Battery pack abnormality early-warning method and battery pack abnormality early-warning apparatus

Abstract

The present application relates to a battery pack abnormality early-warning method and a battery pack abnormality early-warning apparatus. The method comprises: acquiring a plurality of target state-of-charge differences of all battery cells of a plurality of vehicles at different preset times; on the basis of the distribution of the plurality of target state-of-charge differences over time, determining a target state-of-charge difference range for each battery cell; and if the current target state-of-charge difference of the current vehicle is outside the target state-of-charge difference range, issuing an early warning regarding the abnormality of a battery pack of the current vehicle.

Description

Battery Pack Anomaly Early Warning Method and Device

[0001] This application claims priority to Chinese Patent Application No. 202410954624.0, filed on July 16, 2024, the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of battery anomaly warning technology, specifically to a battery pack anomaly warning method and a battery pack anomaly warning device. Background Technology

[0003] With the increasing application of new energy vehicles in the market, big data technology has shown promising prospects in this field. Among these applications, establishing vehicle-level fault risk prediction and early warning systems based on electric vehicle operating data from a big data perspective has become a current research hotspot. Big data platforms can utilize multi-dimensional tags for data collection, and through data filtering, integration, and cleaning, intelligent data processing and analysis can be achieved, thereby enabling the diagnosis and early warning of battery system faults.

[0004] In related technologies, the battery pack risk prediction and fault identification models established by different automakers will vary due to differences in material systems and the performance of the batteries themselves. Invention Overview

[0005] However, the vehicle displays the state of charge of the battery pack, but cannot estimate or identify the state of charge of any individual cell. Therefore, it cannot more effectively identify abnormal risks in the battery pack, resulting in frequent fire accidents and low safety.

[0006] Firstly, this application provides a battery pack anomaly early warning method, the method comprising: acquiring multiple open-circuit voltage data of battery packs of multiple vehicles, the open-circuit voltage data including the open-circuit voltage of each cell in the battery pack; obtaining multiple target state of charge differences of all cells of multiple vehicles at different preset times based on the multiple open-circuit voltage data; determining the target state of charge difference range of each cell according to the distribution of the multiple target state of charge differences over time; determining whether the current target state of charge difference of the current vehicle is within the target state of charge difference range; if the current target state of charge difference of the current vehicle is within the target state of charge difference range, then marking the battery pack of the current vehicle as normal; if the current target state of charge difference of the current vehicle is outside the target state of charge difference range, then issuing an early warning for the battery pack anomaly of the current vehicle.

[0007] Secondly, this application also provides a battery pack abnormality warning device, which includes a processing unit connected to multiple vehicles and is used to implement a battery pack abnormality warning method. Beneficial effects

[0008] This application acquires multiple open-circuit voltage data sets of battery packs from multiple vehicles, and based on these data sets, obtains multiple target state of charge (SOC) differences for all cells in the vehicles at different preset times. Then, it determines the SOC difference range for each cell based on the distribution of these SOC differences over time. Finally, it judges whether the current SOC difference of the current vehicle falls within the target SOC difference range. Based on these aspects, this application enables the determination of whether the SOC is abnormal at the individual cell level, more effectively identifying abnormal risks in the battery pack and improving the overall vehicle safety. Attached Figure Description

[0009] Figure 1 shows a flowchart of the battery pack anomaly warning method in a possible implementation of this application.

[0010] Figure 2 shows a schematic diagram of the distribution of the target state of charge difference over time in this application.

[0011] Figure 3 shows a schematic diagram of the first expression of this application.

[0012] Figure 4 shows a block diagram of the processing unit in a possible implementation of this application. Embodiments of the present invention

[0013] Figure 1 shows a flowchart of a battery pack anomaly warning method in a possible implementation of this application. As shown in Figure 1, the battery pack anomaly warning method includes:

[0014] Step S1: Obtain multiple open-circuit voltage data of battery packs from multiple vehicles. The open-circuit voltage data includes the open-circuit voltage of each cell in the battery pack.

[0015] In one possible implementation, the vehicle is a complete vehicle, and the vehicle's battery pack may include multiple individual battery cells. A Battery Management System (BMS) is electrically connected to each cell in the vehicle's battery pack. Multiple vehicles can connect to a battery pack anomaly warning device via cloud networking. The battery pack anomaly warning device includes a processing unit that implements a battery pack anomaly warning method. Optionally, the battery pack anomaly warning device is a cloud-based control console, which may be a cloud-based big data platform.

[0016] The open-circuit voltage data for each cell includes multiple open-circuit voltages (OCVs) at different times. The open-circuit voltage is the voltage between the positive and negative terminals of the cell when no load is connected. The control console can be electrically connected to the battery management system of each vehicle to receive multiple open-circuit voltage data sets collected by the battery management system.

[0017] Step S2: Based on multiple open-circuit voltage data, obtain the differences in multiple target states of charge of all cells of multiple vehicles at different preset times;

[0018] In possible implementations, the state of charge (SOC) of a battery cell represents the ratio of the cell's usable capacity to its total capacity, and can be expressed as a percentage. At any given moment, there is a one-to-one correspondence between the cell's SOC and its open-circuit voltage. The preset time can be in days.

[0019] In possible implementations, the differences in target state of charge (SOC) of all cells in multiple vehicles at different preset times are obtained based on multiple open-circuit voltage data, including:

[0020] Step S21: Based on multiple open-circuit voltage data, obtain multiple target states of charge for all cells of multiple vehicles;

[0021] The target state of charge (SBC) of the battery cell can characterize the actual SBC of the battery cell. In this application, to make the target SBC more closely resemble reality, the SBC is utilized not only when the vehicle is off, but also when the vehicle is on.

[0022] In possible implementations, multiple target states of charge (SOCs) for all cells in multiple vehicles are obtained based on multiple open-circuit voltage data sets, including:

[0023] Step S211: Obtain the SOC-OCV curves of each cell in multiple vehicles;

[0024] In this context, the SOC-OCV curves of each cell in multiple vehicles can be obtained in advance through the battery management system.

[0025] Step S212: Based on multiple open-circuit voltage data and the SOC-OCV curve of each cell, obtain multiple first states of charge of each cell when the vehicle is turned off;

[0026] With the vehicle powered off and no load connected to the battery pack cells, the control console can determine the first state of charge (SOC) corresponding to different open-circuit voltages for each cell based on the static SOC-OCV curves of each cell in the battery management system. The static SOC-OCV curves can be obtained using the open-circuit voltage method.

[0027] Step S213: Use ampere-hour integration to correct multiple first states of charge to obtain multiple second states of charge for each cell;

[0028] Since the first state of charge obtained using the static SOC-OCV curve does not match the actual situation, the ampere-hour integration method can be used to correct the first state of charge in order to make the estimation of the state of charge more accurate.

[0029] When the vehicle is started, the battery pack cells are connected to a load. At this time, the ampere-hour integral method is used to correct the first state of charge. The formula for the ampere-hour integral method is as follows:

[0030] SOC t =SOC t-1 -(I×Δt / total cell capacity)×100%

[0031] Where I represents the charging and discharging current of the battery cell; the direction of current I is positive when the cell is discharging and negative when it is charging. Δt represents the change in time, I×Δt represents the change in cell capacity over time Δt, and SOC represents the state of charge (SOC). t-1 The second state of charge at time t-1, SOC t This represents the second state of charge at time t. In practical applications, the current I can be detected by the current sensor of the battery management system, and the detected current I can be sent to the control console for further processing.

[0032] Step S214: Determine multiple target states of charge for all cells of multiple vehicles based on multiple first states of charge and multiple second states of charge.

[0033] In a possible implementation, the target state of charge (SOC) can be determined by a first SOC and a second SOC. In a possible implementation, for a fixed cell, a weighting coefficient k1 can be assigned to the first SOC, and a weighting coefficient k2 can be assigned to the second SOC. The target SOC can then be expressed as k1 × first SOC + k2 × second SOC, where both k1 and k2 are greater than 0 and less than 1, with k1 being greater than k2. For example, k1 is 0.8 and k2 is 0.2.

[0034] Step S22: Iterate through multiple open-circuit voltages corresponding to multiple target states of charge to obtain multiple first open-circuit voltage extremes and multiple second open-circuit voltage extremes at different preset times;

[0035] In possible implementations, the first open-circuit voltage extreme value can be the maximum value of all open-circuit voltages of all cells in multiple vehicles at a preset time, and the second open-circuit voltage extreme value can be the minimum value of all open-circuit voltages of all cells in multiple vehicles at a preset time. Both the first and second open-circuit voltage extreme values ​​are determined based on the corresponding preset time.

[0036] Step S23: Obtain multiple first reference states of charge corresponding to multiple first open-circuit voltage extreme values ​​and multiple second reference states of charge corresponding to multiple second open-circuit voltage extreme values ​​by reverse lookup;

[0037] In one possible implementation, the open-circuit voltage and target states of charge at different preset times can be stored in a table in the battery pack anomaly warning device. During a reverse lookup, the table containing the open-circuit voltage and target states of charge at different preset times can be queried.

[0038] For example, the preset time is in days, and multiple open-circuit voltages and target states of charge are obtained for each cell each day. On day 1, the highest open-circuit voltage of each cell in multiple vehicles is OCV. max1 The corresponding target state of charge is SOC. max1 The minimum open-circuit voltage is OCV. min1 The corresponding target state of charge is SOC. min1 On day 1, the highest open-circuit voltage of each cell in multiple vehicles was OCV. max2 The corresponding target state of charge is SOC. max2 The minimum open-circuit voltage is OCV. min2 The corresponding target state of charge is SOC. min2 At this time, SOC max1 and SOC max2 All are in the first reference state of charge (SOC). min1 and SOC min2 All are in the second reference state of charge.

[0039] Step S24: Determine the differences in multiple target states of charge for all cells of multiple vehicles at different preset times based on multiple first reference states of charge and multiple second states of charge.

[0040] In possible implementations, the differences in multiple target states of charge (SOCs) of all cells in multiple vehicles at different preset times are determined based on multiple first reference SOCs and multiple second SOCs, including:

[0041] Step S241: Sum the multiple first reference states of charge of all cells at the same preset time and then average them to obtain the average value of the first reference states of charge;

[0042] Each battery cell corresponds to a first reference state of charge at a preset time. By summing the multiple first reference states of charge of all battery cells at the same preset time and then dividing by the total number of battery cells in multiple vehicles, the average value of the first reference states of charge at the same preset time can be obtained.

[0043] Step S242: Sum the multiple second reference states of charge of all cells at the same preset time and then average them to obtain the average value of the second reference states of charge;

[0044] Each battery cell corresponds to a second reference state of charge at a preset time. By summing the multiple second reference states of charge of all battery cells at the same preset time and then dividing by the total number of battery cells in multiple vehicles, the average value of the second reference states of charge at the same preset time can be obtained.

[0045] Step S243: Obtain the target state of charge difference at a preset time based on the average value of the first reference state of charge and the average value of the second reference state of charge.

[0046] In one possible implementation, the average value of the first reference state of charge and the average value of the second reference state of charge can be subtracted, and the absolute value can be taken to obtain the target state of charge difference over a preset time. Since there are multiple preset times, and each preset time can correspond to one target state of charge difference, there can also be multiple target state of charge differences.

[0047] Step S3: Determine the range of target state of charge differences for each cell based on the distribution of differences between multiple target states of charge over time;

[0048] Figure 2 shows a schematic diagram of the distribution of the target state of charge difference over time in this application. As shown in Figure 2, the horizontal axis represents the reference time difference between the preset time and the market launch time, and the vertical axis represents the target state of charge difference. It can be seen that the density of the target state of charge difference varies at different preset times.

[0049] In possible implementations, the range of target state of charge (SPC) differences for each cell is determined based on the distribution of multiple target SPC differences over time, including:

[0050] Step S31: Obtain the launch time of multiple vehicles and calculate multiple reference time differences between the launch time and multiple preset times;

[0051] The vehicle's launch date can be pre-stored in the battery management system. The launch date can be accurate to the day, so when calculating the reference time difference, since the preset time is in days, the reference time difference will also be in days. The reference time difference can be a positive integer.

[0052] Step S32: Fit multiple target state of charge differences and multiple reference time differences to obtain the first expression for the target state of charge differences and reference time differences;

[0053] In one possible implementation, the distribution of multiple target state-of-charge differences relative to multiple reference time differences can be fitted with the reference time difference as the horizontal axis and the target state-of-charge difference as the vertical axis to obtain a first expression for the target state-of-charge difference and the reference time difference. This first expression can be f1(Δt) d f1 represents the functional relationship between the target state of charge difference and the reference time difference, where Δt is the reference time difference.d The reference time difference is used as a reference. Optionally, the target state of charge difference and the reference time difference have a linear relationship, and the functional relationship between the target state of charge difference and the reference time difference is a linear function.

[0054] Figure 3 shows a schematic diagram of the first expression of this application. As shown in Figure 3, the horizontal axis represents the reference time difference between the preset time and the listing time, and the vertical axis represents the target state of charge difference. Since the density of the target state of charge difference varies at different preset times, a piecewise function with different slopes can be used to fit the relationship between the target state of charge difference and the reference time difference.

[0055] Step S33: Determine the first difference boundary and the second difference boundary based on the first expression to obtain the target state of charge difference range for each cell.

[0056] Wherein, the first difference boundary and the second difference boundary are the end values ​​of the target state of charge difference range, and the first difference boundary is smaller than the second difference boundary.

[0057] In one possible implementation, the first and second difference boundaries can be calculated using a first expression. For example, initial first and second difference boundaries can be set according to requirements. Then, the target state of charge (SBC) differences of multiple unknown cells at a preset time can be calculated using the first expression. It can then be checked whether the SBC differences of the multiple unknown cells fall within the initial first and second difference boundaries. If they do, the initial first and second difference boundaries remain unchanged. If they do not, the initial first and second difference boundaries can be changed with a set step size until the target SBC differences of the multiple unknown cells with the target probability fall within the first and second difference boundaries. At this point, the range of the target SBC differences for each cell is obtained.

[0058] Step S4: Determine whether the current target state of charge difference of the current vehicle is within the target state of charge difference range. If the current target state of charge difference of the current vehicle is within the target state of charge difference range, mark the current vehicle's battery pack as normal; if the current target state of charge difference of the current vehicle is outside the target state of charge difference range, issue a warning for the current vehicle's battery pack as abnormal.

[0059] In possible implementations, determining whether the current target state of charge difference of the current vehicle is within the target state of charge difference range includes:

[0060] Step S41: Obtain the differences in the initial state of charge of multiple vehicles at the time of market launch;

[0061] Step S42: Calculate the rate of change of multiple target state of charge differences based on the initial state of charge difference and the multiple target state of charge differences at different preset times;

[0062] In possible implementations, the rate of change of the target state of charge difference can be calculated using the following formula:

[0063] ΔSOC' = (SOC0 - SOC) td ) / SOC0

[0064] Where SOC0 is the initial state of charge difference at the time of market launch, SOC td The target state of charge difference is defined as the difference in the target state of charge over a preset time period, and ΔSOC' is the rate of change of the target state of charge difference.

[0065] Step S43: Determine the range of the target state of charge difference change rate for each cell based on the distribution of the change rate of multiple target state of charge difference over time.

[0066] In possible implementations, the range of the target state-of-charge difference rate of change for each cell is determined based on the distribution of multiple target state-of-charge difference rates over time, including:

[0067] Step S431: Fit the rate of change of multiple target states of charge and multiple reference time differences to obtain a second expression for the rate of change of multiple target states of charge and multiple reference time differences;

[0068] In one possible implementation, the distribution of multiple target state-of-charge differences relative to multiple reference time differences can be fitted with the reference time difference as the horizontal axis and the rate of change of target state-of-charge differences as the vertical axis, thus obtaining a second expression for the target state-of-charge difference rate of change and the reference time difference. This second expression can be f2(Δt) d f2 represents the functional relationship between the target state of charge difference rate of change and the reference time difference, Δt. d The reference time difference is used as a reference. Optionally, the rate of change of the target state of charge difference and the reference time difference are linearly related, and the functional relationship between the rate of change of the target state of charge difference and the reference time difference is a linear function.

[0069] Step S432: Determine the first rate of change boundary and the second rate of change boundary based on the second expression to obtain the target state of charge difference rate of change range for each cell.

[0070] Wherein, the first rate of change boundary and the second rate of change boundary are the end values ​​of the range of target state of charge difference rate of change, and the first rate of change boundary is smaller than the second rate of change boundary.

[0071] In one possible implementation, the first and second rate of change boundaries can be calculated using a second expression, similar to the method used for the first expression. For example, initial first and second rate of change boundaries can be set according to requirements. Then, the rate of change of the target state of charge difference of multiple unknown cells over a preset time can be calculated using the first expression. It can then be checked whether the rate of change of the target state of charge difference of the multiple unknown cells falls within the initial first and second rate of change boundaries. If it has, the initial first and second rate of change boundaries remain unchanged. If it has not, the initial first and second rate of change boundaries can be changed by a set step size until the target state of charge difference of the multiple unknown cells with the target probability falls within the first and second rate of change boundaries. At this point, the range of the target state of charge difference rate of change for each cell is obtained.

[0072] In possible implementations, before determining whether the current target state of charge difference of the current vehicle is within the target state of charge difference range and whether the current target state of charge difference rate of change is within the target state of charge difference rate of change range, the battery pack anomaly warning method includes:

[0073] Step S441: Calculate the differences in multiple target states of charge and the standard deviation of the rate of change of the differences in multiple target states of charge;

[0074] Step S442: Determine the target state of charge difference range and the target state of charge difference change rate range for each cell based on the standard deviation and the preset confidence interval.

[0075] In one possible implementation, the vehicle's market launch time can be used as a starting point to obtain the target state of charge (SBC) difference and the rate of change of the target SBC difference for all vehicles launched at the same time. Their mean and standard deviation can be calculated. Then, based on a specific confidence interval (e.g., 95%), the upper and lower limits of the target SBC difference and the target SBC difference rate of change can be determined, thereby determining the range of the target SBC difference and the range of the target SBC difference rate of change for each cell. It is worth noting that the range of the target SBC difference rate of change obtained in step S432 and the range of the target SBC difference rate of change obtained in step S442 can be substituted for each other.

[0076] Step S44: Determine whether the current target state of charge difference of the current vehicle is within the target state of charge difference range, and whether the current target state of charge difference change rate is within the target state of charge difference change rate range.

[0077] In possible implementations, determining whether the current target state of charge difference of the current vehicle is within the target state of charge difference range, and whether the current target state of charge difference change rate is within the target state of charge difference change rate range, includes:

[0078] Step S441: If the current target state of charge difference of the current vehicle is within the target state of charge difference range, and the current target state of charge difference change rate is within the target state of charge difference change rate range, then mark the current vehicle's battery pack as normal; if the current target state of charge difference of the current vehicle is outside the target state of charge difference range or the current target state of charge difference change rate is outside the target state of charge difference change rate range, then issue a warning for the current vehicle's battery pack as abnormal.

[0079] In one possible implementation, marking the current vehicle's battery pack as normal can be achieved by sending a normal command to the battery management system, at which point an icon indicating a normal battery pack will be displayed on the vehicle's central control screen. Conversely, if the current vehicle's battery pack is malfunctioning, an malfunction command can be sent to the battery management system, at which point an malfunction icon will be displayed on the vehicle's central control screen, and a voice prompt will alert the driver that the battery pack is in an abnormal state.

[0080] In addition, this application also provides a battery pack abnormality warning device, which includes a processing unit connected to multiple vehicles, and the processing unit is used to implement a battery pack abnormality warning method.

[0081] Figure 4 shows a block diagram of the processing unit in a possible implementation of this application. As shown in Figure 4, the processing unit includes:

[0082] The data acquisition module 41 is used to acquire multiple open-circuit voltage data of battery packs of multiple vehicles. The open-circuit voltage data includes the open-circuit voltage of each cell in the battery pack.

[0083] The difference determination module 42 is connected to the data acquisition module 41 and is used to obtain the differences in multiple target states of charge of all cells of multiple vehicles at different preset times based on multiple open circuit voltage data.

[0084] The range determination module 43 is connected to the difference determination module 42 and is used to determine the range of target state of charge difference for each cell based on the distribution of multiple target state of charge differences over time.

[0085] The judgment module 44, connected to the range determination module 43, is used to determine whether the current target state of charge difference of the current vehicle is within the target state of charge difference range. If the current target state of charge difference of the current vehicle is within the target state of charge difference range, the battery pack of the current vehicle is marked as normal; if the current target state of charge difference of the current vehicle is outside the target state of charge difference range, an early warning is issued for the battery pack of the current vehicle to be abnormal.

[0086] For specific details regarding the processing unit, please refer to the relevant description of the battery pack anomaly warning method, which will not be repeated here.

[0087] In summary, by acquiring multiple open-circuit voltage data sets of battery packs from multiple vehicles, and based on these data sets, obtaining multiple target state of charge (SOC) differences for all cells in the vehicles at different preset times, and then determining the range of target SOC differences for each cell based on the distribution of these differences over time, this application can determine whether the current target SOC difference of the current vehicle falls within this range. This enables the assessment of whether the SOC is abnormal at the individual cell level, more effectively identifying abnormal risks in the battery pack and improving overall vehicle safety. Furthermore, by utilizing a big data platform to record, compare, analyze, and summarize changes in SOC during vehicle operation and static processes, it can identify abnormal self-discharge in the vehicle or certain batteries and vehicles in the battery pack that are experiencing excessively rapid degradation. This allows for the early detection of battery pack anomalies. The use of a big data platform ensures that the model's parameter boundaries are supported by a relatively sufficient amount of data, resulting in strong representativeness.

Claims

1. A battery pack abnormality early warning method, comprising: obtaining multiple open-circuit voltage data of battery packs of multiple vehicles, the open-circuit voltage data comprising open-circuit voltage of each cell in the battery pack; obtaining multiple target state of charge differences of all cells of the multiple vehicles at different preset times based on the multiple open-circuit voltage data; determining a target state of charge difference range of each cell according to distribution of the multiple target state of charge differences over time; judging whether a current target state of charge difference of a current vehicle is within the target state of charge difference range, and if the current target state of charge difference of the current vehicle is within the target state of charge difference range, marking the battery pack of the current vehicle as normal; if the current target state of charge difference of the current vehicle is outside the target state of charge difference range, early warning of an abnormality of the battery pack of the current vehicle.

2. The battery pack abnormality early warning method of claim 1, wherein, The obtaining of the multiple target state of charge differences of all cells of the multiple vehicles at different preset times based on the multiple open-circuit voltage data comprises: obtaining multiple target states of charge of all cells of the multiple vehicles based on the multiple open-circuit voltage data; traversing multiple open-circuit voltages corresponding to the multiple target states of charge to obtain multiple first open-circuit voltage extreme values and multiple second open-circuit voltage extreme values at different preset times; obtaining multiple first reference states of charge corresponding to the multiple first open-circuit voltage extreme values and multiple second reference states of charge corresponding to the multiple second open-circuit voltage extreme values through reverse query; determining the multiple target state of charge differences of all cells of the multiple vehicles at different preset times according to the multiple first reference states of charge and the multiple second states of charge.

3. The battery pack abnormality early warning method of claim 2, wherein, The obtaining of the multiple target states of charge of all cells of the multiple vehicles based on the multiple open-circuit voltage data comprises: obtaining SOC-OCV curves of each cell of the multiple vehicles; obtaining multiple first states of charge of each cell of the multiple vehicles when the vehicles are turned off according to the multiple open-circuit voltage data and the SOC-OCV curves of each cell; correcting the multiple first states of charge by ampere-hour integration to obtain multiple second states of charge of each cell; determining the multiple target states of charge of all cells of the multiple vehicles based on the multiple first states of charge and the multiple second states of charge.

4. The battery pack abnormality early warning method of claim 2, wherein The determining of the multiple target state of charge differences of all cells of the multiple vehicles at different preset times according to the multiple first reference states of charge and the multiple second states of charge comprises: summing and averaging the multiple first reference states of charge of all cells at the same preset time to obtain a first reference state of charge average value; summing and averaging the multiple second reference states of charge of all cells at the same preset time to obtain a second reference state of charge average value; obtaining a target state of charge difference at the preset time according to the first reference state of charge average value and the second reference state of charge average value.

5. The battery pack abnormality pre-warning method of claim 1, wherein, The determining of the target state of charge difference range of each cell according to the distribution of the multiple target state of charge differences over time comprises: Obtaining a time to market of a plurality of vehicles, and calculating a plurality of reference time differences between the time to market and a plurality of preset times; Fitting a plurality of the target state of charge differences and a plurality of reference time differences to obtain a first expression of the target state of charge differences and reference time differences; Determining a first difference boundary and a second difference boundary based on the first expression to obtain a target state of charge difference range of each of the battery cells, wherein the first difference boundary and the second difference boundary are end values of the target state of charge difference range, and the first difference boundary is smaller than the second difference boundary.

6. The battery pack abnormality pre-warning method of claim 5, wherein, The battery pack abnormality early warning method further comprises: Obtaining an initial state of charge difference of a plurality of vehicles at a time to market; Calculating a plurality of target state of charge difference change rates of a plurality of the reference time differences according to the initial state of charge difference and a plurality of target state of charge differences of different preset times; Determining a target state of charge difference change rate range of each of the battery cells according to a distribution of a plurality of the target state of charge difference change rates over time; Determining whether a current target state of charge difference of a current vehicle is located within the target state of charge difference range and whether a current target state of charge difference change rate is located within the target state of charge difference change rate range.

7. The battery pack abnormality pre-warning method of claim 6, wherein The determining a target state of charge difference change rate range of each of the battery cells according to a distribution of a plurality of the target state of charge difference change rates over time comprises: Fitting a plurality of the target state of charge difference change rates and a plurality of reference time differences to obtain a second expression of a plurality of the target state of charge difference change rates and a plurality of reference time differences; Determining a first change rate boundary and a second change rate boundary based on the second expression to obtain a target state of charge difference change rate range of each of the battery cells, wherein the first change rate boundary and the second change rate boundary are end values of the target state of charge difference change rate range, and the first change rate boundary is smaller than the second change rate boundary.

8. The battery pack abnormality early warning method of claim 6, wherein, The determining whether a current target state of charge difference of a current vehicle is located within the target state of charge difference range and whether a current target state of charge difference change rate is located within the target state of charge difference change rate range comprises: If the current target state of charge difference of the current vehicle is located within the target state of charge difference range and the current target state of charge difference change rate is located within the target state of charge difference change rate range, marking a battery pack of the current vehicle as normal; if the current target state of charge difference of the current vehicle is located outside the target state of charge difference range or the current target state of charge difference change rate is located outside the target state of charge difference change rate range, early warning an abnormality of the battery pack of the current vehicle.

9. The battery pack abnormality pre-warning method of claim 6, wherein, Before determining whether a current target state of charge difference of a current vehicle is located within the target state of charge difference range and whether a current target state of charge difference change rate is located within the target state of charge difference change rate range, the battery pack abnormality early warning method comprises: Calculating a standard deviation of a plurality of the target state of charge differences and a plurality of the target state of charge difference change rates; Determine a target state of charge difference range of each of the battery cells and a target state of charge difference change rate range of each of the battery cells based on the standard deviation and a preset confidence interval.

10. A battery pack abnormality early warning device, comprising a processing unit connected to a plurality of vehicles, wherein the processing unit is configured to implement the battery pack abnormality early warning method according to any one of claims 1 to 9.

Images