Battery safety detection method, device, apparatus, and medium

Abstract

The application discloses a battery safety detection method, device, equipment and medium, and the method comprises the following steps: calculating the historical feature value of each candidate detection feature corresponding to each reference battery according to the historical working condition data of a plurality of reference batteries; the type of the reference battery is the same as that of the battery to be detected; determining the detection judgment parameter corresponding to the candidate detection feature based on the historical feature value of the candidate detection feature corresponding to all reference batteries, wherein the detection judgment parameter comprises an optimal error detection balance factor and an optimal detection threshold; selecting the target detection feature of the battery to be detected from all candidate detection features according to the optimal error detection balance factor corresponding to each candidate detection feature; determining the real-time feature value of the target detection feature based on the real-time working condition data of the battery to be detected, and determining the safety detection result of the battery to be detected according to the real-time feature value of the target detection feature and the optimal detection threshold. Through the technical scheme of the embodiment of the application, the accuracy and reliability of detection are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of battery technology, and in particular to battery safety testing methods, devices, equipment and media. Background Technology

[0002] With the rapid development of new energy technologies, lithium batteries, thanks to their high energy density, long cycle life, and wide applicability, have been widely used in new energy vehicles, energy storage power stations, and portable electronic devices. The tabs, as the metal conductors connecting the internal electrodes of the cell to the external terminals, directly determine the battery's internal resistance, rate performance, and safety reliability. However, under conditions such as manufacturing, packaging, extrusion, charge / discharge vibration, and thermal cycling, tabs are highly susceptible to damage such as micro-tears, localized cracks, or the propagation of latent cracks. These defects are among the main safety hazards of batteries. Therefore, achieving high-precision, rapid online detection of tab tears is a core key to ensuring the safe operation of batteries throughout their entire lifecycle.

[0003] Currently, electrode tear detection technologies are mainly divided into two categories: physical non-destructive testing and operating parameter threshold testing. Physical non-destructive testing methods, such as X-rays, industrial computed tomography (CT), and ultrasound, can directly identify electrode structural defects; however, these methods are costly and inefficient, limiting their application to offline sampling and laboratory analysis, and failing to meet the needs of online real-time battery monitoring. Threshold testing methods based on operating parameters rely on characteristics such as voltage fluctuations, internal resistance changes, and localized temperature rises combined with fixed thresholds to determine faults. However, the selection of detection features and the setting of alarm thresholds largely depend on human experience, resulting in strong subjectivity and a lack of quantitative judgment criteria. Furthermore, these methods are susceptible to individual differences in battery cells and fluctuations in operating conditions, leading to a high probability of false positives and false negatives, and insufficient overall detection accuracy and reliability.

[0004] Therefore, there is an urgent need to propose a new method to solve the above problems. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and medium for battery safety testing, which can perform online real-time safety testing of batteries, significantly improving the accuracy and reliability of the testing.

[0006] In a first aspect, embodiments of the present invention provide a battery safety detection method, the method comprising: Based on historical operating data of multiple reference batteries, historical feature values ​​of candidate detection features corresponding to each reference battery are calculated; the type of the reference battery is the same as the type of the battery to be detected. Based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries, the detection judgment parameters of the corresponding candidate detection features are determined. The detection judgment parameters include the optimal error detection balance factor and the optimal detection threshold. Based on the optimal error detection balance factor corresponding to each candidate detection feature, the target detection feature of the battery to be detected is selected from all candidate detection features. The real-time feature value of the target detection feature is determined based on the real-time operating condition data of the battery under test, and the safety detection result of the battery under test is determined based on the real-time feature value of the target detection feature and the optimal detection threshold.

[0007] The technical solution of this invention first calculates the historical feature values ​​of candidate detection features for each benchmark battery based on historical operating condition data of multiple benchmark batteries, laying a data foundation for determining the detection judgment parameters of each candidate detection feature. Then, based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries, the detection judgment parameters for the corresponding candidate detection features are determined, resulting in exclusive detection judgment parameters for each candidate detection feature. This provides a quantitative basis for the selection of target detection features and ensures the objectivity and reliability of the selection results. Next, based on the optimal error detection balance factor corresponding to each candidate detection feature, the target detection features for the battery to be tested are selected from all candidate detection features. This avoids subjective bias caused by relying on manual experience to select features, and allows for a direct comparison of the fault discrimination capabilities of each candidate feature, thereby quickly selecting features with excellent identification performance and effectively improving the accuracy and reliability of battery safety detection. Finally, based on the real-time operating condition data of the battery to be tested, the real-time feature values ​​of the target detection features are determined, and based on the real-time feature values ​​of the target detection features and the optimal detection threshold, the safety detection result of the battery to be tested is determined, which can improve the accuracy, reliability, and real-time performance of the battery tab tear-specific safety hazard detection. Therefore, the technical solution of the present invention can overcome the shortcomings of the prior art in that it is difficult to simultaneously meet the requirements of online real-time battery detection and the requirements of detection accuracy and reliability.

[0008] Secondly, embodiments of the present invention also provide a battery safety detection device, the device comprising: The calculation module is used to calculate the historical feature value of the candidate detection feature corresponding to each reference battery based on the historical operating condition data of multiple reference batteries; the type of the reference battery is the same as the type of the battery to be detected. The determination module is used to determine the detection judgment parameters of the corresponding candidate detection features based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries. The detection judgment parameters include the optimal error detection balance factor and the optimal detection threshold. The filtering module is used to filter out the target detection features of the battery to be detected from all candidate detection features based on the optimal error detection balance factor corresponding to each candidate detection feature. The detection module is used to determine the real-time feature value of the target detection feature based on the real-time operating condition data of the battery under test, and to determine the safety detection result of the battery under test based on the real-time feature value of the target detection feature and the optimal detection threshold.

[0009] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the battery safety detection method according to any embodiment of the present invention.

[0010] Fourthly, embodiments of the present invention also provide a storage medium containing computer-executable instructions, characterized in that the computer-executable instructions, when executed by a computer processor, implement the battery safety detection method described in any embodiment of the present invention.

[0011] It should be noted that the aforementioned computer instructions may be stored, in whole or in part, on a computer-readable storage medium. This computer-readable storage medium may be packaged together with the processor of the battery safety detection device, or it may be packaged separately from the processor of the battery safety detection device; this application does not impose any limitations on this.

[0012] The descriptions of the second, third, and fourth aspects in this application can be referenced to the detailed description of the first aspect; and the beneficial effects described in the second, third, and fourth aspects can be referenced to the analysis of the beneficial effects of the first aspect, which will not be repeated here.

[0013] In this application, the name of the aforementioned battery safety detection device does not limit the device or functional module itself. In actual implementation, these devices or functional modules may appear under other names. As long as the function of each device or functional module is similar to that of this application, it falls within the scope of the claims of this application and its equivalents.

[0014] These or other aspects of this application will become more readily apparent in the following description. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic flowchart of a battery safety testing method provided in an embodiment of the present invention; Figure 2a A schematic flowchart of another battery safety detection method provided in an embodiment of the present invention; Figure 2b This is a schematic diagram showing the comparison between the distribution of candidate detection feature samples and the corresponding optimal detection threshold provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the structure of a battery safety detection device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0018] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0019] The terms "first" and "second," etc., used in the specification and drawings of this application are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0020] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0021] Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the figures. The process can correspond to a method, function, procedure, subroutine, subroutine, etc. Moreover, without conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0022] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0023] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0024] Figure 1 This is a flowchart illustrating a battery safety detection method according to an embodiment of the present invention. This embodiment is applicable to situations requiring real-time battery safety detection. The method can be executed by a battery safety detection device, which can be implemented in software and / or hardware. For example, the device can be an electronic device. (Reference) Figure 1 The battery safety detection method in this embodiment specifically includes the following steps: Step 110: Based on the historical operating data of multiple reference batteries, calculate the historical feature value of the candidate detection feature corresponding to each reference battery.

[0025] Specifically, the reference battery is a standard reference battery sample of the same type as the battery under test (e.g., identical battery model, cell material, rated capacity, and external specifications). This reference battery possesses long-term, complete operating condition data, clearly defined operating status, and includes both fault-free and faulty samples. The battery under test is the target battery object requiring real-time safety testing. Operating condition data is a set of time-series operating parameters synchronously collected throughout the battery's entire lifecycle, such as voltage, current, temperature, capacity, charge / discharge duration, internal resistance, differential voltage, ambient temperature, and cycle count. It is divided into historical operating condition data and real-time operating condition data. Historical operating condition data is reference data stored in time sequence, recording the complete historical operating process of the battery, fully covering all operating parameters under past charging, discharging, and static storage conditions. Real-time operating condition data is dynamic data collected in real-time during the current operation of the battery under test. Candidate detection features are quantitative indicators derived from the original operating condition data that can distinguish between normal battery conditions and faulty conditions. They are alternative feature quantities for battery safety detection, such as time-domain statistical features (e.g., single-cell voltage range, voltage rebound, DC internal resistance increment, temperature rise rate, capacity decay rate, open-circuit voltage offset, etc.), curve morphology features (e.g., quantitative characterization parameters related to charge-discharge curves), and electrochemical model features (e.g., ohmic internal resistance, electrochemical polarization internal resistance, concentration polarization internal resistance, electrochemical polarization capacitance, and concentration polarization capacitance obtained from equivalent circuit fitting, etc.). In this embodiment, candidate detection features are used to identify battery tab tearing faults. Historical feature values ​​refer to the historical values ​​of candidate detection features obtained by extracting a segment or the entire cycle of historical operating condition data for a single reference battery and calculating them using the feature calculation formula. For each candidate feature of each reference battery, a corresponding historical feature value is generated using a segment or the entire cycle of its historical operating condition data.

[0026] In the specific implementation, firstly, historical operating condition data of multiple benchmark batteries are extracted from the database storing battery operating condition data; then, according to the calculation requirements of each candidate detection feature, operating condition segments are extracted from the historical operating condition data; then, through a double-layer traversal (the outer layer traverses each benchmark battery, and the inner layer traverses each candidate detection feature), the extracted operating condition data are substituted into the corresponding feature formulas for sequential calculation, thereby obtaining the historical feature values ​​of each benchmark battery under each candidate detection feature.

[0027] For example, when the candidate detection feature is the average value of DC internal resistance increment, its calculation formula is: the sum of the DC internal resistance increments at each sampling time within the preset calculation period divided by the total number of sampling points in that period; where the calculation expression for a single DC internal resistance increment is: (static voltage - instantaneous discharge voltage) / discharge current - initial internal resistance.

[0028] For example, when the candidate detection feature is the individual cell voltage range, its calculation formula is: the difference between the maximum and minimum values ​​of the individual cell voltage within a preset calculation period.

[0029] In this embodiment, the above steps lay the data foundation for subsequently determining the detection and judgment parameters of each candidate detection feature.

[0030] Step 120: Based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries, determine the detection judgment parameters of the corresponding candidate detection features.

[0031] Specifically, the detection and judgment parameters include the optimal error detection balance factor and the optimal detection threshold. The detection and judgment parameters are a set of core parameters calculated through statistical analysis and optimization, based on the historical feature value distribution of all benchmark batteries (including normal and abnormal samples). The error detection balance factor is a quantitative evaluation index calculated based on the number of detected samples and the number of overkill samples in the benchmark battery dataset for a specific candidate threshold under a single candidate detection feature. It is used to uniformly measure the comprehensive discrimination performance of the threshold between fault detection and false alarm errors, such as the Youden index, F1 score, and weighted comprehensive error. The optimal error detection balance factor is the globally optimal value obtained by traversing all candidate thresholds within the threshold search interval for a single candidate detection feature, calculating the error detection balance factor for each threshold, and then selecting the optimal value based on a preset optimization criterion (such as maximizing the index value / minimizing the comprehensive error). The optimal detection threshold is the detection threshold corresponding to the optimal error detection balance factor, calculated by traversing all candidate thresholds for a single candidate detection feature and calculating the corresponding error detection balance factor for each. Each candidate detection feature corresponds to an independent set of optimal detection thresholds and optimal error detection balance factors.

[0032] In the specific implementation, for each candidate detection feature, all historical feature values ​​of all benchmark batteries (including normal and abnormal samples) under that feature dimension are collected to form a complete feature value distribution set for that candidate detection feature. Based on this feature value distribution set, a threshold search interval is determined, and all candidate thresholds are traversed within the search interval according to a preset fixed step size.

[0033] For each candidate threshold during the traversal process, the historical feature values ​​corresponding to all reference batteries are compared with the current candidate threshold to statistically determine the number of fault detection samples of abnormal reference batteries and the number of over-detection and false alarm samples of normal reference batteries. Based on the statistically obtained sample quantities of the two types, the error detection balance factor corresponding to the current candidate threshold is calculated through a preset quantitative evaluation method. The error detection balance factor can be any evaluation index that can measure the balance between fault detection and false alarms, such as the Youden index, F1 score, or weighted comprehensive error.

[0034] After completing the calculation of all candidate thresholds, the error detection balance factor corresponding to each candidate threshold is obtained. All error detection balance factors are then screened according to preset optimization criteria (such as maximizing the error detection balance index value or minimizing the overall error), and the globally optimal index value is selected as the optimal error detection balance factor for that candidate detection feature. Simultaneously, the candidate threshold corresponding to this optimal error detection balance factor is determined as the optimal detection threshold for that candidate detection feature.

[0035] Finally, the optimal error detection balance factor is paired with the optimal detection threshold to form the detection decision parameters corresponding to the candidate detection feature. Each candidate detection feature independently undergoes the above threshold optimization process, generating unique and independent detection decision parameters for each.

[0036] In this embodiment, the above steps can obtain the exclusive detection judgment parameters corresponding to each candidate detection feature, which not only provides a quantitative basis for the screening of target detection features, but also ensures the objectivity and reliability of the screening results.

[0037] Step 130: Based on the optimal error detection balance factor corresponding to each candidate detection feature, select the target detection feature of the battery to be detected from all candidate detection features.

[0038] Specifically, the target detection features are the optimal subset of features selected and retained from all candidate detection features and ultimately used for real-time safety judgment of the battery to be detected.

[0039] In practice, all candidate detection features are sorted from highest to lowest based on their respective optimal error detection balance factor values. Then, according to a pre-defined filtering strategy, the target detection features to be retained are determined from the sorted feature sequence. For example, the filtering strategy could be: selecting only the top-ranked feature as a single target detection feature, or selecting the top N features. Finally, the selected single feature or combination of features is used as the target detection feature.

[0040] In this embodiment, the above steps can avoid subjective bias caused by relying on human experience to screen features, and intuitively compare the ability of each candidate feature to distinguish faults, thereby quickly screening out features with excellent identification performance and effectively improving the accuracy and reliability of battery safety detection.

[0041] Step 140: Determine the real-time feature value of the target detection feature based on the real-time operating condition data of the battery to be tested, and determine the safety detection result of the battery to be tested based on the real-time feature value of the target detection feature and the optimal detection threshold.

[0042] Specifically, the real-time feature value is the feature value obtained by applying the feature calculation formula corresponding to the target detection feature to the real-time operating condition data currently collected by the battery under test, and then performing quantization calculation. The safety detection result is the final judgment conclusion output after logically comparing the real-time feature value with the optimal detection threshold.

[0043] In the specific implementation, real-time operating condition data of the current operating status of the battery under test is obtained, and the feature calculation formulas corresponding to the selected target detection features are retrieved. The real-time operating condition data is substituted into each calculation formula for quantitative solution to obtain the real-time feature value corresponding to each target detection feature. Then, the optimal detection threshold of each target detection feature obtained through optimization of benchmark battery samples in the offline stage is retrieved, and each set of real-time feature values ​​is compared with the matched optimal detection threshold in turn. If any real-time feature value exceeds the corresponding optimal detection threshold, the safety detection result of the battery under test is determined to be abnormal (here, abnormality specifically refers to tab tearing abnormality), and it is marked as a potential tab tearing battery; if all real-time feature values ​​do not exceed the corresponding optimal detection threshold, the safety detection result of the battery under test is determined to be that the battery under test is in normal condition. Among them, the abnormal benchmark batteries used in the offline optimization stage are all battery samples with tab tearing faults.

[0044] In this embodiment, an online real-time judgment mechanism is constructed based on high-performance target detection features selected in the offline phase and optimized through benchmark sample data to obtain the optimal detection threshold. During online operation, standardized feature calculation formulas are used to quantify real-time operating condition data to obtain real-time feature values, which are then logically compared with the corresponding optimal threshold to output the safety detection result. The overall operation logic is simple and efficient, with low computational overhead, and is easy to deploy in a lightweight manner, fully ensuring the real-time nature of monitoring. Simultaneously, since both the target detection features and the optimal threshold have undergone rigorous optimization and performance verification in the offline phase, the fault identification sensitivity is high, effectively reducing the probability of missed detections. Furthermore, this judgment standard is objective and stable, effectively balancing the conflict between fault detection and false alarms, significantly improving the detection accuracy and engineering reliability for specific safety hazards such as battery tab tearing.

[0045] The battery safety detection method provided in this invention first calculates the historical feature values ​​of candidate detection features for each benchmark battery based on historical operating condition data of multiple benchmark batteries, laying a data foundation for determining the detection judgment parameters of each candidate detection feature. Then, based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries, the detection judgment parameters for the corresponding candidate detection features are determined, resulting in exclusive detection judgment parameters for each candidate detection feature. This provides a quantitative basis for selecting target detection features and ensures the objectivity and reliability of the selection results. Next, based on the optimal error detection balance factor corresponding to each candidate detection feature, the target detection features for the battery to be tested are selected from all candidate detection features. This avoids subjective bias caused by relying on manual experience to select features, and allows for a direct comparison of the fault discrimination capabilities of each candidate feature, thereby quickly selecting features with excellent identification performance and effectively improving the accuracy and reliability of battery safety detection. Finally, based on the real-time operating condition data of the battery to be tested, the real-time feature values ​​of the target detection features are determined, and based on the real-time feature values ​​of the target detection features and the optimal detection threshold, the safety detection result of the battery to be tested is determined. This improves the accuracy, reliability, and real-time performance of the battery tab tear-specific safety hazard detection. Therefore, the technical solution of the present invention can overcome the shortcomings of the prior art in that it is difficult to simultaneously meet the requirements of online real-time battery detection and the requirements of detection accuracy and reliability.

[0046] Figure 2a This is a flowchart illustrating another battery safety detection method provided by an embodiment of the present invention. This embodiment is a specific implementation based on the above embodiment. In this embodiment, the method may further include: Step 210: Based on the historical operating condition data of multiple reference batteries, calculate the historical feature value of the candidate detection feature corresponding to each reference battery.

[0047] Optionally, candidate detection features include curve representation features, and historical operating condition data includes historical dynamic operating condition data.

[0048] Further, step 210 may specifically include: for the current reference battery, constructing a voltage increment sequence for the current reference battery based on the historical voltage data in the historical dynamic operating condition data of the current reference battery; constructing a capacity sequence for the current reference battery based on the historical capacity data corresponding to the historical voltage data in the historical dynamic operating condition data of the current reference battery; generating an incremental capacity curve for the current reference battery based on the voltage increment sequence and the capacity sequence; and performing quantitative feature extraction on the incremental capacity curve to obtain historical feature values ​​of the curve characterization features of the current reference battery.

[0049] Specifically, historical dynamic operating condition data refers to time-series operating condition data continuously collected during the dynamic operation of battery charging and discharging, distinct from static data, and includes synchronously recorded historical voltage and capacity data. The current reference battery is a standard sample battery used for offline modeling and threshold optimization, categorized into normal reference batteries and abnormal reference batteries with tab tearing faults; here, it specifically refers to the single reference battery sample selected for a single computation. The voltage increment sequence is an ordered voltage sequence formed by sorting all historical voltage data of the current reference battery in ascending order. The capacity sequence is an ordered sequence formed by synchronously arranging the measured capacity values ​​corresponding to each voltage sampling point, based on the arrangement order of the voltage increment sequence. The incremental capacity curve is a feature curve plotted with the monotonically increasing voltage as the horizontal axis and the ratio of the filtered capacity difference to the voltage difference as the vertical axis. Quantitative feature extraction involves numerically calculating the morphological parameters of the incremental capacity curve, such as peak value, valley value, peak area, peak position, and curve slope, transforming graphical information into calculable and horizontally comparable quantitative indicators. The historical characteristic value of the curve characterization feature is a quantified value extracted from the incremental capacity curve of a single reference battery. It serves as the sample value of the curve characterization feature corresponding to the reference battery and is used for subsequent offline solution of the optimal detection threshold and optimal detection balance factor.

[0050] In practice, for the current reference battery, historical voltage data synchronized with time is extracted from the historical dynamic operating condition data of the current reference battery based on a preset fixed time window (such as the past day, the past month, etc.). The obtained historical voltage values ​​are sorted in ascending order to construct the voltage increment sequence corresponding to the reference battery. Specifically, if the extracted data belongs to the discharge step, the voltage sequence is inverted to make it monotonically increasing; if it belongs to the charging step, the voltage increment sequence is obtained directly by sorting.

[0051] Next, based on the historical capacity data corresponding to the historical voltage data in the historical dynamic operating condition data of the current reference battery, the capacity sequences are synchronously matched and arranged according to the correspondence of the voltage increment sequence to construct a capacity sequence that strictly corresponds one-to-one with the voltage. Then, a voltage difference sequence is determined based on the voltage increment sequence, which is obtained by subtracting the previous value from the next value in the voltage increment sequence; simultaneously, a capacity difference sequence is determined based on the capacity sequence that corresponds one-to-one with the voltage, which is obtained by subtracting the previous value from the next value in the capacity sequence. Then, the voltage difference sequence and the capacity difference sequence are filtered (e.g., Gaussian filtering, Savitzky-Gore filtering, etc.) to obtain the filtered voltage difference sequence and capacity difference sequence. Finally, the filtered capacity difference sequence is divided by the corresponding voltage difference sequence. Using the normalized, monotonically increasing voltage as the horizontal axis and the division result as the vertical axis, the incremental capacity curve of the current reference battery is plotted. Then, morphological parameters such as peak / valley positions, heights, and areas are extracted from the incremental capacity curve as historical feature values ​​representing the curve characteristics of the current reference battery.

[0052] In this embodiment, by introducing an incremental capacity curve analysis method, the electrochemical state information implicit in the dynamic charging and discharging conditions of the battery is quantified into calculable and horizontally comparable curve characterization features. These features exhibit specific response sensitivity to tab tearing faults, enabling early warnings. They also possess good noise resistance and batch automated processing capabilities, effectively complementing other candidate features (such as time-domain statistical features and electrochemical model features). This provides a differentiated, high-quality candidate feature source for subsequent screening and comprehensive identification of the target detection features with optimal performance.

[0053] Optionally, candidate detection features may also include equivalent circuit features, ohmic internal resistance, first resistor, second resistor, first capacitor, and second capacitor, and historical operating condition data may also include historical static operating condition data.

[0054] Further, step 210 may specifically include: for the current reference battery, determining the historical characteristic values ​​of the equivalent circuit features of the current reference battery based on the historical voltage data and equivalent circuit model in the historical static operating condition data; the equivalent circuit features include a first voltage, a first time constant, a second voltage, and a second time constant; calculating the historical characteristic value of the ohmic internal resistance of the current reference battery based on the historical voltage data and historical current data in the historical operating condition data; determining the historical characteristic value of the first resistance of the current reference battery based on the historical current data and the historical characteristic value of the first voltage in the historical dynamic operating condition data, and determining the historical characteristic value of the second resistance of the current reference battery based on the historical current data and the historical characteristic value of the second voltage in the historical dynamic operating condition data; calculating the ratio of the historical characteristic value of the first time constant to the historical characteristic value of the first resistance to obtain the historical characteristic value of the first capacitance of the current reference battery, and calculating the ratio of the historical characteristic value of the second time constant to the historical characteristic value of the second resistance to obtain the historical characteristic value of the second capacitance of the current reference battery.

[0055] Specifically, the equivalent circuit features are a set of electrochemical parameters obtained by fitting historical operating condition data based on a second-order equivalent circuit model of the battery. These parameters include a first voltage, a first time constant, a second voltage, and a second time constant. The historical static operating condition data is a set of time-series parameters collected from a reference battery during the open-circuit relaxation phase with the charge / discharge circuit disconnected and no load current. The equivalent circuit model is a mathematical model that abstracts the complex electrochemical processes inside the battery into a combination of various electronic components. In this embodiment, the equivalent circuit model can be a second-order equivalent circuit model. The first voltage is the polarization voltage on the first set of parallel resistor-capacitor branches, generated by the electrochemical reaction at the electrode interface. Its voltage relaxation decay rate is fast, and it is also called the electrochemical polarization voltage. The second voltage is the polarization voltage on the second set of parallel resistor-capacitor branches, generated by the lithium-ion diffusion concentration difference. Its voltage relaxation decay rate is slow, and it is also called the concentration polarization voltage.

[0056] The first time constant characterizes the relaxation decay rate of the electrochemical polarization voltage. A smaller value indicates a faster recovery rate of the polarization voltage; it is also called the electrochemical polarization time constant. The second time constant characterizes the relaxation decay rate of the concentration polarization voltage. A larger value indicates a slower recovery rate of the polarization voltage; it is also called the concentration polarization time constant. The ohmic internal resistance is a purely resistive impedance formed by the electrode current collector, electrolyte, diaphragm, and tab contact area. The first resistance corresponds to the polarization internal resistance during the electrochemical polarization process; it is also the resistive element in the first set of parallel resistor-capacitor branches, characterizing the resistance to charge transfer reactions at the electrode interface. The second resistance corresponds to the polarization internal resistance during the concentration polarization process; it is also the resistive element in the second set of parallel resistor-capacitor branches, characterizing the diffusion resistance of lithium ions inside the electrode and the diffusion resistance caused by the electrolyte concentration gradient. The first capacitor is a capacitor element that equivalently simulates the charge storage capacity of the double layer at the electrode / electrolyte interface, used to characterize the charge buffering characteristics of the electrochemical polarization branch; it is also called the electrochemical polarization capacitor. The second capacitor is a capacitor element that is equivalent to the charge buffering and energy storage characteristics corresponding to the lithium ion diffusion inside the solid phase of the electrode and the concentration gradient of the electrolyte liquid phase. It is used to characterize the charge storage capacity of the concentration polarization branch and is also called the concentration polarization capacitor.

[0057] In practice, for the current reference battery, the historical characteristic values ​​corresponding to the equivalent circuit characteristics of the current reference battery are determined based on the historical voltage data in the historical static operating condition data and the equivalent circuit model. Specifically, the static relaxation step data (i.e., the voltage recovery time series curve collected after the load current changes abruptly from a non-zero state to zero) within a preset time window can be selected according to the actual situation or requirements. The equivalent circuit model is used to fit this voltage relaxation process, and the equivalent circuit model can be expressed by the following formula: Where U(t) is the measured voltage across the battery at time t, which is a time-series voltage observation value that varies with the resting time and is taken from historical static operating condition data; OCV is the open circuit voltage; U1 is the first voltage (electrochemical polarization voltage); and U2 is the second voltage (concentration polarization voltage). This is the first time constant (electrochemical polarization time constant); This is the second time constant (concentration polarization time constant).

[0058] Subsequently, based on historical voltage and current data from historical operating condition data, the historical characteristic value of the ohmic internal resistance of the current reference battery is calculated. Specifically, within a preset time window, step switching events are identified from the historical operating condition data of the current reference battery. The preceding step is a constant current charge / discharge step, and the following step is a resting relaxation step. The load current Iend and the instantaneous terminal voltage Uend at the end of the charge / discharge step are obtained. Then, the steady-state terminal voltage U0 after the initial current returns to zero in the resting relaxation step is extracted (the average value of multiple zero-current sampling points after the switch is taken). The historical characteristic value of the ohmic internal resistance is calculated using the formula: R0=(U0-Uend) / Iend. Here, R0 represents the ohmic internal resistance. If there are multiple step switching samples that meet the conditions, the median of all calculation results is taken as the historical characteristic value of the final ohmic internal resistance of the reference battery to suppress the numerical deviation caused by single sampling noise.

[0059] Next, based on the historical current data and the historical characteristic value of the first voltage in the historical dynamic operating condition data, the historical characteristic value of the first resistance of the current reference battery is calculated, and the corresponding calculation formula is R1=U1 / Imean. Similarly, based on the historical current data and the historical characteristic value of the second voltage in the historical dynamic operating condition data, the historical characteristic value of the second resistance of the current reference battery is calculated, and the corresponding calculation formula is R2=U2 / Imean, where Imean is the average current of the constant current charging / discharging step before relaxation, R1 is the first resistance, and R2 is the second resistance.

[0060] Finally, the historical characteristic value of the first time constant is compared with the historical characteristic value of the first resistor to obtain the historical characteristic value of the first capacitor of the current reference battery; the historical characteristic value of the second time constant is compared with the historical characteristic value of the second resistor to obtain the historical characteristic value of the second capacitor of the current reference battery.

[0061] In this embodiment, based on a second-order equivalent circuit model, the voltage rebound curve during the battery relaxation phase and the current excitation response under dynamic conditions are quantitatively extracted to obtain a series of candidate detection features with clear physical meanings, such as ohmic internal resistance, first internal resistance, first capacitance, first voltage, and time constant. This set of features directly responds to the deterioration of the contact state caused by tab tearing, exhibits inherent self-consistency among parameters, and is insensitive to measurement noise. Furthermore, this set of features complements the curve characterization features—the former characterizes the dynamic response, and the latter characterizes the static thermodynamic properties—together constituting a candidate detection feature system covering multidimensional degradation information. This system has a rigorous structure and strong physical interpretability, providing high-quality candidate input for subsequent target detection feature screening based on error detection balance factors.

[0062] Optionally, candidate detection features may also include the open-circuit convergence mean absolute deviation and the future relaxation start point fitting deviation, and equivalent circuit features may also include the open-circuit voltage.

[0063] Furthermore, after obtaining the historical characteristic value of the second capacitor of the current reference battery, the method further includes: determining the voltage rebound prediction sequence of the current reference battery; determining the historical characteristic value of the open-circuit convergence average absolute deviation of the current reference battery based on the voltage rebound prediction sequence and the historical characteristic value of the open-circuit voltage; and determining the historical characteristic value of the future relaxation start fitting deviation of the current reference battery based on the voltage rebound prediction sequence and the relaxation end voltage of the current reference battery.

[0064] Specifically, the open-circuit voltage is the stable terminal voltage of the battery after the internal electrochemical reaction reaches equilibrium and the polarization voltage is completely dissipated in the open-circuit resting state. In this embodiment, the open-circuit voltage is obtained by simultaneously solving the resting relaxation curve by fitting the equivalent circuit model. The average absolute deviation of open-circuit convergence is a quantitative index characterizing the degree of convergence between the full-cycle theoretical predicted voltage trajectory output by the equivalent circuit model and the overall steady-state open-circuit voltage. The fitting deviation of the future relaxation start point is a quantitative index used to evaluate the overall fitting deviation of the equivalent circuit model's entire voltage trajectory towards the final steady-state voltage, calculated from the complete voltage rebound prediction sequence and the measured steady-state voltage at the end of relaxation. The voltage rebound prediction sequence is a set of continuous time-series voltage prediction values ​​that vary with the resting relaxation time, calculated based on the characteristics of the equivalent circuit.

[0065] In the specific implementation, after obtaining the historical characteristic value of the second capacitor of the current reference battery, the historical characteristic value of the equivalent circuit feature can be substituted into the mathematical expression of the equivalent circuit model. The theoretical voltage value is calculated point by point within the preset expected time interval (which can be customized according to the battery model) to obtain the voltage rebound prediction sequence of the current reference battery. Then, the predicted voltage at each moment in the voltage rebound prediction sequence is extracted, and the difference is calculated with the historical characteristic value of the open circuit voltage in the equivalent circuit feature and the absolute value is taken to obtain the absolute deviation at each sampling moment. Then, the arithmetic mean of all absolute deviations is calculated to obtain the historical characteristic value of the open circuit convergence average absolute deviation.

[0066] Simultaneously, the average measured voltage of multiple sampling points at the end of the static relaxation step is obtained as the relaxation end voltage. The predicted voltage at each moment in the voltage rebound prediction sequence is extracted, and the difference between the predicted voltage and the relaxation end voltage is calculated and the absolute value is taken to obtain the absolute deviation at each sampling moment. Then, the absolute deviations at all moments are summed and the arithmetic mean is calculated to obtain the historical characteristic value of the fitting deviation of the future relaxation start point.

[0067] In this embodiment, after completing the fitting of the equivalent circuit model parameters, two types of features are derived based on the same set of operating condition data: the average absolute deviation of open-circuit convergence and the fitting deviation of the future relaxation start point. This allows for the discovery of entirely new dimensions for fault diagnosis without the need for additional battery testing. These two types of deviation features quantify the overall model fitting accuracy from two dimensions: the degree of convergence of the predicted voltage sequence to the steady-state open-circuit voltage and the degree of fit between the predicted voltage sequence and the measured final steady-state voltage. They can sensitively reflect the distortion of battery relaxation behavior caused by tab tearing. Simultaneously, they can be used for automated operating condition data quality verification, providing a richer and more reliable candidate feature set for subsequent feature screening and threshold optimization based on the optimal error detection balance factor.

[0068] Optionally, the multiple reference cells include normal reference cells and abnormal reference cells.

[0069] Step 211: For any candidate detection feature, take the value range of the candidate detection feature as the preset threshold search interval, and traverse each candidate threshold within the preset threshold search interval with a preset fixed step size.

[0070] Specifically, the preset threshold search interval is a numerical range defined based on the maximum and minimum values ​​of a single candidate detection feature, used to traverse and search for the optimal detection threshold. This interval has the minimum value as the lower limit and the maximum value as the upper limit, ensuring that the search range completely covers the feature values ​​of all benchmark battery samples. The preset fixed step size is a constant numerical interval at which the threshold increases single-time during traversal operations within the preset threshold search interval, and can be pre-configured according to actual working conditions and accuracy requirements. Candidate thresholds are various threshold value points generated within the preset threshold search interval, increasing sequentially from the lower limit of the interval according to the preset fixed step size; each candidate threshold serves as a temporary judgment boundary for binary classification evaluation of the benchmark battery samples as normal or abnormal. An abnormal benchmark battery is a faulty sample battery with a target fault (such as tab tearing, contact degradation, etc.). In this embodiment, the abnormal benchmark battery is a battery sample among the benchmark batteries that has been calibrated and confirmed to have tab tearing faults.

[0071] In the specific implementation, after completing the data parsing of all benchmark batteries based on a unified preset time window and solving for the historical feature values ​​of all candidate detection features corresponding to each benchmark battery, for any candidate detection feature, the preset threshold search interval corresponding to this feature is defined according to the value range of the feature; then, using a pre-configured fixed numerical step size, each candidate threshold generated in the preset threshold search interval is traversed sequentially from small to large along the numerical interval to complete the full interval threshold traversal evaluation process.

[0072] In this embodiment, through the above steps, on the one hand, the complete search interval is defined by the value range of the candidate detection features, which can effectively avoid missing the optimal judgment threshold due to the truncation of the search range; on the other hand, by using a preset fixed step size to uniformly traverse all candidate thresholds in the interval, the threshold optimization accuracy and algorithm computation overhead can be flexibly balanced, realizing automated and standardized search of the entire interval without the need for manual subjective configuration of the threshold range and traversal interval, reducing the judgment error introduced by human intervention.

[0073] Step 212: For the current candidate threshold, determine the number of abnormal batteries in the abnormal benchmark batteries whose historical feature values ​​are greater than the current candidate threshold as the detection quantity, and determine the number of normal batteries in the normal benchmark batteries whose historical feature values ​​are greater than the current candidate threshold as the overkill quantity.

[0074] In the specific implementation, for the candidate threshold to be evaluated during the traversal process, the historical feature values ​​of all abnormal baseline batteries corresponding to the candidate detection feature are compared one by one, and the total number of abnormal baseline battery samples whose historical feature values ​​are greater than the current candidate threshold is counted. This statistical value is recorded as the detection quantity. Simultaneously, the historical feature values ​​of all normal baseline batteries corresponding to the candidate detection feature are compared one by one, and the total number of normal baseline battery samples whose historical feature values ​​are greater than the current candidate threshold is counted. This statistical value is recorded as the overkill quantity.

[0075] In this embodiment, the above steps enable bidirectional quantification of the fault identification effects of two types of faults: missed fault detection and misjudgment of healthy samples, providing a reliable data foundation for subsequent unified calculation of the error detection balance factor.

[0076] Step 213: Calculate the ratio of the number of detected cases to the number of over-detected cases to obtain the error detection balance factor for the current candidate threshold.

[0077] In practice, after obtaining the two quantitative indicators, the number of detected cases and the number of overkills, corresponding to the current candidate threshold, a division operation can be performed with the number of detected cases as the numerator and the number of overkills as the denominator to calculate the ratio between the two. This ratio is then used as the error detection balance factor corresponding to the current candidate threshold. In this embodiment, the above steps can integrate fault detection capability and false positive loss into a single quantitative indicator, simultaneously balancing the two contradictory indicators, avoiding the one-sidedness of relying solely on the number of detections or overkills, and helping to screen out the optimal detection threshold with high fault detection rate and low false positives for healthy samples.

[0078] Step 214: After the traversal is completed, the candidate threshold with the largest error detection balance factor is determined as the optimal detection threshold of the candidate detection feature, and the largest error detection balance factor is determined as the optimal error detection balance factor of the candidate detection feature.

[0079] Specifically, the optimal detection threshold is the candidate threshold corresponding to the maximum value of the error detection balance factor after traversing all candidate thresholds within the preset threshold search interval. This threshold is taken as the optimal detection threshold for the candidate detection feature. This threshold achieves the optimal balance between the number of missed faults and the number of over-detected healthy batteries on the benchmark battery dataset, and is the best judgment boundary for distinguishing normal batteries from abnormal faulty batteries. The optimal error detection balance factor is the error detection balance factor value corresponding to the above-mentioned optimal detection threshold, that is, the maximum error detection balance factor obtained during the traversal. This index quantifies the comprehensive distinguishing performance of the candidate detection feature against faulty and normal samples under the optimal judgment boundary. The larger the value, the better the comprehensive performance of the feature in identifying tab tear-type faults, and it can be used as the core basis for subsequent screening and ranking of multiple candidate detection features.

[0080] In the specific implementation, after completing the traversal of all candidate thresholds within the preset threshold search interval and calculating the error detection balance factor, the error detection balance factor values ​​corresponding to all candidate thresholds are compared, and the candidate threshold corresponding to the maximum value of the error detection balance factor is selected as the optimal detection threshold for this type of candidate detection feature; at the same time, the maximum error detection balance factor obtained in the traversal process is set as the optimal error detection balance factor corresponding to the candidate detection feature, which is used to characterize the fault discrimination ability of the feature under the best judgment threshold.

[0081] In this embodiment, the above steps can be used to compare the performance of various features in identifying tab tear faults, providing an objective quantitative ranking basis for screening high-value fault features and feature dimensionality reduction, and ultimately improving the accuracy and efficiency of detection.

[0082] For example, Figure 2b This is a schematic diagram comparing the distribution of candidate detection feature samples with the corresponding optimal detection threshold, provided by an embodiment of the present invention. Figure 2b As shown, the figure contains three sub-figures (1), (2), and (3), corresponding to three candidate detection features: feature 1, feature 2, and feature 3, respectively. The horizontal axis of each sub-figure represents the battery sample number (ranging from 0 to 100), used to represent all baseline battery samples. The vertical axis represents the quantized feature value of the corresponding candidate detection feature. In the legend, triangles represent abnormal battery samples with tab tearing problems, and asterisks represent normal battery samples without faults. The black horizontal dashed line in the sub-figure is the segmentation threshold corresponding to the maximum value of the error detection balance factor after traversing all candidate thresholds for a single target feature and calculating the error detection balance factor corresponding to each threshold. This is the optimal detection threshold for that feature.

[0083] Step 215: Based on the optimal error detection balance factor corresponding to each candidate detection feature, select the target detection feature of the battery to be detected from all candidate detection features.

[0084] In practice, the optimal error detection balance factor corresponding to each candidate detection feature is compared and filtered with a pre-set benchmark error detection balance threshold (such as 0.3). Only candidate detection features with an optimal error detection balance factor value higher than the preset benchmark value are retained and used as target detection features for subsequent battery safety testing.

[0085] In this embodiment, the above steps enable the selection of target detection features that have the best overall performance in distinguishing between the two indicators of fault missed detection and sample overkill.

[0086] Step 216: Determine the real-time feature value of the target detection feature based on the real-time operating condition data of the battery to be tested, and determine the safety detection result of the battery to be tested based on the real-time feature value of the target detection feature and the optimal detection threshold.

[0087] Furthermore, based on the real-time feature value of the target detection feature and the optimal detection threshold, the safety detection result of the battery under test is determined, including: determining whether the real-time feature value of the target detection feature is greater than the optimal detection threshold of the target detection feature; if it is greater, the safety detection result of the battery under test is determined to be a potential fault with tab tearing.

[0088] In practice, after obtaining the real-time feature value of the target detection feature, it is determined whether the real-time feature value is greater than the corresponding optimal detection threshold. If it is greater, the safety detection result of the battery under test is determined to have a potential tab tear fault, and the battery under test is marked as a potential tab tear battery. Conversely, if the real-time feature value is less than or equal to the optimal detection threshold, the safety detection result of the battery under test is determined to be normal, and the battery is marked as a normal battery.

[0089] When the number of target detection features obtained through screening is greater than one, the above-mentioned comparison and judgment logic between the real-time feature value and the optimal detection threshold is executed sequentially for each target detection feature. As long as the real-time feature value of any target detection feature exceeds its corresponding optimal detection threshold, the safety detection result of the battery under test can be determined to have a potential electrode tear fault, and it is classified as a battery with potential electrode tear. Only when the real-time feature values ​​of all target detection features do not exceed their respective matched optimal detection thresholds can the safety detection result of the battery under test be determined to be normal, and it is classified as a normal battery without faults.

[0090] In this embodiment, fault determination can be completed simply by comparing real-time feature values ​​with the optimal detection threshold. The calculation is simple and the response is rapid, fully meeting the needs of online real-time battery monitoring. The optimal detection threshold is automatically optimized based on labeled samples, balancing missed detections and over-detection, effectively avoiding subjective biases from manual thresholds and achieving higher recognition accuracy. Furthermore, the determination logic closely matches the characteristic variation patterns of tab tearing faults, with a clear physical mechanism that is easy to trace. The rules are standardized and modular, stably supporting automated safety testing of large batches of batteries.

[0091] Furthermore, based on the real-time feature values ​​of the target detection features and the optimal detection threshold, the safety detection result of the battery to be tested is determined, including: counting the number of target detection features whose real-time feature values ​​are greater than the corresponding optimal detection threshold to obtain target over-limit data; and determining the safety detection result of the battery to be tested based on the target over-limit data.

[0092] Specifically, the target exceeding the limit data is obtained by comparing the real-time feature values ​​of all target detection features with the corresponding optimal detection threshold one by one, and counting the total number of features whose real-time feature values ​​exceed the corresponding optimal detection threshold.

[0093] In the specific implementation, all selected target detection features are sequentially traversed. The real-time feature value of each target detection feature is compared with its corresponding optimal detection threshold to determine whether the real-time feature value exceeds the threshold. The total number of target detection features with real-time feature values ​​greater than the corresponding optimal detection threshold is counted, and this statistical value is recorded as the target over-limit frequency, i.e., the target over-limit data. After completing the comparison of all features and obtaining the target over-limit data, the target over-limit data is parsed according to the preset over-limit quantity judgment rule to obtain the safety detection result of the current battery to be tested. For example, the preset over-limit quantity judgment rule is: if the target over-limit data is 0, the safety detection result is normal, and the battery is marked as normal; if the target over-limit data is 1, the safety detection result is suspicious, and the battery is marked for continuous tracking and monitoring; if the target over-limit data is greater than or equal to 2, the safety detection result is that there is a potential tab tear fault, the battery is marked as a potential tab tear battery, and an alarm message is sent to the staff's terminal.

[0094] In this embodiment, a progressive judgment mechanism of "independent determination of multiple target features → statistics of the number of features exceeding limits → mapping of hierarchical early warning rules" is adopted to use multiple high-performance target detection features obtained through offline optimization for online fusion decision-making. This mechanism, while ensuring fault detection sensitivity, relies on multi-feature collaborative verification and hierarchical early warning output to significantly reduce the probability of false alarms caused by occasional fluctuations in a single feature; at the same time, it accurately correlates battery safety detection results with maintenance and operation levels, balancing detection reliability and on-site maintenance operability.

[0095] The battery safety detection method provided in this invention first calculates the historical feature values ​​of candidate detection features for each benchmark battery based on historical operating data of multiple benchmark batteries, laying a data foundation for subsequently determining the detection judgment parameters of each candidate detection feature. Then, for any candidate detection feature, the value range of the candidate detection feature is used as a preset threshold search interval. Each candidate threshold within the preset threshold search interval is traversed with a preset fixed step size. On the one hand, defining the complete search interval by the value range of the candidate detection feature effectively avoids missing the optimal judgment threshold due to truncation of the search range; on the other hand, uniformly traversing all candidate thresholds within the interval with a preset fixed step size flexibly balances the threshold optimization accuracy and algorithm computational overhead, achieving automated and standardized search across the entire interval without the need for manual subjective configuration of the threshold range and traversal interval, reducing judgment errors introduced by human intervention. For the current candidate threshold, the number of abnormal batteries in the abnormal benchmark batteries whose historical feature values ​​are greater than the current candidate threshold is determined as the detection count, and the number of normal batteries in the normal benchmark batteries whose historical feature values ​​are greater than the current candidate threshold is determined as the over-detection count. This enables bidirectional quantification of the fault identification effect for both missed fault detection and misjudgment of healthy samples, providing a reliable data foundation for the subsequent unified calculation of the error detection balance factor. The ratio of the detection count to the over-detection count is calculated to obtain the error detection balance factor for the current candidate threshold. This integrates fault detection capability and misjudgment loss into a single quantitative indicator, simultaneously balancing two contradictory indicators and avoiding the one-sidedness of relying solely on the detection count or the over-detection count. This helps to screen for the optimal detection threshold with high fault detection rate and low misjudgment of healthy samples. After traversal, the candidate threshold with the largest error detection balance factor is determined as the optimal detection threshold for the candidate detection feature, and the largest error detection balance factor is determined as the optimal error detection balance factor for the candidate detection feature. This allows for a horizontal comparison of the performance of various features in identifying tab tear faults, providing an objective quantitative ranking basis for the screening of high-value fault features and feature dimensionality reduction, ultimately improving detection accuracy and efficiency. Based on the optimal error detection balance factor corresponding to each candidate detection feature, the target detection feature of the battery to be tested is selected from all candidate detection features. This allows for the selection of the target detection feature with the best comprehensive performance in distinguishing between fault missed detection and sample over-detection. Finally, the real-time feature value of the target detection feature is determined based on the real-time operating data of the battery to be tested. Based on the real-time feature value of the target detection feature and the optimal detection threshold, the safety test result of the battery to be tested is determined, which improves the accuracy, reliability, and real-time performance of the battery tab tear-specific safety hazard detection. Therefore, the technical solution of this invention overcomes the shortcomings of existing technologies that struggle to simultaneously meet the requirements of online real-time battery detection and the requirements of detection accuracy and reliability.

[0096] Figure 3This is a schematic diagram of a battery safety testing device provided in an embodiment of the present invention. This device belongs to the same inventive concept as the battery safety testing methods in the above embodiments. Details not described in detail in the embodiments of the battery safety testing device can be found in the embodiments of the battery safety testing methods described above. Figure 3 As shown, the device includes: like Figure 3 As shown, the device includes: The calculation module 310 is used to calculate the historical feature value of the candidate detection feature corresponding to each reference battery based on the historical operating condition data of multiple reference batteries; the type of the reference battery is the same as the type of the battery to be detected. The determination module 320 is used to determine the detection judgment parameters of the corresponding candidate detection features based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries. The detection judgment parameters include the optimal error detection balance factor and the optimal detection threshold. The screening module 330 is used to screen out the target detection features of the battery to be detected from all candidate detection features based on the optimal error detection balance factor corresponding to each candidate detection feature. The detection module 340 is used to determine the real-time feature value of the target detection feature based on the real-time operating condition data of the battery to be tested, and to determine the safety detection result of the battery to be tested based on the real-time feature value of the target detection feature and the optimal detection threshold.

[0097] Based on the above embodiments, the candidate detection features include curve representation features, the historical operating condition data includes historical dynamic operating condition data, and the calculation module 310 is specifically used for: For the current reference battery, a voltage increment sequence is constructed based on the historical voltage data in the historical dynamic operating condition data of the current reference battery; a capacity sequence is constructed based on the historical capacity data corresponding to the historical voltage data in the historical dynamic operating condition data of the current reference battery; an incremental capacity curve of the current reference battery is generated based on the voltage increment sequence and the capacity sequence; and a quantitative feature extraction is performed on the incremental capacity curve to obtain the historical feature values ​​of the curve characterization features of the current reference battery.

[0098] Based on the above embodiments, the candidate detection features further include equivalent circuit features, ohmic internal resistance, a first resistor, a second resistor, a first capacitor, and a second capacitor. The historical operating condition data also includes historical static operating condition data. The calculation module 310 is specifically used for: For the current reference battery, based on the historical voltage data and equivalent circuit model in the historical static operating condition data, the historical characteristic values ​​of the equivalent circuit features of the current reference battery are determined; the equivalent circuit features include a first voltage, a first time constant, a second voltage, and a second time constant; based on the historical voltage data and historical current data in the historical operating condition data, the historical characteristic value of the ohmic internal resistance of the current reference battery is calculated; based on the historical current data and the historical characteristic value of the first voltage in the historical dynamic operating condition data, the historical characteristic value of the first resistance of the current reference battery is determined, and based on the historical current data and the historical characteristic value of the second voltage in the historical dynamic operating condition data, the historical characteristic value of the second resistance of the current reference battery is determined; the ratio of the historical characteristic value of the first time constant to the historical characteristic value of the first resistance is calculated to obtain the historical characteristic value of the first capacitance of the current reference battery, and the ratio of the historical characteristic value of the second time constant to the historical characteristic value of the second resistance is calculated to obtain the historical characteristic value of the second capacitance of the current reference battery.

[0099] Based on the above embodiments, the candidate detection features further include open-circuit convergence average absolute deviation and future relaxation start point fitting deviation, and the equivalent circuit features further include open-circuit voltage. The device also includes: The feature calculation module is used to determine the voltage rebound prediction sequence of the current reference battery after obtaining the historical feature value of the second capacitor of the current reference battery; determine the historical feature value of the open-circuit convergence average absolute deviation of the current reference battery based on the voltage rebound prediction sequence and the historical feature value of the open-circuit voltage; and determine the historical feature value of the future relaxation start fitting deviation of the current reference battery based on the voltage rebound prediction sequence and the relaxation end voltage of the current reference battery.

[0100] Based on the above embodiments, the multiple reference batteries include normal reference batteries and abnormal reference batteries, and the determination module 320 is specifically used for: For any candidate detection feature, the value range of the candidate detection feature is used as a preset threshold search interval. Each candidate threshold within the preset threshold search interval is traversed with a preset fixed step size. For the current candidate threshold, the number of abnormal batteries in the abnormal reference batteries whose historical feature values ​​are greater than the current candidate threshold is determined as the detection quantity, and the number of normal batteries in the normal reference batteries whose historical feature values ​​are greater than the current candidate threshold is determined as the over-detection quantity. The ratio of the detection quantity to the over-detection quantity is calculated to obtain the error detection balance factor of the current candidate threshold. After traversal, the candidate threshold with the largest error detection balance factor is determined as the optimal detection threshold of the candidate detection feature, and the largest error detection balance factor is determined as the optimal error detection balance factor of the candidate detection feature.

[0101] Based on the above embodiments, the detection module 340 determines the safety detection result of the battery to be detected according to the real-time feature value of the target detection feature and the optimal detection threshold, including: Determine whether the real-time feature value of the target detection feature is greater than the optimal detection threshold of the target detection feature; if it is greater, then determine that the safety detection result of the battery under test is that there is a potential fault of tab tearing.

[0102] Based on the above embodiments, the detection module 340 determines the safety detection result of the battery to be detected according to the real-time feature value of the target detection feature and the optimal detection threshold, including: The number of target detection features whose real-time feature values ​​are greater than the corresponding optimal detection threshold is counted to obtain target out-of-limit data; the safety detection result of the battery to be tested is determined based on the target out-of-limit data.

[0103] The battery safety testing device provided in this embodiment of the invention can execute the battery safety testing method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0104] It is worth noting that in the embodiments of the battery safety detection device described above, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0105] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Figure 4 A block diagram of an exemplary electronic device 4 suitable for implementing embodiments of the present invention is shown. Figure 4 The electronic device 4 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0106] like Figure 4 As shown, electronic device 4 is represented in the form of a general-purpose computing electronic device. The components of electronic device 4 may include, but are not limited to: one or more processors or processing units 16, system memory 28, and bus 18 connecting different system components (including system memory 28 and processing unit 16).

[0107] Bus 18 represents one or more of several bus architectures, including a memory bus or memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any of the various bus architectures. For example, these architectures include, but are not limited to, the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MAC) bus, the Enhanced ISA bus, the Video Electronics Standards Association (VESA) local bus, and the Peripheral Component Interconnect (PCI) bus.

[0108] Electronic device 4 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by electronic device 4, including volatile and non-volatile media, removable and non-removable media.

[0109] System memory 28 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 30 and / or cache memory 32. Electronic device 4 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, storage system 34 may be used to read and write non-removable, non-volatile magnetic media (… Figure 4 Not shown; usually referred to as a "hard drive"). Although Figure 4 Not shown, a disk drive for reading and writing to a removable non-volatile disk (e.g., a "floppy disk") and an optical disk drive for reading and writing to a removable non-volatile optical disk (e.g., a CD-ROM, DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 18 via one or more data media interfaces. System memory 28 may include at least one program product having a set (e.g., at least one) of program modules configured to perform the functions of the embodiments of the present invention.

[0110] A program / utility 40 having a set (at least one) of program modules 42 may be stored, for example, in system memory 28. Such program modules 42 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include an implementation of a network environment. Program modules 42 typically perform the functions and / or methods described in the embodiments of the present invention.

[0111] Electronic device 4 can also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), and with one or more devices that enable a user to interact with electronic device 4, and / or with any device that enables electronic device 4 to communicate with one or more other computing devices (e.g., network card, modem, etc.). This communication can be performed through input / output (I / O) interface 22. Furthermore, electronic device 4 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 20. Figure 4 As shown, network adapter 20 communicates with other modules of electronic device 4 via bus 18. It should be understood that, although... Figure 4 Not shown, it can be combined with electronic device 4 to use other hardware and / or software modules, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0112] Processing unit 16 executes various functional applications and page displays by running programs stored in system memory 28, such as implementing the battery safety detection method provided in this embodiment of the invention, which includes: Based on historical operating data of multiple reference batteries, historical feature values ​​of candidate detection features corresponding to each reference battery are calculated; the type of the reference battery is the same as the type of the battery to be detected. Based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries, the detection judgment parameters of the corresponding candidate detection features are determined. The detection judgment parameters include the optimal error detection balance factor and the optimal detection threshold. Based on the optimal error detection balance factor corresponding to each candidate detection feature, the target detection feature of the battery to be detected is selected from all candidate detection features. The real-time feature value of the target detection feature is determined based on the real-time operating condition data of the battery under test, and the safety detection result of the battery under test is determined based on the real-time feature value of the target detection feature and the optimal detection threshold.

[0113] Of course, those skilled in the art will understand that the processor can also implement the technical solution of the battery safety detection method provided in any embodiment of the present invention.

[0114] This invention provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the program implements, for example, the battery safety detection method provided in this invention, the method comprising: Based on historical operating data of multiple reference batteries, historical feature values ​​of candidate detection features corresponding to each reference battery are calculated; the type of the reference battery is the same as the type of the battery to be detected. Based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries, the detection judgment parameters of the corresponding candidate detection features are determined. The detection judgment parameters include the optimal error detection balance factor and the optimal detection threshold. Based on the optimal error detection balance factor corresponding to each candidate detection feature, the target detection feature of the battery to be detected is selected from all candidate detection features. The real-time feature value of the target detection feature is determined based on the real-time operating condition data of the battery under test, and the safety detection result of the battery under test is determined based on the real-time feature value of the target detection feature and the optimal detection threshold.

[0115] The computer storage medium of this invention can be any combination of one or more computer-readable media. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0116] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, capable of sending, propagating, or transmitting programs for use by or in connection with an instruction execution system, apparatus, or device.

[0117] Program code contained on a computer-readable medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0118] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0119] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computing device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Thus, the present invention is not limited to any particular combination of hardware and software.

[0120] Furthermore, the acquisition, storage, use, and processing of data in the technical solution of this invention all comply with relevant laws and regulations.

[0121] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A method for safety testing of a battery, characterized in that, The method includes: Based on historical operating data of multiple reference batteries, historical feature values ​​of candidate detection features corresponding to each reference battery are calculated; the type of the reference battery is the same as the type of the battery to be detected. Based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries, the detection judgment parameters of the corresponding candidate detection features are determined. The detection judgment parameters include the optimal error detection balance factor and the optimal detection threshold. Based on the optimal error detection balance factor corresponding to each candidate detection feature, the target detection feature of the battery to be detected is selected from all candidate detection features. The real-time feature value of the target detection feature is determined based on the real-time operating condition data of the battery under test, and the safety detection result of the battery under test is determined based on the real-time feature value of the target detection feature and the optimal detection threshold.

2. The method according to claim 1, characterized in that, The candidate detection features include curve representation features, and the historical operating condition data includes historical dynamic operating condition data. Based on the historical operating condition data of multiple benchmark batteries, the historical feature values ​​of the candidate detection features corresponding to each benchmark battery are calculated, including: For the current reference battery, a voltage increment sequence for the current reference battery is constructed based on the historical voltage data in the historical dynamic operating condition data of the current reference battery. Based on the historical capacity data corresponding to the historical voltage data in the historical dynamic operating condition data of the current reference battery, construct the capacity sequence of the current reference battery; Based on the voltage increment sequence and the capacity sequence, the incremental capacity curve of the current reference battery is generated; The incremental capacity curve is subjected to quantitative feature extraction to obtain the historical feature values ​​of the curve characterization features of the current reference battery.

3. The method according to claim 2, characterized in that, The candidate detection features also include equivalent circuit features, ohmic internal resistance, a first resistor, a second resistor, a first capacitor, and a second capacitor. The historical operating condition data also includes historical static operating condition data. Based on the historical operating condition data of multiple reference batteries, the historical feature values ​​of the candidate detection features corresponding to each reference battery are calculated, including: For the current reference battery, based on the historical voltage data and equivalent circuit model in the historical static operating condition data, the historical characteristic values ​​of the equivalent circuit characteristics of the current reference battery are determined; the equivalent circuit characteristics include a first voltage, a first time constant, a second voltage, and a second time constant; Based on the historical voltage and current data in the historical operating condition data, calculate the historical characteristic value of the ohmic internal resistance of the current reference battery. Based on the historical current data and the historical characteristic value of the first voltage in the historical dynamic operating condition data, the historical characteristic value of the first resistance of the current reference battery is determined, and based on the historical current data and the historical characteristic value of the second voltage in the historical dynamic operating condition data, the historical characteristic value of the second resistance of the current reference battery is determined. The historical characteristic value of the first time constant is calculated to the historical characteristic value of the first resistor to obtain the historical characteristic value of the first capacitor of the current reference battery. The historical characteristic value of the second time constant is calculated to the historical characteristic value of the second resistor to obtain the historical characteristic value of the second capacitor of the current reference battery.

4. The method according to claim 3, characterized in that, The candidate detection features also include the open-circuit convergence average absolute deviation and the future relaxation start point fitting deviation. The equivalent circuit features also include the open-circuit voltage. After obtaining the historical feature values ​​of the second capacitance of the current reference battery, the following features are also included: Determine the voltage rebound prediction sequence for the current reference battery; Based on the voltage rebound prediction sequence and the historical characteristic value of the open-circuit voltage, determine the historical characteristic value of the open-circuit convergence mean absolute deviation of the current reference cell; Based on the voltage rebound prediction sequence and the relaxation end voltage of the current reference cell, determine the historical characteristic value of the fitting deviation of the future relaxation start point of the current reference cell.

5. The method according to claim 1, characterized in that, Multiple benchmark cells, including normal and abnormal benchmark cells, are used. Based on the historical feature values ​​of candidate detection features corresponding to all benchmark cells, detection judgment parameters for the corresponding candidate detection features are determined, including: For any candidate detection feature, the value range of the candidate detection feature is used as a preset threshold search interval, and each candidate threshold in the preset threshold search interval is traversed with a preset fixed step size. For the current candidate threshold, the number of abnormal batteries in the abnormal reference batteries whose historical feature values ​​are greater than the current candidate threshold is determined as the detection quantity, and the number of normal batteries in the normal reference batteries whose historical feature values ​​are greater than the current candidate threshold is determined as the overkill quantity. Calculate the ratio of the number of detected cases to the number of overkilled cases to obtain the error detection balance factor for the current candidate threshold; After the traversal is completed, the candidate threshold with the largest error detection balance factor is determined as the optimal detection threshold of the candidate detection feature, and the largest error detection balance factor is determined as the optimal error detection balance factor of the candidate detection feature.

6. The method according to claim 1, characterized in that, Based on the real-time feature values ​​of the target detection features and the optimal detection threshold, the safety detection result of the battery to be detected is determined, including: Determine whether the real-time feature value of the target detection feature is greater than the optimal detection threshold of the target detection feature; If the value is greater than 1, then the safety test result of the battery under test is determined to be a potential fault with a torn tab.

7. The method according to claim 1, characterized in that, Based on the real-time feature values ​​of the target detection features and the optimal detection threshold, the safety detection result of the battery to be detected is determined, including: The number of target detection features whose real-time feature values ​​are greater than the corresponding optimal detection threshold is counted to obtain target out-of-limit data; The safety test result of the battery to be tested is determined based on the target out-of-limit data.

8. A battery safety detection device, characterized in that, The device includes: The calculation module is used to calculate the historical feature value of the candidate detection feature corresponding to each reference battery based on the historical operating condition data of multiple reference batteries; the type of the reference battery is the same as the type of the battery to be detected. The determination module is used to determine the detection judgment parameters of the corresponding candidate detection features based on the historical feature values ​​of the candidate detection features corresponding to all benchmark batteries. The detection judgment parameters include the optimal error detection balance factor and the optimal detection threshold. The filtering module is used to filter out the target detection features of the battery to be detected from all candidate detection features based on the optimal error detection balance factor corresponding to each candidate detection feature. The detection module is used to determine the real-time feature value of the target detection feature based on the real-time operating condition data of the battery under test, and to determine the safety detection result of the battery under test based on the real-time feature value of the target detection feature and the optimal detection threshold.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the battery safety detection method according to any one of claims 1-7.

10. A storage medium containing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a computer processor, are used to perform the battery safety detection method according to any one of claims 1-7.

Images