A method, device, and vehicle for identifying outliers in battery cell capacity.

Abstract

This invention relates to the field of battery cell testing, specifically to a method, apparatus, and vehicle for identifying outlier battery cell capacities. The method provided in this application identifies charging conditions and pre- and post-charging conditions and extracts stable resting points. It directly calculates the change in state of charge (SOC) based on the voltage and temperature of the resting point, eliminating the need for current integration and fundamentally avoiding current acquisition noise and accumulated errors. This solves the problem of existing technologies being susceptible to current integration errors. Simultaneously, the SOC is obtained by combining the resting point voltage and temperature with table lookup and bilinear interpolation, fully incorporating temperature factors to correct the open-circuit voltage, effectively compensating for the shortcomings of existing technologies that do not adequately consider temperature interference. Furthermore, outlier cells are identified through the statistical characteristics of the SOC, making the capacity anomaly determination more stable and accurate, and overall improving the reliability and environmental adaptability of outlier battery cell capacity identification.

Description

Technical Field

[0001] This invention relates to the field of battery cell testing, and specifically to a method, apparatus, and vehicle for identifying outliers in battery cell capacity. Background Technology

[0002] Traditional power battery capacity consistency assessment often uses ampere-hour integration combined with the two-point method to calculate the capacity of a single cell and thus determine capacity differences. However, this method relies on current sensor data collection, and sensor noise is easily amplified during long-term integration, resulting in significant cumulative errors in capacity calculation results and affecting the accuracy of consistency judgment.

[0003] Meanwhile, the open-circuit voltage of a lithium-ion battery is determined by the potential difference between the positive and negative electrode balances, and is significantly affected by temperature. At different temperatures, the solid-phase diffusion coefficient, interfacial reaction rate, and phase transition equilibrium point will all shift, directly changing the characteristics of the open-circuit voltage curve. Existing evaluation methods usually do not systematically compensate for temperature factors, further reducing the accuracy and reliability of capacity consistency analysis. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method, apparatus and vehicle for identifying outliers in battery cell capacity, in order to solve the problems of existing technologies being susceptible to current integration errors and not fully considering the interference of temperature on open-circuit voltage.

[0005] In a first aspect, embodiments of the present invention provide a method for identifying outliers in battery cell capacity, the method comprising:

[0006] Acquire vehicle operating condition data, and identify charging phase and stationary phase before and after the charging phase based on the operating condition data. Extract the charging start point and the charging end point from the static condition stage; The change in state of charge of each individual battery cell of the vehicle between the starting point of charging and the ending point of charging is obtained. The change in state of charge is obtained by bilinear interpolation calculation based on the lookup data of voltage and temperature of the individual battery cell at the starting point of charging and the ending point of charging. Based on the statistical characteristic values ​​associated with the change in the state of charge of the individual cells, outlier individual cells among the individual cells are identified.

[0007] Furthermore, extracting the charging start point and charging end point from the static condition stage includes: Obtain the charging start point and charging end point of the charging stage; The first candidate power-on time point is obtained from the previous resting state ... The second candidate power-on time point is obtained from the next resting state ...

[0008] Furthermore, obtaining the change in state of charge of each individual battery cell of the vehicle between the charging start point and the charging end point includes: The first voltage and first temperature of each individual battery cell of the vehicle at the charging start resting point, and the second voltage and second temperature at the charging end resting point are obtained. Using the first voltage and the first temperature, the first mapping data is obtained by querying a three-dimensional table, and bilinear interpolation is performed based on the obtained first intermediate mapping data to obtain the first state of charge value. Using the second voltage and the second temperature, the second mapping data is obtained by querying a three-dimensional table, and bilinear interpolation is performed based on the obtained second intermediate mapping data to obtain the second state of charge value. Based on the first state of charge value and the second state of charge value, the change in state of charge of the corresponding individual cell is determined.

[0009] Furthermore, the determination of outlier individual cells among the individual cells based on statistical characteristic values ​​associated with the change in state of charge of the individual cells includes: Select the smallest change in state of charge from the changes in state of charge of each individual cell. If the minimum state of charge change is greater than the preset change, the capacity outlier determination mechanism will be triggered. Based on the change in state of charge and the statistical characteristic value, calculate the outlier rate of each individual cell, and determine the individual cells whose outlier rate does not fall within the outlier rate range as the outlier individual cells; and / or, construct outlier determination conditions based on the change in state of charge and the statistical characteristic value of the individual cells, and determine the individual cells that meet the outlier determination conditions as the outlier individual cells.

[0010] Furthermore, the statistical characteristic value includes the mean change in state of charge; the calculation of the outlier rate of each individual cell based on the change in state of charge and the statistical characteristic value includes: Obtain the ratio between the mean change in state of charge and the amount of change in state of charge; The difference between the preset value and the ratio is determined as the outlier rate.

[0011] Furthermore, the statistical feature values ​​include the standard deviation of the state of charge and the mean of the state of charge; the outlier determination criteria based on the change in the state of charge of the individual cell and the statistical feature values ​​include: Obtain the difference between the change in state of charge and the mean state of charge, and obtain the standard deviation of a preset multiple; The outlier determination criteria are constructed based on the relationship between the absolute value of the difference and the standard deviation of the preset multiple.

[0012] Furthermore, the method also includes: If the absolute value of the difference is greater than the standard deviation of the preset multiple, then the outlier determination condition is satisfied; or, if the absolute value of the difference is less than or equal to the standard deviation of the preset multiple, then the outlier determination condition is not satisfied.

[0013] Furthermore, after identifying the outlier cells in the individual battery cells, the method further includes: If the number of outlier individual cells is greater than a preset number, then the spatial distribution characteristics of the outlier individual cells are obtained. The root cause of the fault in the outlier individual cell is located based on the fault detection mode corresponding to the spatial distribution characteristics.

[0014] Secondly, embodiments of the present invention provide a device for identifying outliers in battery cell capacity, the device comprising: The acquisition module is used to acquire vehicle operating condition data and identify the charging stage and the stationary stage before and after the charging stage based on the operating condition data. The extraction module is used to extract the charging start point and the charging end point from the static working condition stage. The calculation module is used to obtain the change in state of charge of each individual battery cell of the vehicle between the charging start rest point and the charging end rest point. The change in state of charge is obtained by bilinear interpolation calculation based on the lookup data of voltage and temperature of the individual battery cell at the charging start rest point and the charging end rest point. The analysis module is used to identify outlier individual cells in the individual cells based on statistical characteristic values ​​associated with the change in the state of charge of the individual cells.

[0015] Thirdly, embodiments of the present invention provide a computer device, including: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method described in the first aspect or any corresponding embodiment thereof.

[0016] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer instructions that cause a computer to perform the method described in the first aspect or any of its corresponding embodiments.

[0017] The method provided in this application identifies charging conditions and pre- and post-charging resting conditions and extracts stable resting points. It directly calculates the change in state of charge (SOC) based on the voltage and temperature of the resting point, eliminating the need for current integration. This fundamentally avoids current acquisition noise and accumulated errors, solving the problem of existing technologies being susceptible to current integration errors. Furthermore, the SOC is obtained by combining the resting point voltage and temperature with table lookup and bilinear interpolation, fully incorporating temperature factors to correct for open-circuit voltage, effectively compensating for the shortcomings of existing technologies that do not adequately consider temperature interference. Finally, the statistical characteristics of the SOC identify outlier cells, making capacity anomaly detection more stable and accurate, and overall improving the reliability and environmental adaptability of outlier cell capacity identification. Attached Figure Description

[0018] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating a method for identifying outliers in battery cell capacity according to some embodiments of the present invention. Figure 2 This is a flowchart illustrating another method for identifying outlier cell capacities according to some embodiments of the present invention; Figure 3 This is a flowchart illustrating another method for identifying outliers in battery cell capacity according to some embodiments of the present invention; Figure 4 This is a structural block diagram of a battery cell capacity outlier identification device according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] According to embodiments of the present invention, a method, apparatus, and vehicle for identifying outliers in battery cell capacity are provided. It should be noted that the steps shown in the flowcharts in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0022] This embodiment provides a method for identifying outliers in battery cell capacity. Figure 1 This is a flowchart of a method for identifying outliers in battery cell capacity according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain vehicle operating condition data, and identify the charging operating condition stage and the stationary operating condition stage before and after the charging operating condition stage based on the operating condition data.

[0023] In this embodiment, online operating condition data of the vehicle over the past month is acquired. The data is then segmented according to the vehicle's actual operating status, dividing the vehicle's operation into charging, discharging, and stationary conditions. Charging conditions are identified by a continuous charging status flag, discharging conditions by continuous discharging current, and stationary conditions consist of adjacent power-off and power-on moments. After segmenting the conditions using these rules, a complete charging phase is accurately identified from the overall operating condition sequence. Simultaneously, the immediately preceding and immediately following stationary phases are located, providing a stable and reliable operating condition basis and time boundary for subsequent selection of effective stationary points.

[0024] Step S102: Extract the charging start point and the charging end point from the static condition stage.

[0025] In this embodiment of the application, extracting the charging start point and the charging end point from the static condition stage includes: Step A1: Obtain the charging start point and charging end point of the charging operation stage.

[0026] Specifically, after completing the division of operating conditions and identifying the complete charging stage, the charging start point and the charging end point are determined based on the transition time of the charging status flag. These charging start and end points serve as key time references, used to retrieve valid power-on times within adjacent static operating conditions, ensuring a clear correspondence between the selection of static points and the charging process.

[0027] Step A2: Obtain the first candidate power-on time point from the previous resting state stage adjacent to the charging state stage, and determine the candidate power-on time point that is closest to the charging start point and whose corresponding resting state stage duration is greater than the preset duration as the charging start resting point.

[0028] Specifically, the first candidate power-on moment is obtained from the preceding resting state adjacent to the charging state. This power-on moment is accurately identified using a pre-trained 1D_CNN_LSTM temporal neural network model. This model automatically extracts features based on raw time-series signals such as current, voltage, and temperature, effectively eliminating misjudgments caused by data loss and noise interference. Among all the first candidate power-on moments, the one closest to the charging start point and corresponding to a resting state duration longer than a preset duration is selected and designated as the charging start resting point, i.e., the low SOC point. This ensures that the battery voltage is sufficiently stable at this moment, accurately reflecting the correspondence between open-circuit voltage and state of charge.

[0029] Step A3: Obtain the second candidate power-on time point from the next resting state ...

[0030] Specifically, a trained and optimized temporal neural network model is used to accurately identify the power-on moment, ensuring the authenticity and reliability of the candidate moments. Among all the second candidate power-on moments, the one closest to the end of charging and whose corresponding resting period is longer than a preset duration is selected and designated as the end-of-charge resting point, i.e., the high SOC point. The charging start and end resting points selected in this way effectively avoid polarization voltage interference, improving the accuracy of subsequent state of charge calculations and capacity consistency analysis.

[0031] Understandably, the charging start resting point represents the time point before the start of the charging operation phase, when the cell voltage and temperature are sufficiently stable and can truly reflect the relationship between open circuit voltage and temperature under low charge conditions; the charging end resting point represents the time point after the end of the charging operation phase, when the cell voltage and temperature are sufficiently stable and can truly reflect the relationship between open circuit voltage and temperature under high charge conditions.

[0032] By obtaining the charging start and end points of each charging stage, the time boundaries of the charging process are clearly defined, providing an accurate benchmark for selecting the resting point. Furthermore, by obtaining the first candidate power-on moment from the preceding resting condition adjacent to the charging condition and selecting the moment closest to the charging start point and with a resting time exceeding a preset threshold as the charging start resting point, and by obtaining the second candidate power-on moment from the subsequent resting condition and selecting the moment closest to the charging end point and with a resting time exceeding a preset threshold as the charging end resting point, it is ensured that the voltage at the selected resting point is sufficiently stable and the polarization effect is essentially eliminated. This ensures that the correspondence between the open-circuit voltage and the state of charge (SOC) is accurately reflected, providing a reliable time benchmark and voltage sampling point for subsequent high-precision SOC calculations. This effectively avoids errors introduced by insufficient resting or improper point selection, improving the accuracy and consistency of SOC inversion.

[0033] Step S103: Obtain the change in state of charge of each individual battery cell of the vehicle between the starting point and the ending point of charging. The change in state of charge is obtained by bilinear interpolation calculation based on the lookup data of voltage and temperature of the individual battery cell at the starting point and the ending point of charging.

[0034] In this embodiment of the application, obtaining the change in state of charge of each individual battery cell of the vehicle between the starting point of charging and the ending point of charging includes: Step B1: Obtain the first voltage and first temperature of each individual battery cell of the vehicle at the starting rest point of charging, and the second voltage and second temperature at the ending rest point of charging.

[0035] Specifically, based on the charging start resting point t1 and the charging end resting point t2, voltage data of all individual cells in the battery pack are collected at the corresponding times to obtain the first voltage and second voltage of each individual cell. The first voltage is the open-circuit voltage OCV1i of the individual cell at the charging start resting point t1, and the second voltage is the open-circuit voltage OCV2i of the individual cell at the charging end resting point t2. Since the above resting points all meet the sufficient resting conditions, the voltages have completed polarization recovery and can truly reflect the electrochemical equilibrium state inside the cell. The first temperature is the bulk stable temperature of the individual cell at time t1; the second temperature is the bulk stable temperature of the individual cell at time t2.

[0036] Step B2: Using the first voltage and the first temperature, query the three-dimensional table to obtain the first mapping data, and perform bilinear interpolation calculation based on the obtained first intermediate mapping data to obtain the first state of charge value.

[0037] Specifically, after obtaining the first voltage and first temperature of each individual cell at the initial resting point t1 of charging, the first voltage and first temperature are used as joint input parameters to call the pre-constructed multi-temperature layer interpolation three-dimensional OCV-SOC-Temp lookup table (non-discrete temperature point). In this three-dimensional lookup table, several discrete reference nodes that are closest to the current first voltage and first temperature are accurately matched, thereby obtaining the corresponding first mapping data. The first mapping data contains the correspondence between the voltage, temperature and state of charge at the reference node.

[0038] Subsequently, based on the obtained first mapping data, a bilinear interpolation algorithm is used for further calculations. This algorithm can perform continuous interpolation operations simultaneously in the voltage and temperature dimensions, effectively eliminating mapping errors caused by discrete reference nodes and achieving smooth fitting between non-fixed temperature points and non-fixed voltage points. Finally, the first state of charge (SOC) value of each individual cell at the initial resting point t1 after temperature dynamic compensation is obtained. This fully incorporates the influence of temperature factors on the mapping relationship between open-circuit voltage and SOC, improving the inversion accuracy of the first SOC value.

[0039] Step B3: Using the second voltage and the second temperature, query the three-dimensional table to obtain the second mapping data, and perform bilinear interpolation calculation based on the obtained second intermediate mapping data to obtain the second state of charge value.

[0040] Specifically, using the same implementation logic as step B2, after obtaining the second voltage and second temperature corresponding to the resting point t2 at the end of charging for each individual cell, the second voltage and second temperature are used as joint input parameters to query the same pre-constructed multi-temperature layer interpolation three-dimensional OCV-SOC-Temp lookup table. The discrete reference node that matches the current second voltage and second temperature is located in the three-dimensional lookup table, and the corresponding second mapping data is extracted. The second mapping data also contains the corresponding correlation information of voltage, temperature and state of charge at the reference node.

[0041] Subsequently, based on the obtained second mapping data, bilinear interpolation calculations are performed. Through continuous interpolation in the voltage and temperature dimensions, the state of charge (POC) values ​​corresponding to the current second voltage and second temperature are smoothly fitted, ultimately obtaining the second POC value of each individual cell at the end of charging and resting point t2 after dynamic temperature compensation. This takes into account the influence of temperature on the cell's open-circuit voltage. By combining three-dimensional lookup tables with bilinear interpolation, the POC calculation deviation caused by temperature shifts is effectively avoided, ensuring the accuracy of the second POC value and establishing a correspondence with the first POC value.

[0042] Step B4: Based on the first state of charge value and the second state of charge value, determine the change in state of charge of the corresponding individual cell.

[0043] Specifically, the calculation formula △SOC is used. i =SOC 2i SOC 1i The difference in state of charge (SOC) before and after charging is taken as the change in SOC of a single cell during this effective charging process. The calculation of the change in SOC of all single cells is completed sequentially, and the changes in SOC of all single cells are combined into a set ΔSOC for subsequent statistical feature calculations, capacity outlier analysis, and dynamic threshold determination.

[0044] Understandably, when constructing a multi-temperature compensated OCV-SOC three-dimensional lookup table, the first step is to obtain the open-circuit voltage and state of charge (SOC) data of lithium-ion batteries under different temperature conditions. This breaks through the limitations of traditional methods that only use discrete temperature points and establishes a three-dimensional continuous mapping model consisting of open-circuit voltage (OCV), SOC, and temperature (Temp). Interpolation is then performed between multiple temperature layers to form a continuous temperature field covering the entire temperature range, rather than fixed temperature nodes. Finally, a bilinear interpolation algorithm is used to smoothly fit the relationship between voltage, temperature, and SOC, thereby forming a high-precision OCV-SOC-Temp three-dimensional lookup table applicable to any temperature. This provides a unified and accurate lookup benchmark for SOC inversion at resting points.

[0045] refer to Figure 2 , Figure 2 The graph shows the SOC-OCV relationship curves at different temperatures, with SOC (State of Charge) on the horizontal axis and OCV (Open Circuit Voltage) on the vertical axis. Different colored curves represent the SOC-OCV mapping relationship under different temperature conditions. The curve trend shows that OCV generally increases with SOC, and at the same SOC, there are significant differences in OCV at different temperatures, demonstrating the significant impact of temperature on the mapping relationship between battery open circuit voltage and state of charge. The distribution and offset of each temperature curve within the SOC range intuitively reflect the necessity of constructing a multi-temperature-layer three-dimensional OCV-SOC-Temp lookup table, and also confirm the rationality of using bilinear interpolation for temperature compensation to improve the accuracy of SOC inversion.

[0046] By acquiring the first voltage and temperature of each individual cell at the initial resting point of charging, and the second voltage and temperature at the end resting point of charging, accurate and synchronous raw data were collected, ensuring the basic quality of subsequent calculations. First mapping data was obtained by querying a pre-constructed three-dimensional OCV-SOC-Temp table using the first voltage and temperature, and the first state of charge (SOC) value was calculated using bilinear interpolation based on the first mapping data. The second SOC value was obtained by performing the same processing using the second voltage and temperature. This achieved high-precision SOC inversion at any temperature and voltage point, overcoming the limitations of traditional discrete temperature point lookup tables and eliminating calculation errors caused by temperature shifts. By calculating the change in SOC based on the first and second SOC values, the SOC increment of each cell during this charging process was accurately quantified, providing an accurate and reliable data foundation for subsequent capacity consistency assessment and outlier identification.

[0047] Step S104: Based on the statistical characteristic values ​​associated with the change in state of charge of individual cells, identify outlier individual cells among the individual cells.

[0048] In this embodiment of the application, the outlier individual cells in a single battery cell are determined based on statistical characteristic values ​​associated with the change in state of charge of the individual cells, including the following steps C1-C5: Step C1: Select the smallest change in state of charge from the changes in state of charge of each individual cell.

[0049] Specifically, after obtaining the state of charge (SOC) changes of each individual cell in the battery pack (i.e., the SOC difference between the starting and ending points of charging for each cell, denoted as ΔSOC_i), these values ​​are sorted or iterated and compared to select the smallest SOC change, denoted as ΔSOC_min. The SOC change refers to the difference in SOC of each individual cell before and after charging, calculated based on bilinear interpolation, reflecting the relative change in the amount of charge received by that cell during the charging process. For example, assuming a battery pack contains 5 individual cells, and the calculated SOC changes are: ΔSOC_1=35%, ΔSOC_2=38%, ΔSOC_3=42%, ΔSOC_4=40%, ΔSOC_5=36%, then the minimum value is selected, i.e., ΔSOC_min=35% (corresponding to cell 1). This minimum value will be compared with a preset change (e.g., 60%) to determine whether the current calculation is valid.

[0050] Step C2: If the change in minimum state of charge is greater than the preset change, the capacity outlier determination mechanism will be triggered.

[0051] Specifically, the minimum change in state of charge (SOC) (i.e., the minimum SOC change ΔSOC_i among all individual cells, denoted as ΔSOC_min) is first compared with a preset threshold. This preset threshold is a pre-defined value (e.g., 60%) used to measure whether the SOC change in this charging event is significant enough to ensure the reliability of subsequent outlier detection. Only when ΔSOC_min is greater than the preset threshold is it considered that the SOC change of all cells during this charging process is large enough to effectively reflect the capacity differences between cells, thus triggering the capacity outlier detection mechanism and continuing to perform subsequent outlier rate calculation and outlier cell identification. Conversely, if ΔSOC_min is less than or equal to the preset threshold, it indicates that the charging amplitude is insufficient and the data validity is low. At this time, the subsequent processing flow is terminated to avoid misjudgment or invalid analysis caused by insufficient charging.

[0052] As an example, assume the preset change rate is set to 60%. In a certain charging event, the battery pack contains 5 individual cells, and the calculated changes in state of charge are ΔSOC1=62%, ΔSOC2=65%, ΔSOC3=58%, ΔSOC4=61%, and ΔSOC5=63%, respectively. The minimum value is ΔSOC_min=58%. Since 58% < 60%, the trigger condition is not met, so this charging data is invalid, and no outlier detection is performed. If another set of data is ΔSOC1=65%, ΔSOC2=68%, ΔSOC3=63%, ΔSOC4=67%, and ΔSOC5=66%, and the minimum value is 63% > 60%, then the capacity outlier detection mechanism is triggered, and subsequent steps are executed.

[0053] Step C3: Based on the change in state of charge and statistical characteristic values, calculate the outlier rate of each individual cell, and identify individual cells whose outlier rates do not fall within the outlier rate range as outlier individual cells; and / or, construct outlier determination conditions based on the change in state of charge and statistical characteristic values ​​of individual cells, and identify individual cells that meet the outlier determination conditions as outlier individual cells.

[0054] In this application embodiment, the method for determining outlier individual cells may include the following three cases: Scenario 1: Based on the change in state of charge and statistical characteristic values, calculate the outlier rate of each individual cell, and identify the individual cells whose outlier rate does not fall within the outlier rate range as outlier individual cells.

[0055] Specifically, the statistical characteristic values ​​include the mean change in state of charge (SOC). Based on the SOC and the statistical characteristic values, the outlier rate of each individual cell is calculated, including: obtaining the ratio between the mean SOC and the SOC; and determining the difference between the preset value and the ratio as the outlier rate.

[0056] Based on the state-of-charge (SOC) changes of all individual cells (i.e., the SOC difference between the starting and ending points of charging for each cell, denoted as ΔSOC_i), the mean SOC change (i.e., the arithmetic mean of all ΔSOC_i, denoted as ΔSOC_μ) is calculated. Then, for each individual cell, the outlier rate is calculated using the formula: Capacity_Outlier_Rate = 1 - (ΔSOC_μ / ΔSOC_i). The preset value here is 1. Subtracting the ratio of the mean to the individual value from 1 gives the degree to which the cell's capacity deviates from the group average. The physical meaning of this outlier rate is: if the calculated result is greater than 0, it means that the cell's ΔSOC_i is less than the mean, indicating that its capacity is below the average level; if the calculated result is less than 0, it means that ΔSOC_i is greater than the mean, indicating that its capacity is above the average level. Subsequently, the outlier rate can be determined by a preset range (such as [-0.1, 0.1]). If the outlier rate of a certain cell exceeds this range, it will be marked as a capacity outlier single cell.

[0057] Specifically, individual cells whose outlier rate does not fall within the outlier rate range are identified as outlier individual cells. This includes: comparing the outlier rate of each individual cell (i.e., Capacity_Outlier_Rate=1-ΔSOC_μ / ΔSOC_i) with a preset outlier rate range (e.g., [-0.1, 0.1]). If the outlier rate falls outside this range, the cell is determined to be an outlier individual cell.

[0058] As an example, assume the battery pack contains four individual cells. The state of charge changes calculated in step B4 are ΔSOC1=62%, ΔSOC2=65%, ΔSOC3=58%, and ΔSOC4=61%, respectively. The mean ΔSOC_μ = (62+65+58+61) / 4 = 61.5%. The outlier rates for each cell are calculated as follows: Cell 1: 1 - 61.5 / 62 ≈ 0.0081; Cell 2: 1 - 61.5 / 65 ≈ 0.0538; Cell 3: 1 - 61.5 / 58 ≈ -0.0603; Cell 4: 1 - 61.5 / 61 ≈ -0.0082. If the preset outlier rate range is [-0.1, 0.1], then the outlier rates of all cells fall within this range, and there are no outlier cells. If a certain cell has a ΔSOC_i of 50% and a mean of 61.5%, then its outlier rate = 1 - 61.5 / 50 = -0.23, which is less than -0.1. This means that the cell's capacity is 23% higher than the average level and it is identified as an outlier cell.

[0059] By obtaining the ratio between the mean change in state of charge (SOC) and the change in SOC of each cell, the relative magnitude of the SOC increment of a single cell relative to the average level of the group is intuitively reflected. By determining the difference between a preset value and this ratio as the outlier rate, a dimensionless capacity deviation index is constructed. The positive and negative values ​​of this index correspond to the cell capacity being lower or higher than the average level, respectively, and its absolute value quantifies the degree of deviation. This provides a clear and interpretable quantitative basis for subsequent outlier determination based on a fixed threshold, simplifies the calculation process, and has a clear physical meaning.

[0060] Scenario 2: Construct outlier determination criteria based on the change in state of charge and statistical characteristic values ​​of individual cells, and identify individual cells that meet the outlier determination criteria as outlier individual cells.

[0061] Specifically, the statistical characteristic values ​​include the standard deviation of the state of charge (SOC) and the mean SOC. Outlier detection criteria are constructed based on the change in SOC of a single cell and the statistical characteristic values, including: obtaining the difference between the change in SOC and the mean SOC, and obtaining the standard deviation of a preset multiple; and constructing outlier detection criteria based on the relationship between the absolute value of the difference and the standard deviation of the preset multiple.

[0062] Statistical characteristic values ​​are calculated based on the state-of-charge (SOC) change (ΔSOC_i) of all individual cells, including the mean SOC_μ and the standard deviation of SOC_σ (the standard deviation reflects the dispersion of ΔSOC_i relative to the mean for all cells). Then, an outlier determination criterion is constructed: for each individual cell, the absolute value of the difference between its ΔSOC_i and the mean ΔSOC_μ is calculated. If this absolute value is greater than a preset multiple (e.g., 3 times) of the standard deviation, i.e., |ΔSOC_i - ΔSOC_μ| > 3 × ΔSOC_σ, then the cell is marked as an outlier. This criterion is based on the Raida criterion (3σ principle) in statistics, which considers cells deviating from the mean by more than 3 times the standard deviation as outliers. A positive ΔSOC_i - ΔSOC_μ indicates that the cell's capacity is lower than the average capacity, while a negative ΔSOC_μ indicates that the capacity is higher than the average capacity. This dynamic threshold method can automatically adjust the judgment boundary according to the current data distribution, avoiding misjudgments caused by fixed thresholds.

[0063] Specifically, if the absolute value of the difference is greater than a preset multiple of the standard deviation, then the outlier determination condition is met; or if the absolute value of the difference is less than or equal to a preset multiple of the standard deviation, then the outlier determination condition is not met.

[0064] Based on the outlier determination criteria, for each individual cell, the absolute value of the difference between its state of charge change ΔSOC_i and the mean ΔSOC_μ of all cells is calculated. If the absolute value is greater than the state of charge standard deviation ΔSOC_σ by a preset multiple (such as 3 times), then the cell is determined to be an outlier individual cell.

[0065] As an example, suppose the battery pack has 6 individual cells, and the calculated state of charge (SOC) changes are: ΔSOC = [62, 65, 58, 61, 70, 63] (unit: %). The mean ΔSOC_μ = (62 + 65 + 58 + 61 + 70 + 63) / 6 ≈ 63.17%, and the standard deviation ΔSOC_σ ≈ 3.97%. Taking a preset multiplier of 3, the lower limit of the judgment threshold is 63.17 - 3 × 3.97 = 51.26%, and the upper limit is 63.17 + 3 × 3.97 = 75.08%. All cells' ΔSOC_i are within the range of 51.26% to 75.08%, therefore there are no outlier cells. If we change the ΔSOC_i of one of the cells to 80%, then 80% > 75.08%, which satisfies |80-63.17|=16.83>3×3.97=11.91. Therefore, this cell is identified as an outlier, and since ΔSOC_i-ΔSOC_μ is positive, it indicates that its capacity is below the average level.

[0066] By comparing the absolute value of the difference with the standard deviation of a preset multiple, if the absolute value is greater than the standard deviation of the preset multiple, the outlier condition is determined to be met; if it is less than or equal to the standard deviation, it is not met. This clarifies the specific mathematical rules for outlier determination, making the determination process objective and reproducible. This comparison rule is based on the Laida criterion in statistics and can effectively identify abnormal cells that deviate from the population distribution beyond the normal fluctuation range. At the same time, the adjustability of the preset multiple adapts to the detection sensitivity requirements of different application scenarios, providing a clear and quantitative execution basis for dynamic threshold outlier detection.

[0067] Scenario 3: Based on the change in state of charge and statistical characteristic values, calculate the outlier rate of each individual cell, and construct outlier determination conditions based on the change in state of charge and statistical characteristic values ​​of the individual cells. Individual cells whose outlier rates do not fall within the outlier rate range and meet the outlier determination conditions are identified as outlier individual cells.

[0068] Specifically, the statistical features include the standard deviation of the state of charge (SOC) and the mean SOC. First, the outlier rate for each individual cell is calculated using the method in Case 1: Outlier rate = 1 - (ΔSOC_μ / ΔSOC_i), where ΔSOC_μ is the arithmetic mean of the SOC changes of all cells, and ΔSOC_i is the SOC change of the i-th cell. The outlier rate is compared with a preset outlier rate interval (e.g., [-0.1, 0.1]), and cells whose outlier rates fall outside this interval are selected. Simultaneously, outlier criteria are constructed using the method in Case 2: For each individual cell, the absolute value of the difference between its SOC change ΔSOC_i and the mean ΔSOC_μ is calculated. If this absolute value is greater than a preset multiple (e.g., 3 times) of the SOC standard deviation ΔSOC_σ, then the outlier criteria are met. Ultimately, individual cells that simultaneously meet the conditions of "outlier rate not falling within the outlier rate range" and "absolute value of the difference greater than a preset multiple of the standard deviation" are identified as outlier cells. This joint determination method combines the advantages of fixed thresholds (outlier rate range) and dynamic thresholds (3σ principle), enabling more reliable identification of cells with significantly abnormal capacity and reducing the possibility of misjudgment or omission by a single method.

[0069] By selecting the minimum value from the state-of-charge (SOC) changes of each individual cell, a lower limit for the SOC change in this charging event was obtained to evaluate the validity of the data. If this minimum value is greater than a preset change value, a capacity outlier detection mechanism is triggered, filtering out events with insufficient charging or poor data quality, thus avoiding invalid calculations and misjudgments. The outlier rate of each cell was calculated based on the SOC changes and their statistical characteristics, quantifying the degree to which each cell's capacity deviates from the group average. Dynamic anomaly identification rules were established by constructing outlier detection conditions based on the SOC changes and their statistical characteristics. Outlier detection based on a fixed threshold was achieved by identifying cells whose outlier rates did not fall within a preset range. Outlier detection based on dynamic statistics was achieved by identifying cells that met the outlier detection conditions. Joint detection based on static thresholds and dynamic statistics was achieved by identifying cells that simultaneously met both of the above conditions. The combination of the above-mentioned multiple judgment methods can be flexibly selected or comprehensively applied according to the actual application scenario, which significantly improves the accuracy and robustness of outlier identification, ensuring that both obviously deviated abnormal cells can be detected and potential problems at the distribution edge can be captured. At the same time, the risk of misjudgment and missed judgment is further reduced through joint judgment.

[0070] In the embodiments of this application, after identifying the outlier cells in a single battery cell, as follows: Figure 3 As shown, the method also includes: Step S201: If the number of outlier cells is greater than a preset number, then obtain the spatial distribution characteristics of the outlier cells.

[0071] Specifically, when the number of identified outlier cells exceeds a preset number (usually 1, meaning multiple cells are out of the group simultaneously), the first step is to obtain a list of these outlier cells' serial numbers, i.e., their physical location numbers within the battery pack (e.g., cell 1, cell 2, ..., cell n). Then, spatial distribution characteristic analysis is performed on these serial numbers to observe whether they exhibit specific regular patterns, such as whether they are continuously concentrated within a certain module area, whether they are evenly spaced, whether they are symmetrically distributed within the battery pack (e.g., the first and last cells), or whether all are odd or even serial numbers. These distribution characteristics can reflect the type of potential fault and provide crucial clues for fault root cause localization in the subsequent step S202.

[0072] Step S202: Locate the root cause of the fault in the outlier individual cells based on the fault detection mode corresponding to the spatial distribution characteristics.

[0073] Specifically, based on the spatial distribution characteristics of the outlier cell serial number list, it is matched with various predefined fault detection modes to locate the root cause of the fault. Spatial distribution characteristics refer to the regularity of the outlier cell serial numbers in the physical arrangement of the battery pack, such as whether they are concentrated within the same module (e.g., IDs 5-8 consecutively), whether they are evenly spaced (e.g., 1, 5, 9), whether they are symmetrically distributed (e.g., the first and last cells 1 and 12), or whether they are all odd / even serial numbers. Based on the matching results, corresponding fault detection actions are triggered: if the outlier cells belong to the same module, module-level fault detection is triggered to check the consistency or connection problems of the cells within the module; if the serial numbers are evenly spaced, it is determined to be a sampling channel synchronization fault, requiring calibration of the corresponding channel of the ADC (analog-to-digital converter) on the BMS sampling board; if they are symmetrically distributed (e.g., the first and last cells), it indicates uneven heat dissipation or uneven balancing current, requiring optimization of thermal management and balancing circuits; if all are odd or all are even, it is determined to be an imbalance in parallel branches, requiring checking the impedance consistency of parallel branches. This pattern matching method enables rapid identification of specific fault causes from outlier phenomena.

[0074] By acquiring the spatial distribution characteristics of outlier cells when the number exceeds a preset limit, the abnormal information of individual cells is analyzed at the level of the overall battery pack layout. This clarifies the potential patterns in the physical locations of multiple outlier cells (such as being in the same module, equally spaced, symmetrically distributed, or having odd or even distribution). By matching the corresponding fault detection mode based on the spatial distribution characteristics to locate the root cause of the fault, a closed loop from anomaly detection to cause diagnosis is achieved. This enables rapid differentiation of different fault types, such as module-level faults, sampling channel faults, heat dissipation problems, or parallel branch imbalances. This provides clear guidance for subsequent precise maintenance and optimized design, significantly improving the diagnosability and operational efficiency of the battery system.

[0075] In one embodiment of this application, the method further includes: obtaining a list of serial numbers of outlier individual cells and the quantity of outlier individual cells; if the quantity of outlier individual cells is equal to 1, then determining that the outlier individual cell is a single cell outlier; based on the determination result of the single cell outlier, determining that the outlier individual cell has an abnormality, the abnormality type including large self-discharge or abnormal cell capacity; and outputting the serial number of the outlier individual cell and the corresponding fault cause prompt.

[0076] Specifically, after outlier detection, the process first obtains the serial numbers of all cells marked as outliers, forming an outlier cell serial number list. The number of elements in this list is then counted to determine the number of outlier cells. Next, the number of outlier cells is compared to a preset quantity of 1. If the quantity equals 1, the single-cell outlier processing flow begins, and the serial number of that outlier cell is extracted. Then, based on the characteristics of outlier cells, it is inferred that the cell may have an abnormal self-discharge rate (e.g., excessive self-discharge due to a micro-short circuit) or a significant deviation between the cell's capacity and its factory specifications (e.g., manufacturing defects or abnormal aging), causing its state of charge change to deviate from the group average, thus being identified as an outlier. Finally, the serial number of the outlier cell and the corresponding fault cause are output, allowing maintenance personnel to focus on inspecting or replacing that cell.

[0077] As an example, suppose a battery pack has 12 cells. After identification, the list of outlier cells is as follows [7], that is, only cell number 7 is marked as an outlier, and the number of outlier cells is equal to 1. Then, the single cell outlier processing procedure is executed to extract cell number 7. Based on the reason for the single cell outlier, it is inferred that the cell may have a large self-discharge (for example, the voltage drops rapidly when the cell is idle due to an internal micro-short circuit) or abnormal cell capacity (for example, the capacity is significantly lower than that of other cells). The diagnostic information is output: "Cell number 7 is an outlier cell. It is recommended to check the self-discharge and cell capacity." Based on this, maintenance personnel can perform offline testing on the cell to further confirm the cause of the fault.

[0078] In one embodiment of this application, the method further includes: acquiring outlier cell trigger records for the current period, and statistically analyzing the trigger count, trigger consistency ratio, and ambient temperature level corresponding to the trigger time for each outlier cell in the current period to obtain outlier cell statistical data for the current period. Verification is performed based on the outlier cell statistical data, specifically including: verifying whether the following conditions are simultaneously met: monthly trigger count > monthly trigger count threshold, trigger consistency ≥ trigger consistency threshold, and ambient temperature coverage of high, medium, and low levels, to obtain a validity determination result.

[0079] Then, the average outlier rate for past historical periods is obtained. The current month's outlier rate is compared with this average to determine if the deviation exceeds the outlier rate deviation threshold, thus obtaining the historical baseline comparison result. Finally, the validity determination result is combined with the historical baseline comparison result to determine whether the capacity outlier identification result for this month is valid. If the baseline deviation exceeds the outlier rate deviation threshold, manual review is triggered.

[0080] As an example, taking the capacity outlier identification verification of a certain power battery pack in a certain month as an example, firstly, the trigger records of all outlier cells in that period are retrieved, and statistics are carried out for cell B: this cell was marked as an outlier cell by the system 10 times in that month. When calculating its trigger consistency ratio, it was found that 9 out of the 10 trigger records were determined to be the same type of capacity outlier problem, with a trigger consistency ratio of 90%. At the same time, the ambient temperature data at the time of each trigger is extracted, covering -15℃ (low temperature range), 22℃ (medium temperature range), and 38℃ (high temperature range), respectively. Finally, complete statistical data of outlier cells for cell B in that month, such as the number of triggers, trigger consistency ratio, and temperature range coverage, are obtained.

[0081] The validity verification was conducted based on the statistical data of cell B. The pre-set monthly trigger frequency threshold was 3 times, the trigger consistency threshold was 80%, and the ambient temperature was required to cover high, medium, and low levels. The verification showed that cell B's monthly trigger frequency of 10 times (>3 times), trigger consistency of 90% (≥80%), and the temperature at the trigger time completely covering high, medium, and low levels all met the three verification conditions. Therefore, the validity judgment result of cell B's outlier triggers this month was determined to be "valid". If cell C only triggered 2 times in the month (below the 3-times threshold), even if the other conditions were met, its validity judgment result would still be "invalid".

[0082] The outlier rate deviation threshold is preset to 50%. First, the outlier rate data of the battery pack in the past two cycles is retrieved, and the historical average outlier rate is calculated to be 6%. Then, the outlier rate of cell B this month is 10%. The deviation is calculated using the formula (this month's outlier rate - historical average) / historical average × 100%, which is (10% - 6%) / 6% × 100% ≈ 66.7%. This deviation value exceeds the 50% outlier rate deviation threshold. Therefore, the historical baseline comparison result of cell B is "deviation exceeds the limit". If the outlier rate of cell B this month is 8%, the calculated deviation is (8% - 6%) / 6% × 100% ≈ 33.3%, which does not exceed the threshold. Therefore, the comparison result is "deviation does not exceed the limit".

[0083] The validity determination result is combined with the historical baseline comparison result for a comprehensive judgment: The validity determination result of cell B is "valid", but the historical baseline comparison result is "deviation exceeds the limit". Therefore, the capacity outlier identification result of cell B this month needs to be manually reviewed, and the technicians shall check whether the cell has problems such as actual capacity decay or abnormal sampling data; If the validity determination of cell D is "valid" and the historical baseline comparison result is "deviation does not exceed the limit", then the capacity outlier identification result of cell D this month is determined to be valid and no manual review is required; If the validity determination of cell C is "invalid", regardless of the baseline comparison result, the outlier identification result of cell C this month is determined to be invalid and no manual review is triggered.

[0084] This embodiment also provides a device for identifying outliers in battery cell capacity. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0085] This embodiment provides a device for identifying outliers in battery cell capacity, such as... Figure 4 As shown, it includes: The acquisition module 401 is used to acquire the vehicle's operating condition data and identify the charging condition stage and the stationary condition stage before and after the charging condition stage based on the operating condition data. Extraction module 402 is used to extract the charging start point and the charging end point from the static condition stage. The calculation module 403 is used to obtain the change in state of charge of each individual cell of the vehicle between the starting point of charging and the ending point of charging. The change in state of charge is obtained by bilinear interpolation calculation based on the lookup data of voltage and temperature of the individual cell at the starting point of charging and the ending point of charging. Analysis module 404 is used to identify outlier individual cells in a single cell based on statistical characteristic values ​​associated with changes in the state of charge of individual cells.

[0086] In this embodiment, the extraction module 402 is used to obtain the charging start point and charging end point of the charging condition stage; obtain a first candidate power-on time point from the previous static condition stage adjacent to the charging condition stage, and determine the candidate power-on time point that is closest to the charging start point and whose corresponding static condition stage duration is greater than a preset duration as the charging start static point; obtain a second candidate power-on time point from the next static condition stage adjacent to the charging condition stage, and determine the second candidate power-on time point that is closest to the charging end point and whose corresponding static condition stage duration is greater than a preset duration as the charging end static point.

[0087] In this embodiment, the calculation module 403 is used to obtain the first voltage and first temperature of each individual battery cell in the vehicle at the starting point of charging and the second voltage and second temperature at the ending point of charging; using the first voltage and first temperature, it queries a three-dimensional table to obtain first mapping data, and performs bilinear interpolation calculation based on the obtained first intermediate mapping data to obtain a first state of charge value; using the second voltage and second temperature, it queries a three-dimensional table to obtain second mapping data, and performs bilinear interpolation calculation based on the obtained second intermediate mapping data to obtain a second state of charge value; based on the first state of charge value and the second state of charge value, it determines the change in state of charge of the corresponding individual battery cell.

[0088] In this embodiment, the analysis module 404 is used to select the minimum state-of-charge change from the state-of-charge changes of each individual cell; if the minimum state-of-charge change is greater than a preset change, the capacity outlier determination mechanism is triggered; based on the state-of-charge change and statistical characteristic values, the outlier rate of each individual cell is calculated, and the individual cells whose outlier rate does not fall within the outlier rate range are determined as outlier individual cells; and / or, based on the state-of-charge change and statistical characteristic values ​​of the individual cells, outlier determination conditions are constructed, and the individual cells that meet the outlier determination conditions are determined as outlier individual cells.

[0089] In this embodiment of the application, the statistical feature value includes the mean change in state of charge; the analysis module 404 is used to obtain the ratio between the mean change in state of charge and the amount of change in state of charge; and the difference between the preset value and the ratio is determined as the outlier rate.

[0090] In this embodiment of the application, the statistical feature values ​​include the standard deviation of the state of charge and the mean of the state of charge; the analysis module 404 is used to obtain the difference between the change in the state of charge and the mean of the state of charge, and to obtain the standard deviation of a preset multiple; based on the relationship between the absolute value of the difference and the standard deviation of the preset multiple, outlier determination conditions are constructed.

[0091] In the embodiments of this application, if the absolute value of the difference is greater than a preset multiple of the standard deviation, it is determined that the outlier determination condition is met; or, if the absolute value of the difference is less than or equal to a preset multiple of the standard deviation, it is determined that the outlier determination condition is not met.

[0092] In this embodiment of the application, the fault analysis module is used to obtain the spatial distribution characteristics of the outlier cells if the number of outlier cells is greater than a preset number; and to locate the root cause of the outlier cells according to the fault detection mode corresponding to the spatial distribution characteristics.

[0093] Please see Figure 5 , Figure 5This is a schematic diagram of the structure of a computer device provided in an optional embodiment of the present invention, such as... Figure 5 As shown, the computer device includes one or more processors 10, memory 20, and interfaces for connecting the components, including high-speed interfaces and low-speed interfaces. The components communicate with each other via different buses and can be mounted on a common motherboard or otherwise installed as needed. The processors can process instructions executed within the computer device, including instructions stored in or on memory to display graphical information of a GUI on external input / output devices (such as display devices coupled to the interfaces). In some alternative implementations, multiple processors and / or multiple buses can be used with multiple memories and multiple memory modules, if desired. Similarly, multiple computer devices can be connected, each providing some of the necessary operations (e.g., as a server array, a group of blade servers, or a multiprocessor system).

[0094] Processor 10 may be a central processing unit, a network processor, or a combination thereof. Processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The programmable logic device may be a complex programmable logic device (CAMP), a field-programmable gate array (FPGA), a general-purpose array logic (GDA), or any combination thereof.

[0095] The memory 20 stores instructions executable by at least one processor 10 to cause at least one processor 10 to perform the method shown in the above embodiments.

[0096] The memory 20 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the computer device as shown by a landing page for an app. Furthermore, the memory 20 may include high-speed random access memory and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some alternative embodiments, the memory 20 may optionally include memory remotely located relative to the processor 10, which can be connected to the computer device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0097] The memory 20 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0098] The computer device also includes a communication interface 30 for communicating with other devices or communication networks.

[0099] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods shown in the above embodiments.

[0100] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for identifying outliers in battery cell capacity, characterized in that, The method includes: Acquire vehicle operating condition data, and identify charging phase and stationary phase before and after the charging phase based on the operating condition data. Extract the charging start point and the charging end point from the static condition stage; The change in state of charge of each individual battery cell of the vehicle between the starting point of charging and the ending point of charging is obtained. The change in state of charge is obtained by bilinear interpolation calculation based on the lookup data of voltage and temperature of the individual battery cell at the starting point of charging and the ending point of charging. Based on the statistical characteristic values ​​associated with the change in the state of charge of the individual cells, outlier individual cells among the individual cells are identified.

2. The method according to claim 1, characterized in that, Extracting the charging start point and charging end point from the static condition stage includes: Obtain the charging start point and charging end point of the charging stage; The first candidate power-on time point is obtained from the previous resting state ... The second candidate power-on time point is obtained from the next resting state ...

3. The method according to claim 1, characterized in that, The step of obtaining the change in state of charge of each individual battery cell of the vehicle between the charging start point and the charging end point includes: The first voltage and first temperature of each individual battery cell of the vehicle at the charging start resting point, and the second voltage and second temperature at the charging end resting point are obtained. Using the first voltage and the first temperature, the first mapping data is obtained by querying a three-dimensional table, and bilinear interpolation is performed based on the obtained first intermediate mapping data to obtain the first state of charge value. Using the second voltage and the second temperature, the second mapping data is obtained by querying a three-dimensional table, and bilinear interpolation is performed based on the obtained second intermediate mapping data to obtain the second state of charge value. Based on the first state of charge value and the second state of charge value, the change in state of charge of the corresponding individual cell is determined.

4. The method according to claim 1, characterized in that, The method of identifying outlier cells among the individual cells based on statistical characteristic values ​​associated with the change in state of charge of the individual cells includes: Select the smallest change in state of charge from the changes in state of charge of each individual cell. If the minimum state of charge change is greater than the preset change, the capacity outlier determination mechanism will be triggered. Based on the change in state of charge and the statistical characteristic value, calculate the outlier rate of each individual cell, and determine the individual cells whose outlier rate does not fall within the outlier rate range as the outlier individual cells; and / or, construct outlier determination conditions based on the change in state of charge and the statistical characteristic value of the individual cells, and determine the individual cells that meet the outlier determination conditions as the outlier individual cells.

5. The method according to claim 4, characterized in that, The statistical characteristic value includes the mean change in state of charge; the calculation of the outlier rate of each individual cell based on the change in state of charge and the statistical characteristic value includes: Obtain the ratio between the mean change in state of charge and the amount of change in state of charge; The difference between the preset value and the ratio is determined as the outlier rate.

6. The method according to claim 4, characterized in that, The statistical feature values ​​include the standard deviation of the state of charge and the mean of the state of charge; the outlier determination criteria based on the change in the state of charge of the individual cell and the statistical feature values ​​include: Obtain the difference between the change in state of charge and the mean state of charge, and obtain the standard deviation of a preset multiple; The outlier determination criteria are constructed based on the relationship between the absolute value of the difference and the standard deviation of the preset multiple.

7. The method according to claim 6, characterized in that, The method further includes: If the absolute value of the difference is greater than the standard deviation of the preset multiple, then the outlier determination condition is satisfied; or, if the absolute value of the difference is less than or equal to the standard deviation of the preset multiple, then the outlier determination condition is not satisfied.

8. The method according to claim 1, characterized in that, After identifying the outlier cells in the individual cells, the method further includes: If the number of outlier individual cells is greater than a preset number, then the spatial distribution characteristics of the outlier individual cells are obtained. The root cause of the fault in the outlier individual cell is located based on the fault detection mode corresponding to the spatial distribution characteristics.

9. A device for identifying outliers in battery cell capacity, characterized in that, The device includes: The acquisition module is used to acquire vehicle operating condition data and identify the charging stage and the stationary stage before and after the charging stage based on the operating condition data. The extraction module is used to extract the charging start point and the charging end point from the static working condition stage. The calculation module is used to obtain the change in state of charge of each individual battery cell of the vehicle between the charging start rest point and the charging end rest point. The change in state of charge is obtained by bilinear interpolation calculation based on the lookup data of voltage and temperature of the individual battery cell at the charging start rest point and the charging end rest point. The analysis module is used to identify outlier individual cells in the individual cells based on statistical characteristic values ​​associated with the change in the state of charge of the individual cells.

10. A vehicle, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 1 to 8.

Images