A method, apparatus, device, and medium for detecting common defects of a battery

Abstract

The application provides a battery common defect detection method, device, equipment and medium, the method comprises: obtaining the charging data of a plurality of vehicles in a target batch, and performing charging event segmentation; for the segmented charging event, calculate the spatial dimension feature and the energy dimension feature; identify abnormal vehicles based on the abnormal occurrence rate of each feature of a single vehicle; determine the proportion of abnormal vehicles and the batch abnormal signal represented by the coordinated occurrence of multiple dimensional abnormal features; in response to the batch abnormal signal, by comparing the differences between the target batch and the preset normal batch in different characteristics under the same user behavior characteristics, verify whether the batch abnormal signal is caused by the vehicle itself defect; if the verification is passed, finally determine that the battery of the target batch exists common defect. Through the above method, the application embodiment effectively identifies and locates the systematic common defect of the battery batch.

Description

Technical Field

[0001] This application relates to the field of power battery quality monitoring and big data analysis technology, specifically to a method, device, equipment and medium for detecting common defects in batteries. Background Technology

[0002] As a core component of electric vehicles, the performance and reliability of the power battery directly affect vehicle safety, driving range, and user experience. During mass production and use, battery systems may experience batch-specific quality issues due to inherent defects in design, manufacturing, software, or hardware. If such batch defects are not detected in a timely manner, they can lead to large-scale vehicle malfunctions, accelerated performance degradation, and even safety risks.

[0003] The inventors discovered in their research that existing technical solutions mainly focus on fault diagnosis of single vehicle power batteries, making it difficult to effectively identify and warn of systemic defect risks that are widespread in batches of vehicles and are caused by common root causes.

[0004] Therefore, a solution is needed. Summary of the Invention

[0005] In view of this, embodiments of this application provide a method, apparatus, device, and medium for detecting common defects in batteries, in order to solve the technical problem in the prior art that it is difficult to effectively identify and locate systematic common defects in batches of batteries.

[0006] In a first aspect, embodiments of this application provide a method for detecting common defects in batteries, the method comprising: Obtain charging data for multiple vehicles in the target batch and segment charging events; For the segmented charging events, calculate the spatial dimension features used to characterize the internal consistency of the battery and the energy dimension features used to characterize the energy conversion efficiency; Abnormal vehicles are identified based on the anomaly rates of the spatial dimension features and the energy dimension features in multiple charging events of a single vehicle. The proportion of the abnormal vehicles and the batch abnormal signal characterized by the coordinated occurrence of multiple abnormal features were determined. In response to the batch anomaly signal, user charging behavior characteristics are introduced as control variables. By comparing the differences between the target batch and the preset normal batch in the spatial dimension characteristics and the energy dimension characteristics under the same user behavior characteristics, it is verified whether the batch anomaly signal originates from a defect in the vehicle itself. If the verification passes, it is ultimately determined that the batteries of multiple vehicles in the target batch have a common defect.

[0007] In one feasible implementation, determining the proportion of the abnormal vehicles and the batch anomaly signal characterized by the coordinated occurrence of multiple dimensions of anomaly features includes: When the proportion of abnormal vehicles exceeds a first threshold, and the proportion of abnormal vehicles that simultaneously exhibit at least two abnormal features exceeds a second threshold, the batch abnormal signal is generated.

[0008] In one feasible implementation, the differences between the target batch and a preset normal batch in the spatial dimension features and the energy dimension features are compared under the same user behavior characteristics, including: All vehicles are grouped according to user behavior characteristics; these user behavior characteristics are used to characterize users' charging habits. Within the same group, calculate the statistical differences between the target batch and the preset normal batch in terms of the spatial dimension feature and the energy dimension feature; If, in at least one group, the statistical difference exceeds a preset fluctuation range determined based on the data of the normal batch, the verification is deemed successful, confirming that the abnormal signal of the batch does not originate from differences in user charging habits.

[0009] In one feasible implementation, the method further includes: Based on the abnormal combinations of specific characteristics exhibited by abnormal vehicles in the target batch, the type of the common defect is located; Output the types of common defects.

[0010] In one feasible implementation, the spatial dimension features include a first spatial dimension feature and / or a second spatial dimension feature; Based on the abnormal combinations of specific characteristics exhibited by abnormal vehicles in the target batch, the type of the common defect is identified, including: If the proportion of vehicles with both first spatial dimension feature anomalies and second spatial dimension feature anomalies among the abnormal vehicles reaches the first coordination threshold, then the common defect is identified as a manufacturing defect. If the proportion of vehicles with both abnormal energy dimension characteristics and abnormal spatial dimension characteristics reaches the second coordination threshold, then the common defect is identified as a design defect. If the proportion of vehicles with abnormal spatial dimension features but normal energy dimension features and other spatial dimension features reaches the third coordination threshold, then the common defect is identified as a software defect. If the proportion of vehicles with abnormal energy dimension characteristics and normal spatial dimension characteristics among the abnormal vehicles reaches the fourth collaborative threshold, then the common defect is identified as a hardware defect.

[0011] In one feasible implementation, the first spatial dimension feature is voltage fluctuation rate; the second spatial dimension feature is voltage uniformity coefficient; and the energy dimension feature is charging efficiency deviation.

[0012] In one feasible implementation, charging data for multiple vehicles in the target batch is acquired, and charging events are segmented, including: Obtain charging data for multiple vehicles in the target batch, and preprocess the charging data; the preprocessing includes at least one of the following: temporal continuity verification, necessary column integrity verification, and filling missing values ​​with a missing rate below a threshold by linear interpolation. Based on the step change of the charging state signal, independent charging event segments are identified from the preprocessed charging data. The identified charging event segments are filtered for validity to exclude invalid events caused by momentary interference or misoperation.

[0013] Secondly, embodiments of this application also provide a common defect detection device for batteries, the device comprising: The segmentation module is used to acquire charging data of multiple vehicles in the target batch and segment charging events. The calculation module is used to calculate the spatial dimension features that characterize the internal consistency of the battery and the energy dimension features that characterize the energy conversion efficiency for the segmented charging events. The identification module is used to identify abnormal vehicles based on the anomaly rates of the spatial dimension features and the energy dimension features in multiple charging events of a single vehicle. The determination module is used to determine the proportion of the abnormal vehicles and the batch abnormal signal characterized by the coordinated occurrence of multiple dimensions of abnormal features. The verification module is used to respond to the batch abnormal signal by introducing user charging behavior characteristics as control variables. By comparing the differences between the target batch and the preset normal batch in the spatial dimension characteristics and the energy dimension characteristics under the same user behavior characteristics, the module verifies whether the batch abnormal signal originates from a defect in the vehicle itself. The determination module is used to determine that if the verification passes, the batteries of multiple vehicles in the target batch have a common defect.

[0014] In one feasible implementation, a determining module is used to determine the proportion of the abnormal vehicles and a batch anomaly signal representing the coordinated occurrence of multiple dimensions of anomaly features, for the following purposes: When the proportion of abnormal vehicles exceeds a first threshold, and the proportion of abnormal vehicles that simultaneously exhibit at least two abnormal features exceeds a second threshold, the batch abnormal signal is generated.

[0015] In one feasible implementation, the verification module is used to compare the differences between the target batch and the preset normal batch in the spatial dimension feature and the energy dimension feature under the same user behavior characteristics, including: All vehicles are grouped according to user behavior characteristics; these user behavior characteristics are used to characterize users' charging habits. Within the same group, calculate the statistical differences between the target batch and the preset normal batch in terms of the spatial dimension feature and the energy dimension feature; If, in at least one group, the statistical difference exceeds a preset fluctuation range determined based on the data of the normal batch, the verification is deemed successful, confirming that the abnormal signal of the batch does not originate from differences in user charging habits.

[0016] In one feasible implementation, the device further includes: The positioning module is used to locate the type of the common defect based on the abnormal combination of specific characteristics exhibited by abnormal vehicles in the target batch; The output module is used to output the types of common defects.

[0017] In one feasible implementation, the spatial dimension features include a first spatial dimension feature and / or a second spatial dimension feature; The positioning module is used to locate the type of the common defect based on the abnormal combination of specific characteristics exhibited by abnormal vehicles in the target batch, for the following purposes: If the proportion of vehicles with both first spatial dimension feature anomalies and second spatial dimension feature anomalies among the abnormal vehicles reaches the first coordination threshold, then the common defect is identified as a manufacturing defect. If the proportion of vehicles with both abnormal energy dimension characteristics and abnormal spatial dimension characteristics reaches the second coordination threshold, then the common defect is identified as a design defect. If the proportion of vehicles with abnormal spatial dimension features but normal energy dimension features and other spatial dimension features reaches the third coordination threshold, then the common defect is identified as a software defect. If the proportion of vehicles with abnormal energy dimension characteristics and normal spatial dimension characteristics among the abnormal vehicles reaches the fourth collaborative threshold, then the common defect is identified as a hardware defect.

[0018] In one feasible implementation, the first spatial dimension feature is voltage fluctuation rate; the second spatial dimension feature is voltage uniformity coefficient; and the energy dimension feature is charging efficiency deviation.

[0019] In one feasible implementation, the segmentation module is used to acquire charging data from multiple vehicles in the target batch and perform charging event segmentation, for the following purposes: Obtain charging data for multiple vehicles in the target batch, and preprocess the charging data; the preprocessing includes at least one of the following: temporal continuity verification, necessary column integrity verification, and filling missing values ​​with a missing rate below a threshold by linear interpolation. Based on the step change of the charging state signal, independent charging event segments are identified from the preprocessed charging data. The identified charging event segments are filtered for validity to exclude invalid events caused by momentary interference or misoperation.

[0020] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the method as described in any one of the first aspects.

[0021] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the steps of the method as described in any one of the first aspects.

[0022] This application provides a method, apparatus, device, and medium for detecting common defects in batteries. By constructing a multi-level analysis framework from "individual anomaly screening" to "group signal generation" and then to "root cause attribution verification," it effectively overcomes the limitations of existing technologies. Compared with existing technologies that typically only focus on single-vehicle diagnostics, this solution first calculates multi-dimensional performance characteristics (spatial and energy dimensions) and analyzes their anomaly incidence rate, achieving a leap from single-vehicle level to batch level risk perception.

[0023] Furthermore, this solution innovatively introduces user behavior characteristics as control variables. By comparing the performance differences between the target batch and the normal batch under the same user habits, it can accurately verify whether the abnormal signals of the batch originate from the defects of the vehicle itself, thereby eliminating misjudgments caused by different user habits. Ultimately, it achieves highly reliable and accurate identification and confirmation of systemic common defects in battery batches.

[0024] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0025] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A flowchart of a common defect detection method for batteries provided in an embodiment of this application is shown.

[0027] Figure 2 A flowchart of another common defect detection method for batteries provided in an embodiment of this application is shown.

[0028] Figure 3 A schematic diagram of the structure of a common defect detection device for batteries provided in an embodiment of this application is shown.

[0029] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely represents selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0031] As a core component of electric vehicles, the performance and reliability of the power battery directly affect vehicle safety, driving range, and user experience. During mass production and use, battery systems may experience batch-specific quality issues due to inherent defects in design, manufacturing, software, or hardware. If such batch defects are not detected in a timely manner, they can lead to large-scale vehicle malfunctions, accelerated performance degradation, and even safety risks, resulting in significant recall costs and damage to the manufacturer's brand reputation.

[0032] The inventors discovered in their research that existing technical solutions mainly focus on fault diagnosis of individual vehicle power batteries, making it difficult to effectively identify and warn of systemic defect risks that are widespread in batches of vehicles due to common root causes. Specifically, existing solutions have the following limitations: First, they lack an effective method for extracting common characteristics of batches from massive amounts of operational data; second, they cannot distinguish the impact of vehicle defects on battery performance from external user habits, leading to misjudgments.

[0033] Based on this, this application provides a common defect detection scheme for batteries. The core of the scheme lies in analyzing charging data from multiple vehicles within a batch. First, performance characteristics are calculated from the spatial dimension representing internal battery consistency and the energy dimension representing energy conversion efficiency, and abnormal vehicles are identified accordingly. Then, by analyzing the group proportion of abnormal vehicles and the synergistic anomalies of multi-dimensional characteristics, batch-related defect signals are preliminarily discovered. Finally, user charging behavior characteristics are innovatively introduced as control variables to verify whether the defect signal truly originates from the vehicle itself, thereby achieving high-precision and high-reliability identification and confirmation of common defects in battery batches.

[0034] The following is a description through examples.

[0035] To facilitate understanding of this embodiment, a method for detecting common defects in batteries disclosed in this application will first be described in detail. For example... Figure 1 As shown, the method includes the following steps: Step 101: Obtain charging data for multiple vehicles in the target batch and perform charging event segmentation.

[0036] This step aims to prepare a standardized, high-quality data foundation for subsequent analysis. Specifically, it includes: Data Acquisition and Preprocessing: Obtain raw charging data from multiple vehicles in the target batch (such as vehicles of the same model and from the same production period) from the vehicle terminal or cloud storage system. The charging data is time-series data and includes at least key fields such as total voltage, total current, individual cell voltage, charging status, and battery state of charge (SOC).

[0037] To ensure data quality, the raw data needs to be preprocessed.

[0038] Charging event segmentation: Based on preprocessed time-series data, independent charging events are identified and segmented. A charging event refers to a complete process from the start to the end of charging.

[0039] For example, charging data for multiple vehicles in a target batch is obtained, and charging events are segmented, such as... Figure 2As shown, the steps include: Step 201: Obtain charging data of multiple vehicles in the target batch and preprocess the charging data; the preprocessing includes at least one of the following: temporal continuity verification, necessary column integrity verification, and filling missing values ​​with a missing rate below a threshold by linear interpolation.

[0040] First, raw charging data from multiple vehicles in the target batch is acquired. This data is typically a time-series data stream collected from onboard terminals or cloud platforms. To ensure the accuracy of subsequent analysis, the raw data must undergo rigorous preprocessing to guarantee its quality. Preprocessing operations mainly include: (1) Time sequence continuity check: Check whether there are abnormal interruptions in the timestamp sequence of data points. For example, check whether the time interval between adjacent data points exceeds a preset reasonable window to eliminate long-term gaps caused by data acquisition or transmission failures.

[0041] For example, suppose the time-series data of the i-th vehicle contains T sampling points, and its timestamp sequence is as follows: (Unit: seconds).

[0042] To ensure data validity, the interval between any two adjacent timestamps in the sequence must meet the following constraints: (This holds true for all j=1 to T-1). This check can effectively identify and exclude long-term missing data segments caused by interruptions in data transmission or recording.

[0043] (2) Required column integrity check: Verify that the data contains all the key fields necessary for analysis, such as total voltage, total current, individual cell voltage, state of charge (SOC), etc. The absence of any required column will cause that part of the data to be marked as incomplete.

[0044] (3) Missing value handling: For data that passes the time series continuity and necessary column integrity checks, its data quality needs to be further evaluated and missing values ​​need to be handled. This solution ensures data reliability through quantitative evaluation and threshold judgment.

[0045] For example, define the missing rate of the data in column k of the i-th vehicle. This is the ratio of the number of missing values ​​in this column to the total number of samples. A preset missing value threshold (e.g., 5%) is set as the criterion for judging data usability, as shown in the following formula:

[0046] When the missing rate When the missing value is less than 5%, the degree of missing data in that column is considered acceptable. In this case, linear interpolation is used to impute the missing values. At this low missing rate, linear interpolation can effectively restore the continuity of the data, and the impact of imputation on the overall statistical distribution of the data is negligible, reliably supporting subsequent analysis.

[0047] When the missing rate If the percentage is greater than 5%, the data in that column is deemed unreliable and contains serious omissions. Such data will not be used to ensure that subsequent feature calculations and defect analyses are not biased or misjudged due to low-quality data.

[0048] Step 202: Based on the step change of the charging state signal, identify independent charging event segments from the preprocessed charging data.

[0049] The core objective of this step is to segment continuous time-series charging data into multiple independent charging events for subsequent refined analysis. The underlying technology is that charging events exhibit significant "state continuity," meaning that during a complete charging process, the vehicle's "charging state" signal remains in a stable non-zero state (e.g., a value of 1). This allows for precise definition of event boundaries by detecting the start and end transition points of this state (from 1 to 0, or from 0 to 1).

[0050] For example, suppose the charging state of the i-th vehicle at time j is... Its value range is Where 0 represents the "non-charging" state and 1 represents the "charging" state. Based on this, a complete charging event is defined as a continuous time interval. At any time within this interval Their charging states all meet the requirements. The formula is as follows:

[0051] The boundaries of charging events are defined by step changes in the charging state signal (i.e., transitions from "state 0→1" and "1→0"). In one embodiment of this application, a differential operator is used to automatically identify these key time points, in conjunction with the step change detection principle.

[0052] The difference in charging state between adjacent time points is calculated using the following formula. :

[0053] Start time recognition: when When the charging state transitions from "non-charging (0)" to "charging (1)", this indicates that the charging state has changed from "non-charging (0)" to "charging (1)". The start time of the charging event was determined. .

[0054] End time identification: when When the charging state transitions from "charging (1)" back to "non-charging (0)", this indicates that the time point is approaching. The time determined as the end time of the charging event. .

[0055] By traversing the entire preprocessed time series data, all data that meet the criteria are identified. Yes, this allows for the precise segmentation of independent charging event fragments. Each segmented event fragment corresponds to a complete process in the physical world from the start to the end of charging, providing a foundation for subsequent calculations.

[0056] Step 203: Validity screening of the identified charging event segments is performed to exclude invalid events caused by momentary interference or misoperation.

[0057] Not all charging event fragments identified in step 202 are valuable for analysis. In real-world scenarios, there may be momentary charging events caused by user misoperation (such as attempting to plug and unplug the charging gun) or connection failure. These events are extremely short-lived and do not constitute a valid, complete charging process. Including them in the analysis would introduce noise and affect the accuracy of subsequent feature calculations and defect determination.

[0058] To address this issue, this step filters the validity of charging events based on their physical duration. The specific filtering criteria are as follows: Duration calculation: For a duration calculated from the start time... and end time The duration of a defined charging event is defined as follows: (Unit: seconds).

[0059] This scheme sets a minimum duration threshold (e.g., 300 seconds, or 5 minutes). When the duration of the charging event meets the following conditions: Only when the event is completed is it considered a valid charging event and retained for subsequent analysis.

[0060] This screening mechanism can efficiently filter out non-real charging processes, ensuring that subsequent multi-dimensional feature calculations, anomaly identification, and common defect analysis are all based on high-quality, representative charging event data, thereby significantly improving the reliability and accuracy of the entire detection method.

[0061] Step 102: For the segmented charging events, calculate the spatial dimension features used to characterize the internal consistency of the battery and the energy dimension features used to characterize the energy conversion efficiency.

[0062] From each individual charging event, key performance indicators that reflect the battery's health status from different physical dimensions are extracted. These quantified features form the basis for subsequent anomaly identification and defect diagnosis.

[0063] Specifically, the following two types of features are calculated: (1) Spatial dimension characteristics: characterizing the internal consistency of the battery These characteristics focus on the voltage dispersion of individual cells within a battery pack, used to assess the consistency between cells. Poor consistency is a typical manifestation of manufacturing defects or uneven aging.

[0064] For example, the spatial dimension features of this solution include: a first spatial dimension feature and a second spatial dimension feature.

[0065] The first spatial dimension feature is voltage fluctuation rate. The second spatial dimension feature is the uniform pressure coefficient. .

[0066] Voltage fluctuation rate of the i-th vehicle during the e-th charging event The calculation formula is as follows:

[0067] Where T represents the total number of time points (total number of sampling points) included in the charging event; t is a summation index representing the t-th time point. t ranges from 1 to T, indicating that we are iterating through every sampling moment of this charging event; n represents the total number of cells in the battery pack; k is a summation index representing the k-th cell, and k ranges from 1 to n, indicating that we are iterating through every cell in the battery pack. This represents the voltage value of the k-th cell at time t during the e-th charging event of the i-th vehicle; This represents the average voltage of all (n) cells at time t during the e-th charging event of the i-th vehicle.

[0068] This feature quantifies the overall fluctuation of all cell voltages relative to the average voltage throughout the charging process. A higher value indicates greater voltage differences between cells and poorer consistency.

[0069] Voltage uniformity coefficient of the e-th charging event of the i-th vehicle The calculation formula is as follows:

[0070] Where n represents the total number of cells in the battery pack; k is a summation index, representing the k-th cell, and k ranges from 1 to n, indicating that we are iterating through every cell in the battery pack; This represents the end time of the e-th charging event for the i-th vehicle, and the voltage value of the k-th battery cell. Let represent the average voltage of all (n) cells at the end of the e-th charging event for the i-th vehicle.

[0071] The fraction in parentheses calculates the relative deviation between the voltage of the k-th cell and the average voltage. It indicates how much the cell deviates from the average value.

[0072] Voltage uniformity coefficient of the e-th charging event of the i-th vehicle The key feature is the dispersion of cell voltages at the end of charging. This directly reflects the effectiveness of the battery management system's (BMS) balancing function. An abnormal value indicates that the BMS is unable to equalize the cell voltages at the end of charging.

[0073] 2) Energy dimension characteristics: characterizing energy conversion efficiency These features focus on the efficiency of energy conversion and storage during charging, and are used to assess the health of the chemical reactions inside the battery.

[0074] For example, this solution calculates the charging efficiency deviation. This feature quantifies abnormal energy loss during the charging process by comparing the deviation between the actual charging efficiency and the theoretical charging efficiency. The larger the deviation in charging efficiency, the more significant the abnormal energy loss during charging, which may originate from problems such as cell material defects, BMS malfunctions, or charging hardware failures.

[0075] Actual charging efficiency Defined as the ratio of the actual amount of charge stored in the battery to the total amount of charge charged into the battery, its calculation formula is as follows:

[0076] in, This represents the change in state of charge (SOC) for the e-th charging event of the i-th vehicle, signifying the "relative value" of the battery's actual stored charge during this charging cycle. Specifically, it's calculated as: the battery SOC at the end of the charging event minus the battery SOC at the beginning of the charging event, as shown in the following formula:

[0077] Indicates the battery's rated capacity (unit: Ah). Molecule Together, they represent the absolute amount of charge actually stored in the battery during this charge.

[0078] This represents the total amount of charge actually deposited into the battery during the charging event. It is calculated by integrating the absolute value of the charging current over time, as shown in the following formula:

[0079] in, Let represent the total current collected at time t for the i-th vehicle during the e-th charging event.

[0080] Charging efficiency deviation of vehicle i in the e-th charging event The calculation formula is as follows:

[0081] in, The theoretical charging efficiency of a lithium-ion battery is an empirical constant (for example, a value of 0.92, or 92%). It is the charging efficiency of the i-th vehicle during the e-th charging event.

[0082] This step transforms each charging event into a set of quantifiable feature values ​​with clear physical meaning, providing data input for the next step of identifying anomalies at the vehicle level.

[0083] Step 103: Identify abnormal vehicles based on the anomaly rates of the spatial dimension features and energy dimension features in multiple charging events of a single vehicle.

[0084] The core idea is to precisely identify individual vehicles from the target batch that exhibit persistent and systematic performance abnormalities. Instead of relying on occasional anomalies from a single charging event, the determination of a vehicle's true "abnormal vehicle" is based on the high frequency of abnormal behavior observed in its historical charging history.

[0085] For example, suppose the implementation process involves two levels: (1) Judgment of anomalies in a single charging event First, for the e-th charging event of the i-th vehicle, an anomaly determination is performed based on the performance characteristics calculated in step 102. The following determination rules are set: if any one of these conditions is met, the charging event is marked as a "single anomaly": Abnormal voltage fluctuation: >5%.

[0086] Abnormal charging efficiency deviation: >15%.

[0087] Abnormal pressure uniformity coefficient >0.005.

[0088] (2) Single vehicle anomaly detection Based on the judgment of a single incident, the overall health status of a single vehicle is assessed.

[0089] Calculate the anomaly occurrence rate: Calculate the "percentage of anomalies" for the i-th vehicle. The calculation formula is as follows:

[0090] in, This represents the total number of events that marked vehicle i as a "single anomaly"; This represents the total number of valid charging events for the i-th vehicle. A threshold is set (e.g., 30%) to determine the percentage of "abnormal events". If the percentage is greater than 30%, the vehicle will be classified as "single vehicle abnormality".

[0091] This two-tiered judgment mechanism based on anomaly occurrence rate effectively distinguishes between "occasional failures" and "systemic defects." The 30% threshold can effectively exclude occasional anomalies caused by random fluctuations or transient interference, ensuring that the identified abnormal vehicles do indeed exhibit the persistence and authenticity of the anomalies.

[0092] The final selected group of "abnormal vehicles" is a collection of vehicles with high confidence and potential defects, providing high-quality and reliable input for the next step of "batch common defects" analysis.

[0093] Step 104: Determine the proportion of the abnormal vehicles and the batch abnormal signal characterized by the coordinated occurrence of multiple dimensions of abnormal features.

[0094] This step is the core of batch-level risk perception. Its purpose is to extract group-wide and systemic failure modes from individual anomalies, thereby generating a batch anomaly signal. This batch anomaly signal serves as a preliminary batch anomaly warning, meaning that further verification is needed to determine whether it represents an anomaly in the target batch of batteries.

[0095] It is important to note that the core idea here is that genuine batch defects will exhibit abnormalities in both quantity and quality.

[0096] "Quantity" anomaly: This refers to the fact that the number of problematic vehicles must reach a certain scale (i.e. proportion) to exclude the impact of occasional or isolated cases.

[0097] The "qualitative" anomaly refers to the fact that the fault modes exhibited by these abnormal vehicles should not be random or singular, but should present specific and complex combined characteristics (i.e., synergistic occurrence). This points to a common and systematic root cause of the fault, rather than a simple superposition of multiple unrelated problems.

[0098] By comprehensively evaluating the proportion of anomalous vehicles in a group and the synergistic occurrence of anomalous features, a reliable "batch anomaly signal" can be constructed. This signal indicates that there may be common defect risks within the target batch that require further investigation. The technical approach to achieving this judgment can be based on logical judgments using proportion thresholds, or it can employ more complex clustering or pattern recognition algorithms.

[0099] For example, determining the proportion of the abnormal vehicles and the batch anomaly signal characterized by the coordinated occurrence of multiple dimensions of anomaly features includes: When the proportion of abnormal vehicles exceeds a first threshold, and the proportion of abnormal vehicles that simultaneously exhibit at least two abnormal features exceeds a second threshold, the batch abnormal signal is generated.

[0100] When the proportion of abnormal vehicles meeting condition (1) exceeds a first threshold (e.g., 30%), this ensures the prevalence of the problem, rather than an isolated incident. The formula for calculating the proportion r of abnormal vehicles is as follows:

[0101] in, This indicates the number of vehicles with "single vehicle anomaly" within the target batch; This represents the total number of vehicles in the target batch. When r ≥ 30%, the problem is considered statistically significant within the batch. This threshold (30%) is set to effectively exclude individual vehicle anomalies caused by sporadic, isolated faults, thereby ensuring that the detected signals represent a widespread, batch-wide problem, rather than isolated cases.

[0102] Furthermore, condition (2) is met: the proportion of vehicles exhibiting at least two anomalous features simultaneously exceeds the second threshold (e.g., 60%). This demonstrates the synergistic and systematic nature of the defect patterns. When both conditions are met simultaneously, the batch anomalous signal is generated. The calculation formula is as follows:

[0103] in, This indicates the number of vehicles exhibiting two or more anomalous characteristics simultaneously. The formula above calculates the proportion of vehicles exhibiting at least two anomalous characteristics to the total number of vehicles with single-vehicle anomalous characteristics within the target batch. When the rate exceeds 60%, it indicates that most problematic vehicles exhibit interconnected, multi-indicator abnormal patterns. This high degree of "synergy" strongly demonstrates the systemic root cause of the defect, as it rules out the possibility of a single, random, and scattered cause of failure, pointing to a common, deep-seated source of defect, such as a unified manufacturing deviation or design flaw.

[0104] Step 105: In response to the batch anomaly signal, user charging behavior characteristics are introduced as control variables. By comparing the differences between the target batch and the preset normal batch in the spatial dimension characteristics and the energy dimension characteristics under the same user behavior characteristics, it is verified whether the batch anomaly signal originates from a defect in the vehicle itself.

[0105] This step is crucial for defect attribution, and its purpose is to eliminate the most significant confounding factor: user habits (in this application embodiment, the user's charging habits are the primary focus), in order to confirm that the abnormal signal discovered in step 104 truly originates from a defect in the vehicle itself, rather than from user behavior.

[0106] The batch anomaly signal discovered in step 104 has an alternative explanation: that owners of the target batch may generally have certain charging habits that lead to battery performance degradation (e.g., extremely high-frequency fast charging). To eliminate this alternative explanation, this step employs the "controlled variable method" from scientific research.

[0107] By introducing user charging behavior characteristics as a grouping criterion, the target batch and normal batch of vehicles are matched according to the same charging habits to form multiple comparable groups.

[0108] Within each group with identical user behavior, compare the target batch with the normal batch in terms of core performance metrics. If, under this "fair" comparison environment, the target batch still shows a significant disadvantage, then it strongly suggests that the root cause of the performance difference lies in the vehicle itself, and is unrelated to how the user uses it.

[0109] For example, by comparing the differences between the target batch and the preset normal batch in the spatial dimension feature and the energy dimension feature under the same user behavior characteristics, including: All vehicles are grouped according to user behavior characteristics; the user behavior characteristics are used to characterize the user's charging habits; within the same group, the statistical differences between the target batch and the preset normal batch in the spatial dimension characteristics and the energy dimension characteristics are calculated; if the statistical difference exceeds the preset fluctuation range determined based on the data of the normal batch in at least one group, the verification is deemed to have passed, thus confirming that the abnormal signal of the batch is not due to the difference in user charging habits.

[0110] Among them, the percentage of fast charging times for vehicle i The calculation formula is as follows:

[0111] in, This represents the number of fast charging cycles for the i-th vehicle; This represents the number of slow charging attempts for the i-th vehicle. The percentage of these fast charging attempts is [not specified]. This reflects users' frequency preference for choosing fast charging mode.

[0112] Fast charging SOC percentage of vehicle i The calculation formula is as follows:

[0113] in, The numerator represents the SOC increment of a single charging event e, i.e., how much battery capacity was added during this charge (expressed as a percentage). For example, charging from 20% to 80% results in a 60% increment. The summation formula indicates that the SOC increments of all events classified as "fast charging" are summed. Therefore, the numerator represents the total battery capacity (expressed as the sum of SOC percentages) gained by vehicle i through all fast charging events over a statistical period.

[0114] Similarly, the denominator represents the total amount of electricity (expressed as the sum of SOC percentages) obtained by the i-th vehicle through all charging methods within the same statistical time period.

[0115] Fast charging SOC percentage of vehicle i It measures the proportion of total electricity a user replenishes through fast charging to their total charging volume. Compared to the percentage of fast charging sessions, it more accurately reflects a user's reliance on fast charging.

[0116] Based on selected user behavior characteristics (such as fast charging SOC percentage), all vehicles (including target batches and normal batches) are divided into several intervals (e.g., [0%-20%], [20%-40%], ..., [80%-100%]).

[0117] Within the same group (i.e., a user group with similar charging habits), calculate the statistical differences between the target batch and the normal batch in terms of spatial and energy dimensions (e.g., calculate the difference between the mean values ​​of a certain indicator of the two batches within the group).

[0118] If, within at least one group, the statistical difference exceeds a preset fluctuation range determined based on normal batch data (for example, the difference is greater than twice the standard deviation of the performance index of the normal batch within that group), the verification is deemed successful.

[0119] When the verification is passed, it is considered that: a significant abnormal signal in the group has been found (step 104). User behavior, the most significant external interference factor, has been ruled out, confirming that the abnormality originates from the vehicle itself (step 105). At this point, the chain of evidence from both aspects is complete, providing sufficient grounds for a final judgment: the target batch has a common defect. This conclusion provides direct and reliable data support for initiating a quality investigation, targeted recall, or production improvement.

[0120] Step 106: If the verification passes, it is finally determined that the batteries of multiple vehicles in the target batch have a common defect.

[0121] The conclusion reached in step 206 is not based on a single piece of evidence, but rather on a summary and logical AND operation of the conclusions from previous steps. Specifically, it means that all of the following conditions have been met: Individual anomaly confirmation (step 103): A considerable number of vehicles exhibiting persistent performance anomalies have been identified at the single-vehicle level.

[0122] Group anomaly signal confirmation (step 104): These abnormal vehicles not only reached a certain group proportion, but their failure modes also showed a systematic feature of multi-indicator synergy, indicating that the problem has batch commonality.

[0123] Self-defect attribution confirmation (step 105): Through rigorous comparative analysis, the interference of the most important external factor, "user charging habits", has been eliminated, confirming that the root cause of the performance difference lies in the vehicle itself.

[0124] This application provides a method, apparatus, device, and medium for detecting common defects in batteries. By constructing a multi-level analysis framework from "individual anomaly screening" to "group signal generation" and then to "root cause attribution verification," it effectively overcomes the limitations of existing technologies. Compared with existing technologies that typically only focus on single-vehicle diagnostics, this solution first calculates multi-dimensional performance characteristics (spatial and energy dimensions) and analyzes their anomaly incidence rate, achieving a leap from single-vehicle level to batch level risk perception.

[0125] Furthermore, this solution innovatively introduces user behavior characteristics as control variables. By comparing the performance differences between the target batch and the normal batch under the same user habits, it can accurately verify whether the abnormal signals of the batch originate from the defects of the vehicle itself, thereby eliminating misjudgments caused by different user habits. Ultimately, it achieves highly reliable and accurate identification and confirmation of systemic common defects in battery batches.

[0126] In one feasible implementation, the method further includes: Based on the abnormal combination of specific characteristics exhibited by abnormal vehicles in the target batch, locate the type of common defect; output the type of common defect.

[0127] This step is a deepening and enhancement of the aforementioned common defect identification process, and its core purpose is to achieve root cause analysis of defects. It not only confirms the existence of defects in the batch, but also diagnoses the most likely type of root cause of the defects, thereby elevating the value of the analysis conclusions from "early warning" to "guided action".

[0128] Different types of defects (such as manufacturing defects, design defects, software defects, and hardware defects) have different physical natures and will have specific impact patterns on the multi-dimensional performance indicators of batteries. These impact patterns will be statistically presented in the form of "abnormal combination of features" in the group of abnormal vehicles.

[0129] This solution pre-establishes a mapping rule base of "defect type-feature combination". This rule base is based on battery mechanism knowledge and historical data. The localization process involves matching the dominant abnormal patterns actually exhibited in the target batch of abnormal vehicles with this rule base.

[0130] For example, the type of the common defect is located based on anomalies in the specific characteristics exhibited by abnormal vehicles in the target batch, including the following four combinations: Combination 1: If the proportion of vehicles with both first spatial dimension feature anomalies and second spatial dimension feature anomalies among the abnormal vehicles reaches the first coordination threshold, then the common defect is identified as a manufacturing defect.

[0131] The core issue of manufacturing defects (such as cell inconsistency defects) is the inherent differences in the properties (such as capacity and internal resistance) of individual cells within the battery pack. This leads to persistently large voltage differences between cells during charging (first spatial dimension characteristic: voltage fluctuation rate); simultaneously, even at the end of charging, the battery management system (BMS) struggles to effectively balance these inherent differences (second spatial dimension characteristic: voltage uniformity coefficient). The simultaneous anomalies in these two spatial dimensions are strong evidence of poor cell consistency and the inability of the BMS to effectively compensate for these differences.

[0132] Therefore, if the problem is a cell inconsistency defect, voltage fluctuation should also occur. >5% and the coefficient of uniform pressure An anomaly of >0.005.

[0133] Judgment rule: It is necessary to verify that combination 1 has statistical significance in the batch, that is, it requires... and The proportion of abnormal vehicles is ≥30%:

[0134] Furthermore, the combined proportion of the two batches is ≥60%, and p represents the distribution difference test (p<0.05 indicates a difference). Among these, the distribution difference tests, such as the KS test, confirm that the target batch and the normal batch are... and / or There are statistically significant differences in the overall distribution.

[0135] If the above conditions are met, it is confirmed that there is a risk of cell inconsistency defects, indicating poor consistency between the electrode and electrolyte, which may be caused by deviations in the manufacturing process.

[0136] Combination 2: If the proportion of vehicles with both abnormal energy dimension characteristics and abnormal spatial dimension characteristics reaches the second coordination threshold, then the common defect is identified as a design defect.

[0137] Design defects (such as cell material defects) typically stem from formulation or structural issues with the core battery materials. This affects the kinetics of lithium-ion insertion / extraction, leading to increased side reactions and internal resistance, resulting in low energy conversion efficiency (energy dimension characteristic: charging efficiency deviation). (Abnormalities); On the other hand, it will also affect the polarization voltage of the battery cell during charging, resulting in a deterioration in overall voltage stability (spatial dimension characteristic: voltage fluctuation rate). (Abnormality). An abnormal correlation between energy efficiency and voltage stability is a typical characteristic of deep-seated material or core design problems.

[0138] If the defect is in the battery cell material, the following should be present simultaneously:

[0139] Require and The proportion of abnormal vehicles was ≥30%, and the combined proportion of the two was ≥60%, p was used for distribution difference test (p<0.05). Among these, the distribution difference test, including the KS test, confirmed that the target batch and the normal batch were... and / or There are statistically significant differences in the overall distribution.

[0140] If the above conditions are met, it is confirmed that the large energy loss and significant voltage fluctuation may be due to uneven doping of the cathode material, which leads to an aggravation of side reactions and is a design-level issue.

[0141] Combination 3: If the proportion of vehicles with abnormal spatial dimension features but normal energy dimension features and other spatial dimension features reaches the third coordination threshold, then the common defect is identified as a software defect.

[0142] Software defects (such as BMS balancing algorithm defects) are characterized by logical errors. They do not directly damage the battery cell itself, nor do they affect the basic energy conversion. Their impact is concentrated on a specific function of the BMS—balancing. Therefore, it only manifests as poor balancing performance at the end of charging (a specific spatial dimension characteristic: voltage uniformity coefficient). (abnormalities), while voltage fluctuation rate reflects the dynamic process of cell consistency. and charging efficiency that reflects the health of materials All showed normal performance. This kind of "isolated" equilibrium failure is a hallmark of software algorithm problems.

[0143] If it's a defect in the BMS load balancing algorithm, it should appear separately. >0.005 (other indicators are normal).

[0144] The proportion of abnormal vehicles is ≥30% and passes the distribution difference test (p<0.05), while other indicators (e.g.) , The abnormal proportion of ) is not statistically significant.

[0145] If the above conditions are met, it is confirmed that there is a risk of defects in the BMS balancing algorithm, that there is a vulnerability in the BMS balancing logic, and that it is a software algorithm problem.

[0146] Combination 4: If the proportion of vehicles with abnormal energy dimension characteristics and normal spatial dimension characteristics among the abnormal vehicles reaches the fourth collaborative threshold, then the common defect is identified as a hardware defect.

[0147] Hardware defects (such as charging module malfunctions) typically occur on the outside of the battery pack or in peripheral components. Failures in these components (such as increased contactor resistance or reduced charger efficiency) lead to additional energy loss, directly manifesting as reduced charging efficiency (energy dimension characteristic: charging efficiency deviation). (Abnormal). However, since the battery cells themselves and the BMS are normal, the voltage consistency and stability within the battery pack (all spatial dimensions characteristics:) are good. and The system remains unaffected. This "simple" decrease in efficiency is a typical characteristic of peripheral hardware failures.

[0148] If the charging module is faulty, it should appear separately. >15% (other indicators are normal).

[0149] That is: requirements The proportion of abnormal vehicles is ≥30% and passes the distribution difference test (p<0.05), while all spatial dimension features ( and The abnormal proportions of (etc.) do not constitute statistical significance.

[0150] If the above conditions are met, it is considered that the low energy conversion efficiency but normal voltage stability may be caused by excessive internal resistance of the charging module, which is a hardware performance problem such as the internal resistance of the charging module.

[0151] In this embodiment, the distribution difference test is a crucial step in verifying whether the performance differences between batches are statistically significant. Its purpose is to confirm, from a statistical perspective, that the differences in performance indicators between the target batch and the normal batch are not caused by random fluctuations, but rather stem from a systematic and fundamental difference. This solution preferably uses the Kolmogorov-Smirnov test as the tool for implementing the distribution difference test.

[0152] The process is as follows: (1) Sample definition: Clearly define the metrics that need to be tested (e.g., voltage fluctuation rate). Let the sample size of the target batch at this performance index be: , where n is the number of vehicles in the target batch. Let be one of the performance metrics values ​​for the nth vehicle (such as the median or mean of that metric across multiple charging events). Similarly, let the reference sample for this performance metric in the normal batch be: , where m is the number of vehicles in the normal batch.

[0153] (2) Calculation of Empirical Distribution Function (EDF): Calculate the empirical distribution function of the target batch sample X :

[0154] Similarly, calculate the empirical distribution function of the normal batch sample Y. :

[0155] Where I(·) is an indicator function, which takes the value of 1 when the condition in the parentheses is met, and 0 otherwise; x is a specific value of the feature being tested, and its value needs to be based on the physical meaning and empirical range of the feature, covering the entire possible range of values ​​of the target feature.

[0156] (3) Calculation of KS statistic Calculate sequentially within the range of values ​​for the eigenvalue x. and The difference. The KS statistic D is defined as the maximum absolute difference between two empirical distribution functions:

[0157] In the formula, sup_x represents the supremum (i.e. the maximum value) for all possible values ​​of x.

[0158] (4) Significance judgment (p-value calculation) Calculate the significance p-value based on the KS statistic D and the sample size. The following approximate formula can be used:

[0159] In practical calculations, when k is between 10 and 20, the sum of the series converges well enough for most practical applications of D value to meet the accuracy requirements.

[0160] If p < 0.05, the null hypothesis that "the two samples come from the same population distribution" is rejected, indicating a statistically significant difference in the distribution of the target batch and the normal batch on this performance index. This suggests a significant difference between the target batch and the normal batch, posing a defect risk.

[0161] If p ≥ 0.05, there is insufficient evidence to reject the null hypothesis, meaning that the two batches cannot be considered to have a statistically significant distributional difference in this indicator.

[0162] Based on the same technical concept, embodiments of this application also provide a common defect detection device for batteries, such as... Figure 3 As shown, the device includes: The segmentation module 301 is used to acquire charging data of multiple vehicles in the target batch and segment charging events.

[0163] The calculation module 302 is used to calculate, for the segmented charging events, the spatial dimension features used to characterize the internal consistency of the battery and the energy dimension features used to characterize the energy conversion efficiency.

[0164] The identification module 303 is used to identify abnormal vehicles based on the anomaly rates of the spatial dimension features and the energy dimension features in multiple charging events of a single vehicle.

[0165] The determination module 304 is used to determine the proportion of the abnormal vehicles and the batch abnormal signal characterized by the coordinated occurrence of multiple dimensions of abnormal features.

[0166] The verification module 305 is used to respond to the batch abnormal signal by introducing user charging behavior characteristics as control variables, and verifying whether the batch abnormal signal originates from a defect in the vehicle itself by comparing the differences between the target batch and the preset normal batch in the spatial dimension characteristics and the energy dimension characteristics under the same user behavior characteristics.

[0167] The determination module 306 is used to determine that if the verification passes, the batteries of multiple vehicles in the target batch have a common defect.

[0168] In one feasible implementation, a determining module is used to determine the proportion of the abnormal vehicles and a batch anomaly signal representing the coordinated occurrence of multiple dimensions of anomaly features, for the following purposes: When the proportion of abnormal vehicles exceeds a first threshold, and the proportion of abnormal vehicles that simultaneously exhibit at least two abnormal features exceeds a second threshold, the batch abnormal signal is generated.

[0169] In one feasible implementation, the verification module is used to compare the differences between the target batch and the preset normal batch in the spatial dimension feature and the energy dimension feature under the same user behavior characteristics, including: All vehicles are grouped according to user behavior characteristics; these user behavior characteristics are used to characterize users' charging habits.

[0170] Within the same group, the statistical differences between the target batch and the preset normal batch in terms of the spatial dimension feature and the energy dimension feature are calculated.

[0171] If, in at least one group, the statistical difference exceeds a preset fluctuation range determined based on the data of the normal batch, the verification is deemed successful, confirming that the abnormal signal of the batch does not originate from differences in user charging habits.

[0172] In one feasible implementation, the device further includes: The positioning module is used to locate the type of the common defect based on the abnormal combination of specific characteristics exhibited by abnormal vehicles in the target batch.

[0173] The output module is used to output the types of common defects.

[0174] In one feasible implementation, the spatial dimension features include a first spatial dimension feature and / or a second spatial dimension feature.

[0175] The positioning module is used to locate the type of the common defect based on the abnormal combination of specific characteristics exhibited by abnormal vehicles in the target batch, for the following purposes: If the proportion of vehicles exhibiting both first spatial dimension feature anomalies and second spatial dimension feature anomalies reaches a first coordination threshold, then the common defect is identified as a manufacturing defect.

[0176] If the proportion of vehicles exhibiting both energy dimension anomalies and spatial dimension anomalies among the abnormal vehicles reaches the second coordination threshold, then the common defect is identified as a design defect.

[0177] If the proportion of vehicles with abnormal spatial dimension features but normal energy dimension features and other spatial dimension features reaches the third coordination threshold, then the common defect is identified as a software defect.

[0178] If the proportion of vehicles with abnormal energy dimension characteristics and normal spatial dimension characteristics among the abnormal vehicles reaches the fourth collaborative threshold, then the common defect is identified as a hardware defect.

[0179] In one feasible implementation, the first spatial dimension feature is voltage fluctuation rate; the second spatial dimension feature is voltage uniformity coefficient; and the energy dimension feature is charging efficiency deviation.

[0180] In one feasible implementation, the segmentation module is used to acquire charging data from multiple vehicles in the target batch and perform charging event segmentation, for the following purposes: Obtain charging data for multiple vehicles in the target batch, and preprocess the charging data; the preprocessing includes at least one of the following: temporal continuity verification, necessary column integrity verification, and filling missing values ​​with a missing rate below a threshold by linear interpolation.

[0181] Based on the step change of the charging state signal, independent charging event segments are identified from the preprocessed charging data.

[0182] The identified charging event segments are filtered for validity to exclude invalid events caused by momentary interference or misoperation.

[0183] Figure 4 A schematic diagram of an electronic device provided in this application embodiment includes: a processor 401, a storage medium 402, and a bus 403. The storage medium 402 stores machine-readable instructions executable by the processor 401. When the electronic device runs the common defect detection method for batteries as described in the embodiment, the processor 401 communicates with the storage medium 402 via the bus 403, and the processor 401 executes the machine-readable instructions to perform the steps as described in the embodiment.

[0184] In this embodiment, the storage medium 402 may also execute other machine-readable instructions to perform other methods as described in the embodiment. For details on the specific execution steps and principles, please refer to the description of the embodiment, which will not be repeated here.

[0185] This application also provides a computer-readable storage medium storing a computer program that is executed by a processor to perform the steps as described in the embodiments.

[0186] In this embodiment, the computer program, when run by the processor, can also execute other machine-readable instructions to perform other methods as described in the embodiments. For details on the specific execution steps and principles, please refer to the description of the embodiments, which will not be repeated here.

[0187] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Additionally, the coupling or direct coupling or communication connection shown or discussed may be through some communication interface; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0188] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0189] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0190] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0191] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for detecting common defects in batteries, characterized in that, The method includes: Obtain charging data for multiple vehicles in the target batch and segment charging events; For the segmented charging events, calculate the spatial dimension features used to characterize the internal consistency of the battery and the energy dimension features used to characterize the energy conversion efficiency; Abnormal vehicles are identified based on the anomaly rates of the spatial dimension features and the energy dimension features in multiple charging events of a single vehicle. The proportion of the abnormal vehicles and the batch abnormal signal characterized by the coordinated occurrence of multiple abnormal features were determined. In response to the batch anomaly signal, user charging behavior characteristics are introduced as control variables. By comparing the differences between the target batch and the preset normal batch in the spatial dimension characteristics and the energy dimension characteristics under the same user behavior characteristics, it is verified whether the batch anomaly signal originates from a defect in the vehicle itself. If the verification passes, it is ultimately determined that the batteries of multiple vehicles in the target batch have a common defect. The batch anomaly signal, characterized by the proportion of the abnormal vehicles and the coordinated occurrence of multiple dimensions of anomaly features, includes: When the proportion of abnormal vehicles exceeds a first threshold, and the proportion of abnormal vehicles that simultaneously exhibit at least two abnormal features exceeds a second threshold, the batch abnormal signal is generated. The method further includes: Based on the abnormal combinations of specific characteristics exhibited by abnormal vehicles in the target batch, the type of the common defect is located; Output the types of common defects; The spatial dimension features include a first spatial dimension feature and / or a second spatial dimension feature; Based on the abnormal combinations of specific characteristics exhibited by abnormal vehicles in the target batch, the type of the common defect is identified, including: If the proportion of vehicles with both first spatial dimension feature anomalies and second spatial dimension feature anomalies among the abnormal vehicles reaches the first coordination threshold, then the common defect is identified as a manufacturing defect. If the proportion of vehicles with both abnormal energy dimension characteristics and abnormal spatial dimension characteristics reaches the second coordination threshold, then the common defect is identified as a design defect. If the proportion of vehicles with abnormal spatial dimension features but normal energy dimension features and other spatial dimension features reaches the third coordination threshold, then the common defect is identified as a software defect. If the proportion of vehicles with abnormal energy dimension characteristics and normal spatial dimension characteristics among the abnormal vehicles reaches the fourth collaborative threshold, then the common defect is identified as a hardware defect. The first spatial dimension feature is voltage fluctuation rate; the second spatial dimension feature is voltage uniformity coefficient; and the energy dimension feature is charging efficiency deviation.

2. The method according to claim 1, characterized in that, By comparing the differences between the target batch and the preset normal batch in the spatial dimension features and the energy dimension features under the same user behavior characteristics, including: All vehicles are grouped according to user behavior characteristics; these user behavior characteristics are used to characterize users' charging habits. Within the same group, calculate the statistical differences between the target batch and the preset normal batch in terms of the spatial dimension feature and the energy dimension feature; If, in at least one group, the statistical difference exceeds a preset fluctuation range determined based on the data of the normal batch, the verification is deemed successful, confirming that the abnormal signal of the batch does not originate from differences in user charging habits.

3. The method according to claim 1, characterized in that, Obtain charging data for multiple vehicles in the target batch and segment charging events, including: Obtain charging data for multiple vehicles in the target batch, and preprocess the charging data; the preprocessing includes at least one of the following: temporal continuity verification, necessary column integrity verification, and filling missing values ​​with a missing rate below a threshold by linear interpolation. Based on the step change of the charging state signal, independent charging event segments are identified from the preprocessed charging data. The identified charging event segments are filtered for validity to exclude invalid events caused by momentary interference or misoperation.

4. A common defect detection device for batteries, characterized in that, The device includes: The segmentation module is used to acquire charging data of multiple vehicles in the target batch and segment charging events. The calculation module is used to calculate the spatial dimension features that characterize the internal consistency of the battery and the energy dimension features that characterize the energy conversion efficiency for the segmented charging events. The identification module is used to identify abnormal vehicles based on the anomaly rates of the spatial dimension features and the energy dimension features in multiple charging events of a single vehicle. The determination module is used to determine the proportion of the abnormal vehicles and the batch abnormal signal characterized by the coordinated occurrence of multiple dimensions of abnormal features. The verification module is used to respond to the batch abnormal signal by introducing user charging behavior characteristics as control variables. By comparing the differences between the target batch and the preset normal batch in the spatial dimension characteristics and the energy dimension characteristics under the same user behavior characteristics, the module verifies whether the batch abnormal signal originates from a defect in the vehicle itself. The determination module is used to determine that if the verification passes, the batteries of multiple vehicles in the target batch have a common defect. The batch anomaly signal, characterized by the proportion of the abnormal vehicles and the coordinated occurrence of multiple dimensions of anomaly features, includes: When the proportion of abnormal vehicles exceeds a first threshold, and the proportion of abnormal vehicles that simultaneously exhibit at least two abnormal features exceeds a second threshold, the batch abnormal signal is generated. The device further includes: The positioning module is used to locate the type of the common defect based on the abnormal combination of specific characteristics exhibited by abnormal vehicles in the target batch; The output module is used to output the types of common defects; The spatial dimension features include a first spatial dimension feature and / or a second spatial dimension feature; Based on the abnormal combinations of specific characteristics exhibited by abnormal vehicles in the target batch, the type of the common defect is identified, including: If the proportion of vehicles with both first spatial dimension feature anomalies and second spatial dimension feature anomalies among the abnormal vehicles reaches the first coordination threshold, then the common defect is identified as a manufacturing defect. If the proportion of vehicles with both abnormal energy dimension characteristics and abnormal spatial dimension characteristics reaches the second coordination threshold, then the common defect is identified as a design defect. If the proportion of vehicles with abnormal spatial dimension features but normal energy dimension features and other spatial dimension features reaches the third coordination threshold, then the common defect is identified as a software defect. If the proportion of vehicles with abnormal energy dimension characteristics and normal spatial dimension characteristics among the abnormal vehicles reaches the fourth collaborative threshold, then the common defect is identified as a hardware defect. The first spatial dimension feature is voltage fluctuation rate; the second spatial dimension feature is voltage uniformity coefficient; and the energy dimension feature is charging efficiency deviation.

5. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the steps of the common defect detection method for a battery as described in any one of claims 1 to 3.

6. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the steps of the common defect detection method for batteries as described in any one of claims 1 to 3.

Images