Power battery safety early warning method, equipment and medium

Abstract

The invention discloses a power battery safety early warning method and device and a medium, and relates to the field of safety monitoring, and the method comprises the steps: processing the real-time operation data of a vehicle according to the vehicle and a timestamp, and determining the processed data; acquiring the charging state of the power battery, segmenting the processed data according to the charging state of the power battery, determining the charging data of the vehicle, and dividing the charging data into a plurality of charging segments according to time; constructing a voltage matrix of the charging fragments; according to the voltage matrix, carrying out battery monomer early warning to obtain a threshold early warning result, a change early warning result and an outlier degree result; and performing safety early warning on the power battery according to the threshold early warning result, the change early warning result and the outlier degree result. According to the application, the real-time response of the battery management system to the battery fault is realized.

Description

Technical Field

[0001] This application relates to the field of safety monitoring, and in particular to a method, device and medium for early warning of power battery safety. Background Technology

[0002] With the widespread application of electric vehicles in the transportation sector, the safety of power batteries has become a critical factor in the design and operation of electric vehicles. During charging and discharging, especially during charging, the voltage performance of individual battery cells directly reflects the safety status and abnormalities of each cell. However, the battery discharge process is subject to numerous uncertainties, such as the operating environment and driver habits, resulting in significant randomness, fluctuations, and noise in the discharge curve. Due to aging and wear during prolonged use, the voltage of different individual cells may deviate significantly, leading to a decline in the overall battery pack performance and potentially triggering serious safety issues such as thermal runaway. Therefore, timely detection of abnormal changes in individual cell voltages and taking corresponding measures are crucial for ensuring the stability and safety of the battery pack. Traditional battery safety monitoring methods typically rely on individual cell voltage acquisition and equalization control within the battery management system, periodically monitoring the voltage values ​​of each cell. However, conventional voltage monitoring methods used by battery management systems often neglect the dynamic changes in individual cell voltages during charging, particularly failing to consider the relative differences between individual cell voltage changes and the overall battery pack voltage changes. Furthermore, when a voltage deviation occurs in a particular cell, it may not be detected accurately and promptly, leading to a delayed system response to battery faults. Summary of the Invention

[0003] The purpose of this application is to provide a power battery safety early warning method, device, equipment, medium and product based on voltage parameter “threshold-rate of change-outlier degree” coupled analysis, so as to solve the problem of battery management system response lag in battery faults.

[0004] To achieve the above objectives, this application provides the following solution:

[0005] Firstly, this application provides a method for safety early warning of power batteries, including:

[0006] Based on the vehicle and timestamp, the real-time operation data of the vehicle is processed to determine the processed data;

[0007] The charging status of the power battery is obtained, and the processed data is segmented according to the charging status of the power battery to determine the vehicle's charging data. The charging data is then divided into multiple charging segments according to time.

[0008] Construct the voltage matrix of the charging segment; each row of the voltage matrix represents a time frame, and each column represents a single battery cell; the power battery is composed of multiple battery cells;

[0009] Based on the voltage matrix, a battery cell early warning is performed to obtain threshold early warning results, change early warning results, and outlier results;

[0010] Safety warnings for power batteries are issued based on the threshold warning results, the change warning results, and the outlier results.

[0011] In one embodiment, the charging state of the power battery is acquired, the processed data is segmented according to the charging state of the power battery to determine the vehicle's charging data, and the charging data is divided into multiple charging segments according to time, specifically including:

[0012] The charging status of the power battery is obtained based on parameters such as vehicle status indicator, current and vehicle speed.

[0013] Based on the charging status of the power battery, the processed data is segmented to determine the vehicle's charging data;

[0014] The charging data is divided into multiple charging segments according to the time sequence.

[0015] In one embodiment, based on the voltage matrix of the charging segment, a battery cell early warning is performed to obtain threshold early warning results, change early warning results, and outlier results, specifically including:

[0016] The first abnormal battery cell in the voltage matrix is ​​selected. Among the selected first abnormal battery cells, the first abnormal battery cells that exceed the overvoltage and undervoltage thresholds are given a threshold warning, and the threshold warning result is determined. The first abnormal battery cell is an overvoltage battery cell or an undervoltage battery cell.

[0017] Obtain the voltage values ​​of all abnormal battery cells in the voltage matrix within a set time window, and calculate the variance of the voltage values ​​within the set time window;

[0018] The volatility evaluation criteria are determined based on the variance of voltage values ​​within a set time window.

[0019] Based on the volatility evaluation criteria, the second abnormal battery cell that exceeds the standard deviation threshold is selected.

[0020] Based on the selected second abnormal battery cells, a change rate warning is issued, and the change rate warning result is determined.

[0021] Based on the voltage matrix, the outlier characterization parameters of the charging segment are determined; the outlier characterization parameters include a first outlier characterization parameter and a second outlier characterization parameter.

[0022] The relationship between the first outlier characterization parameter and the second outlier characterization parameter is determined by principal component analysis (PCA) to screen abnormal charging segments.

[0023] Calculate the offset of the largest similarity between the voltage curve of the battery cell corresponding to all the screened abnormal charging segments and the reference voltage curve;

[0024] The maximum offset is compared with the offset threshold to filter out the third abnormal battery cell that exceeds the offset threshold;

[0025] Outlier warnings are issued based on the selected third abnormal battery cells, and the outlier warning results are determined.

[0026] In one embodiment, a first abnormal battery cell is selected from the voltage matrix. Among the selected first abnormal battery cells, a threshold warning is issued for the first abnormal battery cells that exceed the overvoltage and undervoltage thresholds, and the threshold warning result is determined. Specifically, this includes:

[0027] Based on the voltage matrix, the first abnormal battery cell and the time frame corresponding to the first abnormal battery cell are selected by comparing the voltage value of each cell in each time frame with the overvoltage and undervoltage thresholds.

[0028] The first abnormal battery cell that exceeds the overvoltage and undervoltage thresholds is given a graded warning, and the threshold warning result is determined.

[0029] In one embodiment, a rate of change warning is issued based on the selected second abnormal battery cells, and the rate of change warning result is determined, specifically including:

[0030] Calculate the voltage change amplitude of the second abnormal battery cell in two time frames before and after, and filter out the second abnormal battery cells whose voltage change amplitude in the two frames exceeds the voltage change amplitude threshold and the abnormal time frames corresponding to the second abnormal battery cells.

[0031] A time window is selected based on the second abnormal battery cell and the abnormal time frame corresponding to the second abnormal battery cell.

[0032] Within the time window, the median is used to detrend the voltage of the selected second abnormal battery cells, and the variance of the second abnormal battery cells after detrending is determined.

[0033] Based on the variance of the second abnormal battery cell, filter out the second abnormal battery cells that exceed the variance threshold, perform a rate of change warning, and determine the rate of change warning result.

[0034] In one embodiment, determining the outlier characterization parameters of the charging segment based on the voltage matrix specifically includes:

[0035] The average value of the battery cell voltage in each time frame is calculated, and the average value is used as the reference voltage value to obtain the reference voltage curve of the charging segment.

[0036] Calculate the average value of each column of the voltage matrix to determine the average voltage of the individual battery cells in the charging segment;

[0037] Based on the average voltage of the individual battery cells, the voltage variance of the individual battery cells is calculated, and the voltage variance of the individual battery cells is used as the first outlier characterization parameter of the charging segment.

[0038] Calculate the voltage difference between the voltage curve of each battery cell and the reference voltage curve, and sum all voltage differences; the battery cell curve is formed by connecting the voltage values ​​obtained for each battery cell with a smooth curve.

[0039] Divide the summed pressure difference by the time frame to obtain the average value;

[0040] The maximum value among the average values ​​is taken as the second outlier characterization parameter of the charging segment.

[0041] In one embodiment, the relationship between the first outlier characterization parameter and the second outlier characterization parameter is determined using principal component analysis (PCA) to screen for abnormal charging segments, specifically including:

[0042] The distribution relationship between the first outlier characterization parameter and the second outlier characterization parameter was determined using principal component analysis (PCA).

[0043] Based on the aforementioned distribution relationship, the degree of abnormality of the charging segment is determined according to the Mahalanobis distance, and abnormal charging segments are screened.

[0044] In one embodiment, the maximum offset is compared with an offset threshold to filter out a third abnormal battery cell that exceeds the offset threshold, specifically including:

[0045] Based on the selected abnormal charging segments, the voltage curve of a single battery cell is shifted by a first specified number of frames. Null values ​​generated during the shift of the voltage curve of the single battery cell and voltage values ​​of the reference voltage curve in the same time frame are deleted. The similarity between the shifted voltage curve of the single battery cell and the processed reference voltage curve is then determined.

[0046] The voltage curve of the individual battery cell is shifted by a second specified number of frames. The similarity is calculated repeatedly until the number of shifted frames reaches a preset number of shifted frames. The similarity corresponding to the preset number of shifted frames is then determined.

[0047] Obtain the offset corresponding to the maximum similarity between the voltage curve of each battery cell and the reference voltage curve. Compare the offset corresponding to the maximum similarity with a preset threshold. The battery cells corresponding to the offset corresponding to the maximum similarity exceeding the preset threshold are identified as battery cells with outlier anomalous degree, and the third anomalous battery cells are determined.

[0048] In a second aspect, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the power battery safety warning method described in any one of the above.

[0049] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the power battery safety warning method described in any one of the above.

[0050] According to the specific embodiments provided in this application, the following technical effects are disclosed:

[0051] This application provides a power battery safety early warning method, device, equipment, medium, and product. It processes real-time vehicle operation data based on the vehicle and timestamp; segments the processed data according to battery status to determine the vehicle's charging data, and divides the charging data into multiple charging segments based on time; constructs a voltage matrix for each charging segment; when performing threshold early warning based on the voltage matrix, it compares the voltage range of each battery frame with the threshold, issuing warnings for batteries with large voltage ranges, enabling timely and accurate identification of voltage deviations in individual batteries; when performing rate of change warning, it filters out batteries exceeding a preset standard deviation, enabling real-time detection of the dynamic changes in individual cell voltage during charging; when performing outlier warning, it fully considers the relative differences between the voltage changes of each individual cell and the overall battery pack voltage changes, completing the power battery safety early warning. This solves the problem of delayed response of the battery management system to battery faults, significantly improving the accuracy and reliability of early anomaly detection in the battery pack. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A schematic diagram of a power battery safety early warning method based on voltage parameter “threshold-rate of change-outlier degree” coupled analysis provided in one embodiment of this application;

[0054] Figure 2 This is a flowchart illustrating a power battery safety early warning method according to an embodiment of this application;

[0055] Figure 3 This is a schematic diagram of a charging segment feature parameter outlier warning method provided in an embodiment of this application;

[0056] Figure 4 This is a schematic diagram of a method for early warning of outlier trends provided in an embodiment of this application;

[0057] Figure 5 This application provides a schematic diagram of the structure of a computer device according to one embodiment. Detailed Implementation

[0058] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0059] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] To address the limitations of traditional power battery safety early warning methods, such as the single-dimensional monitoring indicators, the lag in abnormal state identification, and the insufficient adaptability to complex operating conditions, this invention proposes a power battery safety early warning method based on the coupled analysis of voltage parameters "threshold-rate of change-outlier degree". This method focuses on solving the following technical challenges: (1) Existing detection methods based on fixed thresholds struggle to accurately identify early-stage progressive anomalies in batteries and cannot effectively distinguish between normal operating condition fluctuations and potential fault characteristics; (2) Conventional rate of change monitoring technology lacks sensitivity to dynamic consistency degradation in battery packs, easily leading to missed or incorrect judgments of voltage drop mutations; (3) Existing outlier degree analysis methods lack multi-dimensional feature fusion capabilities, exhibiting technical blind spots in dynamic matching of charging sequences and collaborative analysis of SOC trends. By constructing a three-dimensional coupled analysis model of "threshold-rate of change-outlier degree", this invention achieves multi-dimensional collaborative diagnosis of individual battery voltage characteristics during charging, significantly improving the accuracy and reliability of early-stage anomaly detection in battery packs, and providing a complete technical solution for power battery systems, from feature extraction and trend analysis to safety early warning.

[0061] With the rapid development of the electric vehicle industry, the safety monitoring of power batteries, as the core component of electric vehicles, has become a key area of ​​technological research and development. The operation of numerous vehicles generates massive amounts of real-world data, such as battery voltage, current, state of charge, temperature, and operating conditions, laying the foundation for a comprehensive understanding of battery status and ensuring its safe operation. The operating environment of electric vehicles is complex and variable, including different driving behaviors, operating conditions, climate influences, and road conditions. This complexity makes it difficult for battery monitoring to fully capture all potential problems. Data from onboard terminals records key information about vehicle operation in time-series format, providing rich resources for subsequent data mining and analysis. Among these, the voltage data of individual battery cells is a core indicator for judging battery status. By analyzing the voltage data of individual battery cells during actual vehicle operation, safety risks such as battery aging and malfunctions can be identified.

[0062] In actual vehicle operation, batteries are used in various scenarios, including charging, discharging, and resting. Data from the charging process is generally considered an ideal data source for analyzing battery status. During charging, the battery is typically in a static state, free from other load influences (such as fluctuations in power demand), the charging current is relatively stable, and environmental factors have minimal interference with voltage data. This makes the data from the charging segment more reliable and facilitates accurate analysis of voltage changes. The battery charging curve contains several key characteristic points, such as the linear voltage increase during the constant current charging phase and the voltage stability during the constant voltage charging phase. These characteristics reflect the battery's internal resistance, capacity, and health status, providing rich information for identifying abnormal individual cell voltages.

[0063] Existing power battery safety early warning methods mainly rely on the absolute value or statistical characteristics of single cell voltage to judge anomalies. In practical applications, they generally suffer from high false alarm rates and low detection accuracy. Specifically, the methods are as follows: (1) The absolute threshold method, as the current mainstream voltage inconsistency detection method, judges anomalies by setting the upper and lower limits of single cell voltage. However, its threshold setting is fixed and cannot dynamically adapt to complex operating conditions such as temperature changes, current fluctuations and load changes, resulting in frequent false alarms triggered by the threshold. At the same time, this method only focuses on the instantaneous over-limit of the absolute voltage value and ignores the correlation analysis between the single cell voltage and the overall average voltage trend of the battery pack. It is not sensitive enough to early progressive faults (such as slowly increasing SOC deviation or internal resistance decay). (2) Statistical characteristic-based analysis methods (such as the range method and standard deviation method) detect anomalies by quantifying the outlier degree of voltage distribution. However, its model assumes that voltage changes conform to a specific statistical distribution (such as normal distribution). In actual operating environments, the voltage dynamic characteristics of the battery pack are affected by multiple factors and exhibit non-stationary and nonlinear characteristics, resulting in poor adaptability of the statistical model and a significant increase in the false alarm rate. (3) While machine learning-based anomaly detection methods can uncover hidden patterns in voltage data, their performance is highly dependent on the completeness and labeling quality of the training data. In practical applications, the following bottlenecks exist: training data struggles to cover the entire battery lifecycle degradation patterns and extreme operating conditions; the model is prone to overfitting, leading to insufficient generalization ability; and supervised learning paradigms relying on high-precision data labeling face challenges of high labeling costs and poor timeliness in engineering practice. Because these methods are limited to single-dimensional feature extraction or static threshold determination, they struggle to effectively integrate the dynamic evolution of voltage parameters, intra-group consistency correlation features, and temporal trend coupling relationships, resulting in insufficient ability to identify early-stage latent faults in battery packs, severely restricting the reliability improvement of power battery safety early warning systems.

[0064] Solution (1) CN116442786A relates to a method, device, server, and storage medium for identifying abnormal voltage differences in power batteries. The method includes: acquiring historical charging data for each charge of a vehicle within a preset time period; extracting data from the historical charging data that meets preset single-cell voltage conditions, tagging the data to obtain final tag data, and filtering single-frame data that meets preset filtering conditions from the final tag data; calculating the single-cell voltage difference between each single cell in the power battery and a preset standard single cell based on the single-frame data, obtaining the voltage difference change trend of each single cell, and determining that the power battery voltage difference is abnormal when the voltage difference change trend reaches a preset abnormal condition. This application embodiment can calculate the single-cell voltage difference between the voltage of each single cell in the power battery during each charge and a preset standard single cell, identify single-cell voltage difference abnormalities based on the voltage difference change trend, achieve high accuracy in the identification results, and directly identify and locate abnormal cells.

[0065] In response to the fact that the voltage of a single battery cell is affected by dynamic operating conditions such as temperature, current fluctuations and load changes, the voltage difference alone may not accurately reflect the abnormality. Under normal operating conditions, short-term voltage fluctuations may be misjudged as abnormalities. The traditional voltage difference method relies on a preset fixed threshold. When the battery aging degree and operating conditions change, the threshold may lose its applicability, leading to misjudgment or omission. The voltage difference method is usually based on only a single frame or short period of data and cannot capture the long-term trend or periodic abnormality of voltage changes.

[0066] Scheme (2) discloses a method and device for safety early warning of vehicle power battery using entropy value CN116224065A, including: firstly, obtaining the voltage value and battery temperature value of the target vehicle's vehicle power battery within a preset time; then, calculating the battery voltage difference entropy and battery temperature difference entropy of the target vehicle's vehicle power battery within the preset time based on the voltage value and battery temperature value of the target vehicle's vehicle power battery within the preset time; then, determining whether to issue an early warning for the target vehicle's vehicle power battery based on the battery voltage difference entropy and battery temperature difference entropy of the target vehicle's vehicle power battery within the preset time, thereby combining the common changes of the voltage difference and battery temperature difference of the target vehicle's vehicle power battery to issue an early warning for the safety of the target vehicle's vehicle power battery, thereby improving the safety of the target vehicle.

[0067] The entropy calculation in scheme (2) is sensitive to noise in the data, especially in dynamic operating conditions. Normal random fluctuations may cause abnormal changes in entropy, thereby increasing the false alarm rate. As a mathematical indicator, the change of entropy is difficult to be directly related to the specific problems of battery cells and lacks an intuitive physical explanation. The voltage change patterns under different operating conditions are significantly different, and a unified entropy threshold may not be able to effectively cover all scenarios.

[0068] Scheme (3) utilizes the rate of change amplitude CN117054916A to disclose an anomaly detection method, system and cloud server for power batteries. The method includes: acquiring parameter data of the power battery within a target time period; the parameter data includes voltage data sequence and current data sequence of each individual cell in the power battery; for each individual cell in the power battery, determining the voltage change rate amplitude and current change rate amplitude at each frequency point based on the voltage data sequence and current data sequence of the individual cell; then determining the amplitude ratio of the voltage change rate amplitude and current change rate amplitude of the individual cell at each frequency point; and determining whether each individual cell has an abnormal resistance value based on the amplitude ratio of each individual cell at each frequency point. This can accurately determine whether the individual cell has an abnormal resistance value, avoid the manual definition of voltage abnormality threshold or current abnormality threshold, reduce the cost of anomaly detection for power batteries, and improve the efficiency of anomaly detection.

[0069] In scheme (3), the voltage change rate may increase naturally during constant current charging or fast charging. At this time, the change rate amplitude of a single battery cell may be misjudged as abnormal. The change rate amplitude method focuses on local features but ignores the voltage trend relationship between a single battery cell and the overall battery pack, which may miss some abnormal modes such as gradual voltage shift. This method usually requires a high sampling frequency to accurately capture voltage changes, while the data upload frequency of the actual vehicle cloud is low, which may limit the detection effect.

[0070] In summary, most existing methods analyze the absolute value of individual cell voltage in isolation, ignoring its dynamic correlation with the average voltage of the battery pack. They only perform superficial analysis on a single threshold or statistical index, failing to capture the long-term trend and cumulative effect of voltage changes. This invention, based on threshold and rate of change anomaly detection, compares the dynamic differences between the individual cell voltage curve and the average voltage curve in real time, extracting the cumulative effect characteristics of voltage deviation, effectively suppressing false triggering caused by transient disturbances such as temperature fluctuations and current jumps. Simultaneously, by constructing a voltage trend similarity evaluation function, it captures the long-term deviation between the voltage evolution trajectory of individual cells and the overall battery pack, significantly improving the detection sensitivity for early consistency degradation.

[0071] This application proposes a power battery safety early warning method based on the coupled analysis of voltage parameters "threshold-rate of change-outlier degree", the flowchart of which is shown below. Figure 1 As shown, firstly, the voltage sequences of multiple individual charging segments are obtained; then, the reference voltage curve is calculated and the difference between each individual segment and the reference voltage curve is calculated to obtain the characteristic parameters characterizing the safety status of the charging segment; finally, principal component analysis is performed based on the characteristic parameters characterizing the safety status of the charging segment, and faulty segments are screened out using Mahalanobis distance.

[0072] like Figure 2 As shown in the figure, this application provides a power battery safety early warning method, the specific content of which is as follows.

[0073] S1: Process the real-time operating data of the vehicle based on the vehicle and timestamp to determine the processed data.

[0074] S2: Obtain the charging status of the power battery, segment the processed data according to the charging status of the power battery, determine the vehicle's charging data, and divide the charging data into multiple charging segments according to time.

[0075] S3: Construct the voltage matrix of the charging segment; each row of the voltage matrix represents a time frame, and each column represents a battery cell; the power battery is composed of multiple battery cells.

[0076] S4: Based on the voltage matrix, perform battery cell early warning to obtain threshold early warning results, change early warning results, and outlier results.

[0077] S5: Provide a safety warning for the power battery based on the threshold warning result, the change warning result, and the outlier result.

[0078] Furthermore, in an exemplary embodiment, S2 can be replaced by the following steps.

[0079] S201: Obtain the charging status of the power battery based on parameters such as vehicle status indicator, current, and vehicle speed.

[0080] S202: Based on the charging status of the power battery, the processed data is segmented to determine the vehicle's charging data.

[0081] S203: Divide the charging data into multiple charging segments according to the time sequence.

[0082] The system uses languages ​​such as Structured Query Language (SQL) to retrieve real-time vehicle operation data from the database and integrates the data according to the vehicle and timestamp. It determines whether the vehicle is in a charging state based on the vehicle status identifier or parameters such as current and speed, and performs state segmentation on the real vehicle operation data. Based on the segmented vehicle status, it selects data within a specified time range, extracts the individual unit voltage data, and constructs a voltage matrix for the entire charging segment, where each row represents a frame and each column represents an individual unit.

[0083] Furthermore, in an exemplary embodiment, S4 can be replaced by the following steps.

[0084] S401: Select the first abnormal battery cell in the voltage matrix. Among the selected first abnormal battery cells, perform threshold warning for the first abnormal battery cells that exceed the overvoltage and undervoltage thresholds, and determine the threshold warning result; the first abnormal battery cell is an overvoltage battery cell or an undervoltage battery cell.

[0085] S402: Obtain the voltage values ​​of all individual cells in the voltage matrix under the set time window, and calculate the variance of the voltage value of each individual cell under the set time window.

[0086] S403: Determine the volatility evaluation criteria based on the variance of each voltage value within the set time window.

[0087] S404: Based on the volatility evaluation criteria, select the second abnormal battery cell that exceeds the standard deviation threshold.

[0088] S405: Based on the selected second abnormal battery cell, issue a rate of change warning and determine the rate of change warning result.

[0089] S406: Determine the outlier characterization parameters of the charging segment based on the voltage matrix; the outlier characterization parameters include a first outlier characterization parameter and a second outlier characterization parameter.

[0090] S407: Determine the relationship between the first outlier characterization parameter and the second outlier characterization parameter using PCA principal component analysis to screen for abnormal charging segments.

[0091] S408: Calculate the offset of the battery cell voltage curve corresponding to all selected abnormal charging segments with the largest similarity to the reference voltage curve.

[0092] S409: Compare the maximum offset with the offset threshold to filter out the third abnormal battery cell that exceeds the offset threshold.

[0093] S410: Based on the selected third abnormal battery cell, perform outlier warning and determine the outlier warning result.

[0094] Furthermore, S401 specifically includes:

[0095] S4011: Based on the voltage matrix, by comparing the voltage value of each cell in each time frame with the overvoltage and undervoltage thresholds, the first abnormal battery cell and the time frame corresponding to the first abnormal battery cell are selected. A time window is determined (i.e., a time window is set) within the time frame corresponding to the first abnormal battery cell.

[0096] S4012: Perform graded early warning for the first abnormal battery cell that exceeds the overvoltage and undervoltage thresholds, and determine the threshold early warning result.

[0097] Based on the obtained voltage matrix, threshold judgment is performed on the voltage values ​​of all cells in each frame to filter out cells with overvoltage or undervoltage and their corresponding time frames; the voltage range of all cells in each frame is calculated and compared with the threshold to issue warnings for batteries with excessively large ranges or poor cell inconsistency.

[0098] Furthermore, S405 specifically includes:

[0099] S4051: Calculate the voltage change amplitude of the second abnormal battery cell in two time frames before and after, and filter out the second abnormal battery cell and the abnormal time frame corresponding to the second abnormal battery cell whose voltage change amplitude in two frames exceeds the voltage change amplitude threshold.

[0100] S4052: Select a time window based on the second abnormal battery cell and the abnormal time frame corresponding to the second abnormal battery cell.

[0101] S4053: Based on the variance of the second abnormal battery cell, filter out the second abnormal battery cells that exceed the variance threshold, perform a rate of change warning, and determine the rate of change warning result.

[0102] S4054: Based on the variance of the second abnormal battery cell, filter out the second abnormal battery cells that exceed the variance threshold, perform a rate of change warning, and determine the rate of change warning result.

[0103] The voltage change of each cell between two consecutive frames is calculated and a threshold judgment is made to filter out cells with excessive voltage changes between two consecutive frames. A time window (10 frames) is extracted, and the cell voltage values ​​in this time window are used as the object. The voltage of all cells is de-trended using the median. The variance of each cell in this time window is calculated to form a volatility evaluation standard. Cells exceeding 3 times the standard deviation are screened out and frequency statistics are performed to assess the health status of the battery.

[0104] Furthermore, S406 specifically includes:

[0105] S4061: Calculate the average value of the individual battery cell voltage for each time frame, and use the average value as the reference voltage value to obtain the reference voltage curve of the charging segment.

[0106] S4062: Calculate the average value of each column of the voltage matrix to determine the average voltage of the individual battery cells in the charging segment.

[0107] S4063: Calculate the voltage variance of the battery cells based on the average voltage of the battery cells, and use the voltage variance of the battery cells as the first outlier characterization parameter of the charging segment.

[0108] S4064: Calculate the voltage difference between the voltage curve of each battery cell and the reference voltage curve, and sum all voltage differences; the battery cell curve is formed by connecting the voltage values ​​obtained for each battery cell with a smooth curve.

[0109] S4065: Divide the summed pressure difference by the time frame to calculate the average value.

[0110] S4066: Take the maximum value among the average values ​​as the second outlier characterization parameter of the charging segment.

[0111] First, outlier warning is performed on the characteristic parameters of the charging segment. The average value of each row in the obtained voltage matrix is ​​calculated, and the average value of all individual voltages in each frame is used as the reference voltage value to obtain the reference voltage curve for the charging segment. The average value of the individual voltages in each column is calculated to obtain the average value of the individual voltages in the charging segment. The variance of the average values ​​of all individual voltages is calculated as the first outlier characteristic parameter of the charging segment. The voltage difference between each individual voltage curve and the reference voltage curve is calculated, the absolute values ​​are summed, and then averaged over time frames to characterize the degree to which each individual deviates from the reference voltage curve within the charging segment. The maximum value is taken as the second outlier characteristic parameter of the charging segment. The first and second outlier characteristic parameters are taken as the outlier characteristic parameters of the charging segment.

[0112] Furthermore, S407 specifically includes:

[0113] S4071: Use PCA principal component analysis to determine the distribution relationship between the first outlier characterization parameter and the second outlier characterization parameter.

[0114] S4072: Based on the distribution relationship, determine the degree of abnormality of the charging segment according to the Mahalanobis distance, and screen abnormal charging segments.

[0115] Since battery failures are often caused by an abnormality in a single cell, and often manifest as low voltage, high voltage, or abnormal fluctuations, a second outlier parameter can be used to reflect the presence of abnormal cells in a charging segment. However, due to variations in vehicle operating conditions and external environments, the second outlier parameter fluctuates significantly and cannot effectively reflect abnormal situations. Therefore, a first outlier parameter is selected as a reference to reflect the average voltage performance of all cells within a charging segment under different operating conditions and external conditions. Based on the two outlier parameters reflecting the charging segment, Principal Component Analysis (PCA) is used to obtain the influence relationship between the two outlier parameters. Mahalanobis distance is used to calculate the degree of abnormality, and abnormal charging segments are screened. Figure 3 As shown.

[0116] Furthermore, S409 specifically includes:

[0117] S4091: Based on the selected abnormal charging segments, shift the battery cell voltage curve by a first specified number of frames, delete the null values ​​generated during the shift of the battery cell voltage curve and the voltage values ​​of the reference voltage curve in the same time frame, and determine the similarity between the shifted battery cell voltage curve and the processed reference voltage curve.

[0118] S4092: Shift the voltage curve of the battery cell by a second specified number of frames, and repeat the similarity calculation until the number of shifted frames reaches a preset number of shifted frames, and determine the similarity corresponding to the preset number of shifted frames.

[0119] S4093: Obtain the offset corresponding to the maximum similarity between the voltage curve of each battery cell and the reference voltage curve, compare the offset corresponding to the maximum similarity with a preset threshold, and identify the battery cells corresponding to the offset corresponding to the maximum similarity that exceeds the preset threshold as battery cells with outlier anomalous degree, and determine the third anomalous battery cell.

[0120] Secondly, outlier warnings based on changing trends are used to screen for inconsistencies in the state of charge (SOC) of individual cells within the battery pack. The process is as follows: Figure 4 As shown. Since the time intervals of the acquired individual voltage sequences are not fixed, directly shifting the voltage sequences will cause significant changes in the voltage charging curve. Therefore, it is necessary to fit the charging curve using individual voltage sequences with variable time intervals, and then take voltage data at the same time interval. The average voltage value is obtained by averaging the individual voltages of each frame of data, resulting in a reference voltage sequence for the charging segment. Each individual voltage sequence is shifted for a specified number of frames (the shift step size is variable to adapt to voltage sequences with different time intervals), resulting in the shifted individual voltage sequence. Null values ​​generated during the shift process are deleted, and data from the same time frame of the reference voltage sequence are also deleted to maintain the same sequence length. The similarity between the shifted individual voltage sequence and the processed reference voltage sequence is calculated, such as Pearson similarity, cosine similarity, etc. The individual is then shifted for a certain number of frames, and the similarity calculation is repeated until the number of shifted frames reaches the preset value. Based on the shift amount with the largest similarity between each individual voltage sequence and the reference voltage sequence within the charging segment, a preset threshold is compared, and individuals with abnormal SOC are selected for alarm.

[0121] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 5 As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database is used for power battery safety warnings. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements power battery safety warnings.

[0122] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0123] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0124] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0125] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0126] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0127] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0128] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0129] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for early warning of power battery safety, characterized in that, The power battery safety early warning method includes: Based on the vehicle and timestamp, the real-time operation data of the vehicle is processed to determine the processed data; The charging status of the power battery is obtained, and the processed data is segmented according to the charging status of the power battery to determine the vehicle's charging data. The charging data is then divided into multiple charging segments according to time. Construct the voltage matrix of the charging segment; each row of the voltage matrix represents a time frame, and each column represents a single battery cell; the power battery is composed of multiple battery cells; Based on the voltage matrix, a battery cell early warning is performed to obtain threshold early warning results, change early warning results, and outlier results; Safety warnings for power batteries are issued based on the threshold warning results, the change warning results, and the outlier results.

2. The power battery safety early warning method according to claim 1, characterized in that, The charging status of the power battery is obtained, and the processed data is segmented according to the charging status of the power battery to determine the vehicle's charging data. The charging data is then divided into multiple charging segments based on time, specifically including: The charging status of the power battery is obtained based on parameters such as vehicle status indicator, current and vehicle speed. Based on the charging status of the power battery, the processed data is segmented to determine the vehicle's charging data; The charging data is divided into multiple charging segments according to the time sequence.

3. The power battery safety early warning method according to claim 1, characterized in that, Based on the voltage matrix of the charging segment, a battery cell early warning is performed, yielding threshold early warning results, change early warning results, and outlier results, specifically including: The first abnormal battery cell in the voltage matrix is ​​selected. Among the selected first abnormal battery cells, the first abnormal battery cells that exceed the overvoltage and undervoltage thresholds are given a threshold warning, and the threshold warning result is determined. The first abnormal battery cell is an overvoltage battery cell or an undervoltage battery cell. Obtain the voltage values ​​of all individual cells in the voltage matrix within a set time window, and calculate the variance of the voltage value of each individual cell within the set time window; The volatility evaluation criteria are determined based on the variance of the voltage value of each individual unit within a set time window. Based on the volatility evaluation criteria, the second abnormal battery cell that exceeds the standard deviation threshold is selected. Based on the selected second abnormal battery cells, a change rate warning is issued, and the change rate warning result is determined. Based on the voltage matrix, the outlier characterization parameters of the charging segment are determined; the outlier characterization parameters include a first outlier characterization parameter and a second outlier characterization parameter. The relationship between the first outlier characterization parameter and the second outlier characterization parameter is determined by principal component analysis (PCA) to screen abnormal charging segments. Calculate the offset of the largest similarity between the voltage curve of the battery cell corresponding to all the screened abnormal charging segments and the reference voltage curve; The maximum offset is compared with the offset threshold to filter out the third abnormal battery cell that exceeds the offset threshold; Outlier warnings are issued based on the selected third abnormal battery cells, and the outlier warning results are determined.

4. The power battery safety early warning method according to claim 3, characterized in that, The first abnormal battery cell in the voltage matrix is ​​selected. Among the selected first abnormal battery cells, threshold warnings are issued for those exceeding the overvoltage and undervoltage thresholds. The threshold warning results are determined, specifically including: Based on the voltage matrix, the first abnormal battery cell and the time frame corresponding to the first abnormal battery cell are selected by comparing the voltage value of each cell in each time frame with the overvoltage and undervoltage thresholds. The first abnormal battery cell that exceeds the overvoltage and undervoltage thresholds is given a graded warning, and the threshold warning result is determined.

5. The power battery safety early warning method according to claim 3, characterized in that, Based on the selected second abnormal battery cells, a rate of change warning is issued, and the result of the rate of change warning is determined, specifically including: Calculate the voltage change amplitude of the second abnormal battery cell in two time frames before and after, and filter out the second abnormal battery cells whose voltage change amplitude in the two frames exceeds the voltage change amplitude threshold and the abnormal time frames corresponding to the second abnormal battery cells. A time window is selected based on the second abnormal battery cell and the abnormal time frame corresponding to the second abnormal battery cell. Within the time window, the median is used to detrend the voltage of the selected second abnormal battery cells, and the variance of the second abnormal battery cells after detrending is determined. Based on the variance of the second abnormal battery cell, filter out the second abnormal battery cells that exceed the variance threshold, perform a rate of change warning, and determine the rate of change warning result.

6. The power battery safety early warning method according to claim 3, characterized in that, Based on the voltage matrix, the outlier characterization parameters of the charging segment are determined, specifically including: The average value of the individual battery cell voltage in each time frame is calculated, and the average value is used as the reference voltage value to obtain the reference voltage curve of the charging segment. Calculate the average value of each column of the voltage matrix to determine the average voltage of the individual battery cells in the charging segment; Based on the average voltage of the individual battery cells, the voltage variance of the individual battery cells is calculated, and the voltage variance of the individual battery cells is used as the first outlier characterization parameter of the charging segment. Calculate the voltage difference between the voltage curve of each battery cell and the reference voltage curve, and sum all voltage differences; the battery cell curve is formed by connecting the voltage values ​​obtained for each battery cell with a smooth curve. Divide the summed pressure difference by the time frame to obtain the average value; The maximum value among the average values ​​is taken as the second outlier characterization parameter of the charging segment.

7. The power battery safety early warning method according to claim 3, characterized in that, The relationship between the first outlier characteristic parameter and the second outlier characteristic parameter is determined using principal component analysis (PCA) to screen for abnormal charging segments, specifically including: The distribution relationship between the first outlier characterization parameter and the second outlier characterization parameter was determined using principal component analysis (PCA). Based on the aforementioned distribution relationship, the degree of abnormality of the charging segment is determined according to the Mahalanobis distance, and abnormal charging segments are then screened.

8. The power battery safety early warning method according to claim 3, characterized in that, The maximum offset is compared with an offset threshold to filter out third abnormal battery cells that exceed the offset threshold, specifically including: Based on the selected abnormal charging segments, the voltage curve of a single battery cell is shifted by a first specified number of frames. Null values ​​generated during the shift of the voltage curve of the single battery cell and voltage values ​​of the reference voltage curve in the same time frame are deleted. The similarity between the shifted voltage curve of the single battery cell and the processed reference voltage curve is then determined. The voltage curve of the individual battery cell is shifted by a second specified number of frames. The similarity is calculated repeatedly until the number of shifted frames reaches a preset number of shifted frames. The similarity corresponding to the preset number of shifted frames is then determined. Obtain the offset corresponding to the maximum similarity between the voltage curve of each battery cell and the reference voltage curve. Compare the offset corresponding to the maximum similarity with a preset threshold. The battery cells corresponding to the offset corresponding to the maximum similarity exceeding the preset threshold are identified as battery cells with outlier anomalous degree, and the third anomalous battery cells are determined.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the power battery safety warning method according to any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the power battery safety early warning method according to any one of claims 1-8.

Images