial current response parameter in the state of charge interval in the power limiting mode includes: Collect the total load parameters of the charging pile and the charging demand parameters of all currently connected vehicles, and calculate the theoretical current value that the vehicle should obtain under ideal non-power-limited conditions based on the total load parameters and charging demand parameters and the power allocation strategy of the charging pile; Calculating the ratio of the actual charging current obtained by the vehicle to the theoretical current value to obtain a power correction coefficient; For each sampling point in the state of charge interval in the power limiting mode, dividing the current response characteristic value at the corresponding moment by the power correction coefficient to obtain a corrected current response characteristic value; The corrected current response characteristic value is used to replace the corresponding value in the initial current response parameter to obtain the modified current response parameter.

7. The vehicle battery health status assessment method according to claim 1, characterized in that: The current response parameter and the cell consistency index are combined with historical charging records to form a time series feature, and the time series feature is compared and analyzed with a preset reference pattern to obtain the battery current health status index, including: According to the current response parameters and the cell consistency index, a time series feature is formed in combination with historical charging records; According to the ambient temperature of the charging process, the time series characteristics are normalized by a temperature correction algorithm, and the data collected under different temperature conditions are mapped to a unified reference temperature condition; Retrieving a reference pattern set corresponding to the vehicle's battery model from a preset reference battery database, and using a weighted dynamic time warping algorithm to calculate a similarity score between the normalized time series features and each pattern in the reference pattern set; The most matching reference mode is determined based on the similarity score, and the current health status index of the battery is calculated based on the battery usage time and the number of charge and discharge cycles.

8. A vehicle battery health status assessment system based on a charging pile, characterized in that: The vehicle battery health status assessment system includes: The data acquisition module is used to collect electrical parameters during the connection between the charging pile and the vehicle, and mark key events in the charging process to obtain a structured data stream; A waveform decomposition module is used to decompose the charging current waveform according to the structured data stream, and generate a basic charging current curve and a cell consistency index based on the decomposed basic charging current component and the pulsating component caused by the balancing circuit; A response characteristic module is used to calculate the current response characteristics in different state of charge intervals based on the basic charging current curve, and perform corrections in combination with the power distribution state of the charging pile to obtain the current response parameters of the battery; The health assessment module is used to form a time series feature based on the current response parameter and the cell consistency index in combination with historical charging records, and compare and analyze the time series feature with a preset reference pattern to obtain the current health status index of the battery.

Images