Ultrasonic Monitoring-Based Early Warning Method for Critical Charging of Extremely Charged Batteries

Abstract

This invention provides a method for early warning of critical charging of ultra-fast charging batteries based on ultrasonic monitoring, belonging to the field of charging early warning technology. It includes: real-time acquisition of the amplitude and propagation time of echo signals to infer the electrolyte concentration and polarization degree inside the battery cell; cross-validation of the electrolyte concentration, polarization degree, and SOC estimate from the BMS to correct cumulative errors caused by temperature and aging; establishing a full-charge critical criterion based on accurate SOC calibration values ​​and impedance change characteristics at the end of cell charging and discharging via ultrasonic monitoring; calculating the remaining range requirement under different operating conditions based on actual cell capacity decay data, the full-charge critical criterion, and user driving conditions, and adjusting the low-battery warning threshold; triggering an ultra-fast charging safety warning and pushing the remaining range to the vehicle terminal when the real-time SOC of the battery cell is less than the warning threshold. This invention provides accurate, dynamic, and safe early warning protection for ultra-fast charging scenarios through a closed-loop design across the entire chain.

Description

Technical Field

[0001] This invention relates to the field of charging early warning technology, and in particular to a critical charging early warning method for ultra-high-capacity batteries based on ultrasonic monitoring. Background Technology

[0002] With the rapid development of new energy technologies, high-speed batteries have been widely used in the field of two-wheeled electric vehicles due to their advantage of rapid energy replenishment. During the charging process, for safety reasons, a charging critical state warning module is often set in the charging system.

[0003] However, current methods for warning of critical battery levels have significant limitations, with core issues concentrated in three areas: reliance on SOC values ​​estimated by the BMS (Battery Management System). However, SOC estimation is susceptible to various factors, such as temperature fluctuations, increased battery aging, and changes in charge / discharge rates, all of which can lead to cumulative errors in estimation. This can cause misjudgments in the timing of low or full charge warnings, potentially resulting in premature false alarms that interfere with usage, or delayed alarms that lead to the risk of battery depletion. Second, full charge warnings often use fixed voltage thresholds for triggering, failing to consider the drift characteristics of the actual full charge voltage after cell aging. This can easily lead to false full charge warnings or stopping charging before the battery is actually fully charged, severely impacting charging efficiency. Third, low battery warnings rely solely on SOC values ​​without considering real-time driving conditions. For example, in scenarios like high-speed driving or climbing hills where power consumption is high, it can easily lead to misjudgments where the remaining battery power seems sufficient but is insufficient to reach a charging station, significantly reducing the practicality of the warnings and failing to meet actual user needs. Summary of the Invention

[0004] This invention provides a method for early warning of critical charging of ultra-high-capacity batteries based on ultrasonic monitoring, in order to overcome the deficiencies in the prior art.

[0005] On one hand, the present invention provides a method for early warning of critical charging of ultra-high-capacity batteries based on ultrasonic monitoring, comprising:

[0006] S1: Emit high-frequency ultrasonic waves to the battery cell and collect the amplitude and propagation time of the echo signal in real time, thereby inferring the electrolyte concentration and polarization degree inside the battery cell;

[0007] S2: Cross-validate the electrolyte concentration, polarization degree and the SOC estimate of BMS to correct the cumulative error caused by temperature and aging, and obtain an accurate SOC calibration value.

[0008] S3: Based on the accurate SOC calibration value and the extracted echo signal characteristics, establish a full-charge critical criterion for non-fixed voltage;

[0009] S4: Calculate the actual capacity decay data of the battery cell, and based on the decay data and the full charge threshold, link the user's driving conditions, calculate the remaining range demand under different conditions through the energy consumption model, and dynamically adjust the low battery warning threshold.

[0010] S5: When the real-time SOC of the battery cell is less than the warning threshold, the Extreme Charge safety warning is triggered, the BMS is linked to limit the initial charging power, and range and charging suggestions are pushed to the vehicle terminal.

[0011] S6: Record battery change data after each warning is triggered, input the change data into the cross-validation model and energy consumption model, and update the parameters of the cross-validation model and energy consumption model.

[0012] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-capacity batteries provided by the present invention, the specific steps in step S1 for calculating the electrolyte concentration and polarization degree of the battery cell based on the echo signal replication and propagation time are as follows:

[0013] S11: Collect historical echo amplitude values ​​of electrolytes of different concentrations at various temperatures, covering the set concentration and temperature range, to obtain the basic dataset.

[0014] S12: Training using the base dataset The model is optimized by parameters to ensure that the model error meets the requirements, thus obtaining a concentration estimation model.

[0015] S13: The propagation time of the cell's non-polarized state is collected as a reference. Obtain the real-time propagation time under different polarization operations Calculate the polarization propagation time difference Δt1 and correlate it with the degree of polarization to construct a polarization dataset.

[0016] S14: Input the polarization dataset into the BP neural network, with Δt1 and temperature as inputs and polarization coefficient as output. Set the number of neurons in the hidden layer to train the model and obtain the polarization estimation model.

[0017] S15: Input the real-time acquired echo signal into the concentration estimation model to obtain the current electrolyte concentration and calculate the real-time propagation time. According to the benchmark Calculate the real-time polarization propagation time difference Δt2, input Δt2 and the current temperature into a BP neural network, and output the polarization degree coefficient.

[0018] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-capacity batteries provided by the present invention, in the steps The specific steps to correct the cumulative error caused by temperature and aging are as follows:

[0019] S21: Integrate the electrolyte concentration, polarization degree, and the current SOC estimate output by the BMS into an initial dataset for cross-validation.

[0020] S22: Construct a Kalman filter fusion model based on the initial dataset, using electrolyte concentration and polarization as state observation variables, and the SOC estimate of the BMS as the initial state value. Establish dynamic correlations between variables through filtering equations to form a multi-source data fusion framework.

[0021] S23: Introduce temperature and aging factors into the multi-source data fusion framework, transform the variation law of electrolyte concentration with temperature and the correlation characteristics between polarization degree and aging degree into error correction coefficients and embed them into the prediction-update link of the Kalman filter fusion model to quantitatively compensate for the influence of temperature and aging, and obtain the optimized Kalman filter fusion model.

[0022] S24: The real-time electrolyte concentration and polarization are optimized by Kalman filter fusion model to iteratively calculate and update the observed values, correct the SOC estimation results, and obtain the corrected SOC value.

[0023] S25: If the difference between the corrected SOC value and the original BMS estimate exceeds the threshold, a forced calibration mechanism is performed to feed back the corrected SOC value to the BMS and update its internal ampere-hour integration benchmark.

[0024] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-capacity batteries provided by the present invention, the specific steps for obtaining the current SOC estimate output by the BMS in step S21 are as follows:

[0025] S211: By using the ampere-hour integration algorithm built into the BMS, combined with the real-time current collected by the current sensor in the cell charging and discharging circuit, the change in charge is calculated by integrating the current over time. The open-circuit voltage method is used to calibrate the SOC reference for the initial and fully charged states, and a preliminary SOC estimate is output.

[0026] S212: Integrate the preliminary SOC estimate with the electrolyte concentration and polarization data obtained from ultrasonic testing to form an initial input dataset for cross-validation, providing a complete set of basic parameters containing the original BMS estimation results for subsequent error correction using the Kalman filter algorithm.

[0027] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-capacity batteries provided by the present invention, the specific steps for establishing a non-fixed voltage full-charge critical criterion in step S3 are as follows:

[0028] S31: Extract signal features. Extract the impedance data of the echo signal and establish a relational dataset based on the accurate SOC calibration value and impedance data.

[0029] S32: Analyze the impedance variation pattern within different SOC ranges based on the relational dataset, capture the impedance mutation characteristics when the SOC is close to full charge, and determine the starting point and peak value of the impedance mutation.

[0030] S33: Based on the accurate SOC calibration value, the impedance mutation characteristics are correlated with the cell voltage data to construct a preliminary full-charge criterion model with impedance mutation and SOC threshold as the core.

[0031] S34: Collect charging and discharging data of cells with different aging levels and input them into the preliminary full-charge criterion model for training and optimization. Correct the impact of aging on impedance and voltage to form a non-fixed voltage full-charge critical criterion that is suitable for the entire life cycle.

[0032] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-charge batteries provided by the present invention, the specific steps in step S32, which involve analyzing the impedance variation law within different SOC intervals based on the relational dataset, are as follows:

[0033] The relational dataset is divided into SOC-level intervals based on the SOC value.

[0034] Based on the impedance data within the SOC-level interval analysis interval, the average impedance and the magnitude of change in each interval are calculated to obtain the overall trend of impedance change as SOC increases.

[0035] Zoom in on the impedance data within the high-range of SOC to determine if the impedance is rising rapidly. If so, mark the initial position where the impedance begins to deviate from the stable trend.

[0036] Starting from the initial marked position, analyze the changes in impedance data, find the maximum value point during the impedance rise process, and confirm that this point is a stable abrupt peak based on the data before and after.

[0037] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-charge batteries provided by the present invention, the specific steps in step S33 for constructing a preliminary full-charge criterion model are as follows:

[0038] The precise SOC calibration value, impedance change characteristics, and corresponding cell voltage data are integrated into a correlated dataset.

[0039] Based on the associated dataset, the SOC and voltage values ​​corresponding to the impedance mutation start point and peak value were selected to obtain the parameter co-variance law.

[0040] By setting impedance mutation characteristics as the core triggering condition and setting the SOC threshold based on the accurate SOC calibration value, a preliminary logical correlation framework for the parameters is constructed.

[0041] The logical framework is transformed into a mathematical model, the weights and judgment rules of each parameter are defined, and the parameter weights are adjusted to form a preliminary full-charge criterion model.

[0042] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-capacity batteries provided by the present invention, the specific steps for dynamically adjusting the low battery warning threshold in step S4 are as follows:

[0043] S41: Combine the actual capacity decay data of the battery cells, the full-charge criticality criteria, and the user's driving condition information to form an analysis dataset.

[0044] S42: Collect users' historical driving data, classify and extract energy consumption characteristics of typical operating conditions, and construct energy consumption models under different operating conditions.

[0045] S43: Extract key energy consumption features from the dataset under different operating conditions and input them into the energy consumption model to calculate the actual energy consumption value per unit SOC under various operating conditions, thus obtaining the operating condition energy consumption data.

[0046] S44: Input the operating condition energy consumption data into the energy consumption model to calculate the minimum remaining power requirement to ensure the driving range under different operating conditions.

[0047] S45: Based on the minimum remaining power requirement under different operating conditions, compare the actual range of the current remaining power and dynamically adjust the low power warning threshold.

[0048] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-capacity batteries provided by the present invention, the specific steps for obtaining operating condition energy consumption data in step S43 are as follows:

[0049] Real-time acquisition of current driving parameters, comparison with historical operating condition feature database to identify the current operating condition type, and extraction of key real-time energy consumption features under this operating condition.

[0050] By inputting the key features of real-time energy consumption into the energy consumption model for the corresponding operating condition, and calculating the actual energy consumption value per unit SOC under the current actual usable capacity of the battery cell, a correspondence between energy consumption and SOC is established.

[0051] The unit SOC energy consumption calculation results under various operating conditions are labeled with the corresponding operating condition type and environmental conditions to form a complete operating condition energy consumption dataset.

[0052] According to the ultrasonic monitoring-based critical charging early warning method for ultra-high-charge batteries provided by the present invention, the specific steps for triggering the ultra-high-charge safety early warning in step S5 are as follows:

[0053] S51: Monitor the current SOC value of the battery cell in real time, continuously compare it with the dynamically adjusted warning threshold, and determine whether the real-time SOC of the battery cell is less than the warning threshold. If so, generate a trigger judgment result.

[0054] S52: Based on the judgment result, activate the Extreme Charge safety early warning mechanism, simultaneously send an early warning signal to the BMS, and activate the power limit command preparation process.

[0055] S53: After receiving the warning signal, the BMS limits the initial charging power according to the current polarization degree and temperature status parameters of the battery cell, based on a preset strategy.

[0056] S54: Generate charging suggestions based on the current remaining range and information on nearby charging stations.

[0057] The ultrasonic monitoring-based critical charging early warning method for ultra-high-charge batteries provided by this invention offers accurate, dynamic, and safe early warning protection for ultra-high-charge scenarios through a closed-loop design across the entire chain. Its core advantages are as follows: Based on high-frequency ultrasonic monitoring, this invention uses real-time back-calculation of the cell's electrolyte concentration and polarization degree. It then constructs a Kalman filter fusion model by combining the original SOC estimate from the BMS, introducing temperature and aging factors to quantify and compensate for errors. A forced calibration mechanism updates the ampere-hour integral benchmark, completely resolving the cumulative deviation problem caused by environmental factors and aging in traditional SOC estimation. This provides high-precision SOC data support for early warning and avoids timing deviations in early warning caused by misjudgments of SOC.

[0058] Based on precise SOC calibration values ​​and impedance change characteristics monitored by ultrasound, the solution establishes a non-fixed voltage full-charge critical criterion. By capturing impedance mutation characteristics before full charge and associating them with the entire life cycle data of the battery cell, it dynamically adapts to parameter drift after battery cell aging, effectively avoiding false full-charge warnings or stopping charging before reaching actual full charge. This ensures energy replenishment efficiency and full utilization of battery capacity in extreme charging scenarios, while avoiding the risk of overcharging.

[0059] By linking user driving conditions, the solution calculates the minimum remaining power requirement under different driving conditions based on the cell capacity decay model and the energy consumption model under different driving conditions, and dynamically adjusts the low power warning threshold. This solves the problem of misjudgment caused by the traditional fixed threshold not taking into account the differences in driving conditions, which may result in the remaining power appearing sufficient but not enough to support the journey to the charging station. It ensures that the warning is adapted to the common and complex driving conditions of users who use high-speed and hill-climbing charging stations, thereby improving the practicality of the warning and the user's sense of security when traveling.

[0060] Upon triggering the warning, the solution, in conjunction with the BMS, limits the initial charging power based on the real-time status of the battery cell. It also provides precise suggestions based on remaining range and information on nearby charging stations. This avoids damage to low-charge cells caused by high power charging at the beginning of the process, reducing safety risks, and provides users with personalized charging guidance, enhancing the charging experience. Simultaneously, by recording actual data after the warning and updating cross-validation and energy consumption models, a dynamic optimization loop is formed, continuously improving model adaptability and warning accuracy. This ensures the solution maintains high reliability throughout the battery's entire lifecycle, providing end-to-end technical support for the safe and efficient use of high-charging batteries. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of the steps of the ultrasonic monitoring-based critical charging early warning method for ultra-high-capacity batteries provided in an embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of the steps for establishing a full-charge critical criterion with a non-fixed voltage, provided in an embodiment of the present invention. Detailed Implementation

[0064] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0065] The following is combined Figures 1-2 This invention describes a method for early warning of critical charging of ultra-high-capacity batteries based on ultrasonic monitoring.

[0066] like Figures 1-2 As shown in the embodiment of the present invention, the critical charging early warning method for ultra-high-capacity batteries based on ultrasonic monitoring includes:

[0067] S1: Emit high-frequency ultrasonic waves to the battery cell and collect the amplitude and propagation time of the echo signal in real time, thereby inferring the electrolyte concentration and polarization degree inside the battery cell.

[0068] S11: Collect historical echo amplitude values ​​of electrolytes of different concentrations at various temperatures, covering the set concentration and temperature range, to obtain the basic dataset.

[0069] S12: Train the SVR model using the basic dataset, optimize the parameters to make the model error meet the requirements, and obtain the concentration estimation model.

[0070] S13: Collect the propagation time of the cell in the non-polarized state as a reference t0, and obtain the real-time propagation time t under different polarization operations. real Calculate the polarization propagation time difference Δt1 and correlate it with the degree of polarization to construct a polarization dataset.

[0071] The formula for calculating the polarization propagation time difference Δt1 is:

[0072]

[0073] in This represents the real-time propagation time of ultrasonic waves inside the battery cell. This is the reference propagation time for the cell in its non-polarized state.

[0074] S14: Input the polarization dataset into the BP neural network, with Δt1 and temperature as inputs and polarization coefficient as output. Set the number of neurons in the hidden layer to train the model and obtain the polarization estimation model.

[0075] S15: Input the real-time acquired echo signal into the concentration estimation model to obtain the current electrolyte concentration and calculate the real-time propagation time. Combine the reference t0 to obtain Δt2, input Δt2 and the current temperature into the BP neural network, and output the polarization coefficient.

[0076] S2: Based on the collected electrolyte concentration and polarization parameters, cross-validate with the SOC (State of Charge) estimate from the BMS (Battery Management System) to correct the cumulative errors caused by temperature and aging.

[0077] S21: Obtain the electrolyte concentration and polarization obtained by ultrasonic testing, as well as the current SOC estimate output by BMS, and integrate these three types of data into the initial input dataset for cross-validation.

[0078] S211: By integrating the current over time using the ampere-hour integral algorithm built into the BMS and the real-time current collected by the current sensor in the cell charging and discharging circuit, the change in charge is calculated. The open-circuit voltage method is used to calibrate the SOC reference for the initial and fully charged states, and a preliminary SOC estimate is output.

[0079] The formula for the ampere-hour integral algorithm is expressed as follows:

[0080]

[0081] In the formula, This represents the percentage of remaining battery power initially estimated by the BMS at time t. The initial SOC baseline value. This refers to the rated capacity of the battery cell, which is the total amount of electricity it can release when fully charged. To improve the charging and discharging efficiency of the battery cell, Let τ be the charging and discharging current. This represents the cumulative charge and discharge capacity from time t0 to time t.

[0082] The open-circuit voltage (OCV) calibration formula is expressed as follows:

[0083]

[0084] In the formula, The open-circuit voltage of the cell after it has been left to rest at time t0 reflects the actual energy storage state of the cell. Given the cell temperature at time t0, the effect of temperature on the OCV-SOC mapping relationship is corrected. The pre-calibrated open-circuit voltage-temperature-SOC mapping function ensures that the initial SOC reference matches the actual energy storage state of the cell.

[0085] S212: Integrate the preliminary SOC estimate with the electrolyte concentration and polarization data obtained from ultrasonic testing to form an initial input dataset for cross-validation, providing a complete set of basic parameters containing the original BMS estimation results for subsequent error correction using the Kalman filter algorithm.

[0086] S22: Based on the integrated dataset, a Kalman filter fusion model is constructed, with electrolyte concentration and polarization as state observation variables, and the SOC estimate of BMS as the initial state value. The dynamic relationship between variables is established through the filtering equation, forming a multi-source data fusion framework.

[0087] Electrolyte concentration and polarization are defined as state observation variables to reflect the real-time internal state of the cell. The current SOC estimate output by the BMS is set as the initial state value of the model to provide the starting point for calculation.

[0088] The state equation and observation equation of the Kalman filter fusion model are constructed. The dynamic relationship between the observed variables and the initial state values ​​is established through the equations. The influence weight of each variable on the SOC estimation is quantified, and a multi-source data fusion framework is formed.

[0089] The model is initialized and debugged using historical data from the initial dataset to determine the initial values ​​of the process noise and observation noise covariance matrices in the filtering equation, ensuring the stability and accuracy of the model in its initial state.

[0090] S23: Introduce temperature and aging factors into the fusion model, and transform the variation law of electrolyte concentration with temperature and the correlation characteristics between polarization degree and aging degree into error correction coefficients. Embed the prediction-update link of the filter to achieve quantitative compensation for the effects of temperature and aging.

[0091] Electrolyte concentration data under different temperature conditions were collected, and the variation of electrolyte concentration with temperature was analyzed. Polarization data of cells with different aging levels were recorded simultaneously, and the correlation between polarization and aging levels was explored to establish a quantitative relationship between the two types of characteristics.

[0092] The above-mentioned temperature-related laws and aging-related characteristics are transformed into specific error correction coefficients. The coefficient values ​​are dynamically adjusted according to the temperature and aging degree to ensure that the coefficients can accurately reflect the influence of the two types of factors on the SOC estimation.

[0093] By embedding the error correction coefficient into the prediction-update stage of the Kalman filter model within the multi-source data fusion framework, errors that may be caused by temperature and aging are compensated in advance during the prediction stage, and observation bias is further corrected during the update stage, resulting in an optimized Kalman filter fusion model.

[0094] S24: Run Kalman filtering for iterative calculation. In each iteration, the observed values ​​are updated using the real-time acquired electrolyte concentration and polarization. The estimated SOC results are continuously corrected, and the deviation from the true value is gradually reduced until the error stabilizes within the set range.

[0095] The electrolyte concentration and polarization of the battery cell are collected in real time by high-frequency ultrasonic monitoring. The data is preprocessed in real time to ensure the timeliness and accuracy of the data, which is then used as the real-time input parameters for the model.

[0096] The pre-processed real-time electrolyte concentration and polarization degree are input into the optimized Kalman filter fusion model. The model performs iterative calculations according to the preset equations, continuously updating the estimated values ​​of the observed variables and dynamically correcting the initial SOC estimation results.

[0097] After the iterative calculation converges, the final SOC correction value is output. This value integrates the real-time internal state and the dynamic error compensation results, and the accuracy is significantly improved compared with the original estimate, providing reliable data support for subsequent steps.

[0098] S25: When the difference between the SOC correction value of the filter output and the original estimated value of the BMS exceeds the threshold, the forced calibration mechanism is triggered, and the corrected SOC value is fed back to the BMS to update its internal ampere-hour integration reference, thus completely eliminating the accumulated temperature and aging errors.

[0099] Extract the SOC correction value and the original SOC estimate from the BMS, calculate the difference between the two, compare the difference with the preset error threshold, and determine whether a forced calibration mechanism needs to be activated.

[0100] If the difference exceeds the preset threshold, a forced calibration process is immediately triggered. The system generates a calibration command and feeds back the accurate SOC correction value as the standard value to the BMS control system.

[0101] After receiving the SOC correction value, BMS updates the initial benchmark of its internal ampere-hour integration algorithm based on this value, corrects historical accumulated errors, ensures that subsequent SOC estimations are based on an accurate benchmark, and improves the long-term operating accuracy of BMS.

[0102] S3: Based on the accurate SOC calibration value and the impedance change characteristics at the end of the cell's charging and discharging period monitored by ultrasound, a full-charge critical criterion with a non-fixed voltage is established.

[0103] S31: By combining the precise SOC calibration value and the echo signal at the end of the cell's charge and discharge period monitored by high-frequency ultrasonic equipment, and extracting the impedance data through signal characteristics, a dataset relating impedance and SOC is established.

[0104] S32: Analyze the impedance variation pattern within different SOC ranges based on the relational dataset, capture the impedance mutation characteristics when the SOC is close to full charge, and determine the starting point and peak value of the impedance mutation.

[0105] The dataset of correlated impedance and SOC relationships is divided into intervals according to the SOC values.

[0106] Based on the defined SOC intervals, the impedance data within each interval is analyzed segment by segment. The average impedance and the magnitude of change in each interval are calculated, and the overall trend of impedance change with the increase of SOC is identified.

[0107] Focus on the range with higher SOC values, zoom in on the impedance data details within that range, observe whether there are any sudden changes in impedance with rapid increases, and mark the initial position where the impedance begins to deviate from the stable trend.

[0108] Starting from the initial marked position, the changes in subsequent impedance data are tracked to find the maximum value point during the impedance rise process. The data before and after the point are combined to confirm that the point is a stable abrupt peak, thus completing the core feature capture.

[0109] S33: Based on the accurate SOC calibration value, the impedance mutation characteristics are correlated with the cell voltage data to construct a preliminary full-charge criterion model with impedance mutation and SOC threshold as the core.

[0110] First, we organize accurate SOC calibration values, impedance change characteristics, and corresponding cell voltage data to ensure that the three are completely matched in the time dimension, forming a related dataset containing multi-dimensional parameters, providing a unified benchmark for model construction.

[0111] Based on the associated dataset, the SOC and voltage values ​​corresponding to the impedance change initiation point and peak value were selected, and the correspondence between the parameters at these key nodes was analyzed to clarify the coordinated change law of each parameter during the full-charge process.

[0112] Based on the identified parameter relationships, impedance mutation characteristics are set as the core triggering condition, and the SOC threshold is set in combination with the corresponding accurate SOC calibration value, thus initially establishing a logical relationship framework between parameters.

[0113] The constructed logical framework is transformed into a mathematical model, the weights of each parameter and the judgment rules are defined, the model output results are verified through partial samples of the dataset, and the parameter weights are adjusted to form a preliminary full-charge criterion model.

[0114] S34: Collect charging and discharging data of cells with different aging levels and input them into the preliminary full-charge criterion model for training and optimization. Correct the impact of aging on impedance and voltage to form a non-fixed voltage full-charge critical criterion that is suitable for the entire life cycle.

[0115] S4: Based on the actual capacity decay data of the battery cells and the full charge threshold, and in conjunction with the user's driving conditions, calculate the remaining range demand under different conditions through the energy consumption model, and dynamically adjust the low battery warning threshold.

[0116] S41: Collect capacity decay data throughout the entire life cycle of the battery cell, establish a capacity decay model, and output the current actual usable capacity of the battery cell.

[0117] S42: Collect users' historical driving data, classify and extract energy consumption characteristics of typical operating conditions, and construct energy consumption models under different operating conditions.

[0118] The system collects data on the user's historical driving speed, acceleration and deceleration frequency, and air conditioning usage status through the in-vehicle system, and classifies and labels the data according to urban congestion, highway cruising, and suburban road scenarios.

[0119] For each type of operating condition data, the average energy consumption, peak energy consumption, and energy consumption fluctuation coefficient characteristic values ​​are calculated. The correlation between energy consumption and driving behavior under different operating conditions is analyzed to form a characteristic dataset.

[0120] Using operating condition characteristics as input and actual energy consumption as output, a gradient boosting algorithm is used to construct a sub-operating condition energy consumption model. The model structure is optimized through cross-validation to ensure that the model can accurately reflect the energy consumption characteristics of different operating conditions.

[0121] S43: Extract key energy consumption features for different operating conditions and input them into the energy consumption model to calculate the actual energy consumption value per unit SOC under various operating conditions, thus obtaining operating condition energy consumption data.

[0122] By capturing current driving parameters in real time through onboard sensors, comparing them with historical operating condition feature databases to identify the current operating condition type, and extracting key real-time energy consumption features under that operating condition.

[0123] The extracted key features are input into the energy consumption model for the corresponding operating condition. Combined with the current actual usable capacity of the battery cell, the actual energy consumption value corresponding to the unit SOC under the operating condition is calculated, and the correspondence between energy consumption and SOC is established.

[0124] The formula for calculating actual energy consumption is as follows:

[0125]

[0126] In the formula, The actual energy consumption per unit SOC under the target operating condition. The energy consumption per unit mileage of the vehicle under the target operating conditions. This represents the current actual usable capacity of the battery cell. For the overall vehicle energy conversion efficiency, This is the dynamic full-charge SOC threshold.

[0127] The unit SOC energy consumption calculation results under various operating conditions are summarized, and the corresponding operating condition type and environmental conditions are labeled to form a complete operating condition energy consumption dataset, which provides support for subsequent range requirement calculations.

[0128] S44: Operating condition energy consumption data, actual cell capacity decay and full charge threshold criteria, calculate the minimum remaining power requirement to ensure range under different operating conditions through energy consumption model.

[0129] By integrating operating condition energy consumption data, the current actual usable capacity of the battery cell, and the SOC boundary value in the full-charge critical criterion, the correlation logic of each parameter in energy consumption calculation is clarified, and a multi-parameter input system is constructed.

[0130] A basic range guarantee target is set based on the user's daily single-trip mileage statistics. The amount of electricity required to meet this target under different operating conditions is calculated through an energy consumption model, and the calculation results are corrected by taking into account the capacity decay.

[0131] Based on energy consumption fluctuations and charging convenience under extreme operating conditions, the back-calculated power demand is redundantly corrected to ultimately determine the minimum remaining power demand to ensure battery life under different operating conditions.

[0132] S45: Based on the minimum remaining power requirement under different operating conditions, compare the actual range of the current remaining power and dynamically adjust the low power warning threshold.

[0133] The current remaining SOC of the battery cell is obtained through the BMS, and the actual driving range corresponding to the current remaining power is calculated by combining the unit SOC energy consumption value under the current operating conditions.

[0134] The formula for calculating the actual driving range corresponding to the current remaining battery power is expressed as follows:

[0135]

[0136] in, This represents the actual driving range corresponding to the current remaining battery power. This represents the current actual usable capacity of the battery cell. This is the rated voltage of the battery cell. For the overall vehicle energy conversion efficiency, The current remaining SOC represents the proportion of the current remaining power capacity to the actual available capacity. This represents the actual energy consumption per unit SOC under the current operating conditions.

[0137] The current actual battery life is compared with the minimum remaining battery power requirement under this operating condition. If the actual battery life is close to the required value, the low battery warning threshold is initially adjusted upward or downward according to preset rules.

[0138] Continuously monitor changes in operating conditions and the decay of remaining battery power. When operating conditions change or the remaining battery power changes rapidly, update the minimum remaining battery power requirement in real time and adjust the warning threshold to ensure that the threshold always adapts to the current driving status.

[0139] S5: When the real-time SOC of the battery cell is less than the warning threshold, the Extreme Charge safety warning is triggered, the BMS is linked to limit the initial charging power, and range and charging suggestions are pushed to the vehicle terminal.

[0140] S51: Monitor the current SOC value of the battery cell in real time, continuously compare it with the dynamically adjusted warning threshold, and determine whether the real-time SOC of the battery cell is less than the warning threshold. If so, generate a trigger judgment result.

[0141] S52: Based on the judgment result, activate the Extreme Charge safety early warning mechanism, simultaneously send an early warning signal to the BMS, and activate the power limit command preparation process.

[0142] S53: After receiving the warning signal, the BMS limits the initial charging power according to the current polarization degree and temperature status parameters of the battery cell, based on a preset strategy.

[0143] S54: Generate charging suggestions based on the current remaining range and information on nearby charging stations.

[0144] S6: Records the actual range data, cell temperature rise, polarization change, and full charge criterion execution effect after each warning is triggered. These data are then input into the cross-validation model and the energy consumption model, and the model parameters are updated periodically.

[0145] After each warning is triggered, the system automatically records actual range data, cell temperature rise curves, polarization trend changes, and execution deviation values ​​of the full-charge criterion, organizing these into a structured feedback dataset in chronological order. Based on this dataset, cell state parameters relevant to the cross-validation model are selected, and effective features are extracted and input into the cross-validation model's parameter optimization module. Simultaneously, operating condition energy consumption, range deviation, and energy consumption model-related data are extracted from the feedback dataset, transformed into training samples recognizable by the model, and input into the energy consumption model's iterative update module. At fixed intervals, the input feedback data is used to recalibrate the parameters of both the cross-validation model and the energy consumption model. The model weights are adjusted using an error minimization algorithm to improve the accuracy of subsequent state assessments and energy consumption predictions.

[0146] In summary, this embodiment provides a critical charging early warning method for ultra-high-charge batteries based on ultrasonic monitoring. Through a closed-loop design across the entire chain, it provides accurate, dynamic, and safe early warning protection for ultra-high-charge scenarios. The core benefits are reflected in the following aspects: The solution is based on high-frequency ultrasonic monitoring, which back-calculates the electrolyte concentration and polarization degree of the battery cell in real time. It constructs a Kalman filter fusion model by combining the original SOC estimate from the BMS, introduces temperature and aging factors to quantify and compensate for errors, and updates the ampere-hour integral benchmark through a forced calibration mechanism. This completely solves the problem of cumulative deviation caused by environmental factors and aging in traditional SOC estimation, providing high-precision SOC data support for early warning and avoiding early warning timing deviations caused by SOC misjudgment.

[0147] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus necessary general-purpose hardware platforms, and of course, it can also be implemented using hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as... A magnetic disk, an optical disk, including a number of instructions for causing a computer device (which may be a personal computer, a server, or a network device) to perform methods of various embodiments or parts thereof.

[0148] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or substitutions can be made to some of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for critical charging early warning of polarized battery based on ultrasonic monitoring, characterized in that, Comprise: S1: emit high-frequency ultrasonic waves to the battery cell, real-time collect the amplitude and propagation time of echo signal, and inversely deduce the electrolyte concentration and polarization degree inside the battery cell; The specific steps for calculating the electrolyte concentration and polarization degree of the battery cell according to the amplitude and propagation time of the echo signal are: S11: collect historical echo amplitudes of electrolytes with different concentrations at different temperatures, cover the set concentration and temperature range, and obtain a basic data set; S12: train the SVR model using the basic data set, optimize the parameters to make the model error meet the requirements, and obtain the concentration estimation model; S13: Collect the propagation time of the battery cell in the non-polarization state as the reference t0, and obtain the real-time propagation time t under different polarization operations real , calculate the polarization propagation time difference At1 and correlate the polarization degree to construct a polarization dataset; S14: input the polarization data set into the BP neural network, take Δt1 and temperature as input, and take polarization coefficient as output, set the number of hidden layer neurons to train the model, and obtain the polarization degree estimation model; S15: input the real-time collected echo signal into the concentration estimation model to obtain the current electrolyte concentration, and calculate the real-time propagation time t real According to the reference t0, calculate the real-time polarization propagation time difference Δt2, input Δt2 and the current temperature into the BP neural network, and output the polarization degree coefficient; S2: cross-check the electrolyte concentration, polarization degree and BMS SOC estimation value, correct the cumulative error caused by temperature and aging, and obtain the accurate SOC calibration value; The specific steps for correcting the cumulative error caused by temperature and aging are: S21: integrate the electrolyte concentration, polarization degree and the current SOC estimation value output by the BMS into an initial data set for cross-checking; S22: build a Kalman filter fusion model according to the initial data set, take the electrolyte concentration and polarization degree as state observation variables, and take the BMS SOC estimation value as the initial state value, establish the dynamic correlation between variables through the filter equation, and form a multi-source data fusion framework; S23: introduce temperature and aging factors into the multi-source data fusion framework, convert the temperature variation law of electrolyte concentration and the correlation characteristics between polarization degree and aging degree into error correction coefficients, embed the prediction-update link of the Kalman filter fusion model, and quantitatively compensate the temperature and aging effects to obtain an optimized Kalman filter fusion model; S24: input real-time electrolyte concentration and polarization degree into the optimized Kalman filter fusion model for iteration, calculate the observation value, correct the estimation result of SOC, and obtain the SOC correction value; S25: determine whether the difference between the SOC correction value and the original BMS estimation value exceeds the threshold value, and if so, perform a forced calibration mechanism; S3: establish a full-charge critical criterion of non-fixed voltage according to the accurate SOC calibration value and the extracted echo signal characteristics; S4: calculate the actual capacity attenuation data of the battery cell, link the attenuation data and the full-charge critical criterion with the user driving conditions, calculate the remaining endurance demand under different conditions through the energy consumption model, and dynamically adjust the low-power warning threshold; S5: when the real-time SOC of the battery cell is less than the warning threshold, trigger the extreme charging safety warning, link the BMS to limit the initial charging power, and push the endurance and charging suggestions to the vehicle terminal; S6: record the battery change data after each warning trigger, input the change data into the cross-checking model and the energy consumption model, and update the parameters of the cross-checking model and the energy consumption model.

2. The ultrasonic monitoring based critical charging pre-alarm method for polarized battery according to claim 1, characterized in that, In step S21, the specific steps for obtaining the current SOC estimation value output by the BMS are: S211: Through the ampere-hour integration algorithm built-in the BMS, combined with the real-time current collected by the current sensor of the battery charging and discharging circuit, the time integral of the current is calculated to obtain the change of electric quantity, the open circuit voltage method is used to calibrate the initial and full SOC reference, and the preliminary SOC estimation value is output; S212: The preliminary SOC estimation value is integrated with the electrolyte concentration and polarization degree data obtained by ultrasonic detection to form an initial input data set for cross verification.

3. The ultrasonic monitoring based critical charging pre-alarm method for polarized battery of claim 1, wherein, In step S3, the specific steps of establishing the full charge critical criterion of non-fixed voltage are: S31: Extracting impedance data of echo signals, establishing a relationship data set according to the accurate SOC calibration value and the impedance data; S32: According to the relationship data set, the change rule of impedance in different SOC intervals is analyzed, the impedance mutation characteristics are captured, and the starting point and peak value of impedance mutation are determined; S33: According to the accurate SOC calibration value, the impedance mutation characteristics are associated with the battery voltage data to construct a preliminary full charge criterion model with impedance mutation and SOC threshold as the core; S34: Collecting charging and discharging data of battery cells with different aging degrees and inputting them into the preliminary full charge criterion model for training and optimization to correct the influence of aging on impedance and voltage, and form a non-fixed voltage full charge critical criterion suitable for the whole life cycle.

4. The ultrasonic monitoring based critical charging pre-alarm method for polarized battery according to claim 3, characterized in that, In step S32, the specific steps of analyzing the change rule of impedance in different SOC intervals according to the relationship data set are: Divide the relationship data set into SOC level intervals according to the SOC value; According to the SOC level interval, the impedance data in the interval is analyzed, the average value and the change amplitude of the impedance in each interval are calculated, and the overall change trend of the impedance with the increase of SOC is obtained; Amplify the impedance data in the high-level SOC interval to determine whether the impedance rises rapidly. If yes, mark the initial position where the impedance begins to deviate from the stable trend; Taking the marked initial position as the starting point, analyze the change of impedance data, and find the maximum point in the rising process of impedance as the stable mutation peak.

5. The ultrasonic monitoring based critical charging pre-alarm method for polarized battery of claim 3, wherein, In step S33, the specific steps of constructing the preliminary full charge criterion model are: Integrate the accurate SOC calibration value, impedance mutation characteristics and battery voltage data at the corresponding time into a correlation data set; According to the correlation data set, the SOC value and voltage value corresponding to the starting point and peak value of impedance mutation are selected to obtain the parameter collaborative change rule; Taking the impedance mutation characteristics as the core trigger condition, setting the SOC threshold according to the accurate SOC calibration value, and preliminarily constructing the logical association framework of parameters; Convert the logical association framework into a mathematical model, define the weight and judgment rule of each parameter, and adjust the parameter weight to form a preliminary full charge criterion model.

6. The ultrasonic monitoring based critical charging pre-alarm method for polarized battery of claim 1, wherein, In step S4, the specific steps of dynamically adjusting the low power warning threshold are: S41: Forming an analysis data set by combining the actual capacity attenuation data of the battery, the full charge critical criterion and the user driving condition information collected; S42: Collecting user historical driving data, classifying and extracting energy consumption characteristics of typical working conditions, and constructing energy consumption models under different working conditions; S43: Extract the energy consumption key features of different working conditions in the analysis data set, input the energy consumption model, calculate the actual energy consumption value corresponding to the unit SOC under each type of working condition, and obtain the working condition energy consumption data; S44: Input the working condition energy consumption data into the energy consumption model to calculate the minimum remaining power demand for ensuring the endurance under different working conditions; S45: According to the minimum remaining power demand under different working conditions, compare the actual endurance of the current remaining power, and dynamically adjust the low power warning threshold.

7. The ultrasonic monitoring based critical charging pre-alarm method for polarized battery according to claim 6, characterized in that, In step S43, the specific steps for obtaining working condition energy consumption data are as follows: Real-time acquisition of current driving parameters, comparison with historical working condition feature library to identify current working condition type, extraction of real-time energy consumption key features under current working condition; Input the real-time energy consumption key features into the energy consumption model corresponding to the working condition, calculate the actual energy consumption value corresponding to the unit SOC under the current working condition according to the current actual available capacity of the battery cell, and establish the corresponding relationship between energy consumption and SOC; Label the working condition type and environmental conditions corresponding to the data for each type of working condition under the unit SOC energy consumption calculation result, and form a complete working condition energy consumption data set.

8. The ultrasonic monitoring based critical charging pre-alarm method for polarized battery of claim 1, wherein, In step S5, the specific steps for triggering the extreme charging safety warning are as follows: S51: Real-time monitoring of the current SOC value of the battery cell and continuous comparison with the dynamically adjusted warning threshold to determine whether the real-time SOC of the battery cell is less than the warning threshold. If yes, generate a trigger determination result; S52: Start the extreme charging safety warning mechanism according to the determination result, send a warning signal to the BMS synchronously, and activate the power limitation instruction preparation process; S53: After receiving the warning signal, the BMS limits the initial charging power according to the current polarization degree and temperature state parameters of the battery cell according to the preset strategy; S54: Generate a charging suggestion according to the current remaining endurance data and surrounding charging station information.

Images