Method and device for detecting abnormal single cells in battery pack

Abstract

The invention provides a method and a device for detecting abnormal single cells in a battery pack. The method comprises the following steps: a data acquisition step: acquiring state related data of the battery pack; a feature extraction step: extracting features from the battery pack state related data, the features including kurtosis and skewness of voltage distribution of all single cells in the battery pack; and a feature grouping step: grouping the features according to the combination of one or more of the influence factors of the voltage distribution.

Description

Technical Field

[0001] The present invention generally relates to detecting abnormal single cells, and more particularly, to a method and apparatus for detecting abnormal single cells in a battery pack. Background Art

[0002] Herein, battery refers to all elements and devices that have the function of storing electricity. For example, batteries include lithium-ion batteries, lithium-ion capacitors, nickel-hydrogen batteries, graphene batteries, all-solid-state batteries, double-layer capacitors, air batteries, etc.

[0003] In recent years, various batteries have been actively developed. In particular, with the development of products such as mobile phones, smartphones, tablet computers, notebook personal computers and other portable information terminals, game devices, portable music players, digital cameras, medical equipment, hybrid electric vehicles (HEV), electric vehicles (EV) or plug-in hybrid electric vehicles (PHEV) and electric motorcycles, the demand for high-quality and high-reliability batteries has increased dramatically.

[0004] Many electronic devices (such as electric vehicles, hybrid vehicles, plug-in hybrid electric vehicles (PHEVs), or electric motorcycles, etc.) generally use a battery pack (usually simply referred to as a battery) composed of hundreds of single cells connected in parallel or in series.

[0005] The voltages of multiple cells in a battery pack may be inconsistent. For example, the voltages of three cells are 3.223V, 3.229V, and 3.226V, but the differences are not large and are considered normal inconsistencies. Normal inconsistencies are usually not caused by cell failures.

[0006] However, due to various reasons (for example, impurities are added to the single cells during the production process, the single cells leak, or differences in usage conditions such as charging times, discharge depth, charging current, charging environment (for example, temperature changes), etc.), compared with other normal single cells in the same battery pack, abnormal single cells may degrade or fail earlier, showing a voltage that is significantly different from other normal single cells (for example, 3.123 volts), that is, abnormal inconsistency.

[0007] It is desirable to be able to distinguish between normal and abnormal inconsistencies in the voltage of a single cell. However, due to the unevenness in production, the voltages of multiple single cells in a common battery pack are inherently uneven. Therefore, unless the abnormality of a single cell is very obvious, it will not be possible to identify the abnormal single cell in time by simply looking for the abnormal single cell voltage value. Therefore, in order to reduce safety risks, it is desirable to detect abnormal single cells early and issue timely warnings.

[0008] Traditional abnormal single-cell detection techniques involve detecting the battery behavior of the battery pack and then comparing the measured battery behavior with a predefined battery trend or threshold. However, optimistically setting the threshold may lead to delayed detection, while conservative thresholds may lead to frequent false alarms. In addition, setting a predefined battery trend or threshold usually requires costly testing of the battery pack. Summary of the invention

[0009] The following summary is provided to introduce selected concepts in a simplified form that are further described in the detailed description that follows. This summary is not intended to highlight key features or essential features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter.

[0010] An embodiment of the present invention provides a method for detecting abnormal single cells in a battery pack, comprising: a data acquisition step of acquiring data related to the battery pack status; a feature extraction step of extracting features from the data related to the battery pack status, wherein the features include the kurtosis and skewness of the voltage distribution of all single cells in the battery pack; a feature grouping step of grouping the features according to a combination of one or more of the influencing factors of the voltage distribution.

[0011] An embodiment of the present invention provides a device for detecting abnormal single cells in a battery pack, comprising: a data acquisition module, used to obtain battery pack status-related data; a feature extraction module, used to extract features from the battery pack status-related data, wherein the features include the kurtosis and skewness of the voltage distribution of all single cells in the battery pack; and a feature grouping module, used to group the features according to a combination of one or more of the influencing factors of the voltage distribution.

[0012] An embodiment of the present invention provides a non-transitory machine-readable storage medium storing machine-executable instructions, which, when executed by one or more processors, enable the one or more processors to perform the aforementioned method.

[0013] An embodiment of the present invention provides a computer program product, which includes processor-executable instructions. When the processor-executable instructions are executed by one or more processors, the one or more processors are caused to perform the aforementioned method.

[0014] Compared with the traditional abnormal single cell detection technology, the method for detecting abnormal single cells in a battery pack proposed in the present invention includes: first extracting and grouping the features of the battery pack status related data, and then inputting them into the machine learning model for analysis, so as to more effectively train and apply the machine learning model, thereby detecting abnormal single cells more quickly and efficiently, greatly improving the battery fault detection rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] A further understanding of the nature and advantages of the present invention may be realized by referring to the following drawings.In the accompanying drawings, similar components or features may have the same reference numerals.

[0016] Figure 1 A block diagram of an apparatus for detecting abnormal single cells in a battery pack using a machine learning model according to an embodiment of the present invention is shown.

[0017] Figure 2 A flow chart of a method for using a machine learning model to detect abnormal single cells in a battery pack according to an embodiment of the present invention is shown.

[0018] Figure 3 A schematic diagram of extracting the kurtosis of voltage distribution of all single cells in a battery pack as a feature according to an embodiment of the present invention is shown.

[0019] Figure 4 A schematic diagram of extracting the skewness of the voltage distribution of all single cells in a battery pack as a feature according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0020] The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that the discussion of these embodiments is only to enable those skilled in the art to better understand and implement the subject matter described herein, and is not a limitation of the scope of protection, applicability or examples set forth in the claims. The function and arrangement of the elements discussed can be changed without departing from the scope of protection of the present disclosure. Various examples can omit, replace or add various processes or components as needed. For example, the described method can be performed in an order different from the described order, and various steps can be added, omitted or combined. In addition, the features described relative to some examples can also be combined in other examples.

[0021] Figure 1 A block diagram of an apparatus 100 for detecting abnormal single cells in a battery pack using a machine learning model according to an embodiment of the present invention is shown.

[0022] Figure 1 The device 100 for detecting abnormal single cells in a battery pack includes a feature engineering part and a machine learning part. The feature engineering part includes: a data acquisition module 102, a feature extraction module 104, and a feature grouping module 106. The machine learning part includes a model training module 108, a model application module 109, and an output module 110.

[0023] The data acquisition module 102 is used to obtain data related to the battery pack status.

[0024] The feature extraction module 104 is used to extract features from the battery pack state related data. For example, the features may include the kurtosis and skewness of the voltage distribution of all the single cells in the battery pack.

[0025] The feature grouping module 106 is used to group the features according to a combination of one or more of the influencing factors of the voltage distribution. In the feature grouping module 106, the setting of the group range can be uniform or non-uniform, which will be further described later. Preferably, the feature grouping module 106 can also include a step for calculating the mean and / or maximum value of the features in each group.

[0026] The model training module 108 is used to train a machine learning model by utilizing features or the mean and / or maximum value of features.

[0027] After training the machine learning model, the model application module 109 applies the machine learning model to new battery pack status related data to detect the differences between each single cell in each group, thereby identifying abnormal single cells.

[0028] The output module 110 is used to output abnormal single cell detection results through the machine learning model.

[0029] Figure 2 FIG. 2 is a flow chart showing a method 200 for detecting abnormal single cells in a battery pack according to an embodiment of the present invention. Figure 1 The device 100 shown is implemented.

[0030] For ease of description, the following is an exemplary description of abnormal single cell detection in a battery pack of a new energy vehicle (e.g., an electric vehicle or a plug-in hybrid vehicle, etc.). However, this is only an example and not a limitation, and it should be understood that the present invention can also be applied to abnormal single cell detection in a battery pack of other electronic devices.

[0031] In the data collection step, the battery pack status related data (also called raw data) 202 is recorded at a certain frequency (for example, 1Hz to 10Hz, etc.) during the operation of the vehicle, and then the battery pack status related data is sent to a remote server (for example, a cloud server) via the vehicle's telecommunications unit and stored. The battery pack status related data includes: timestamp, battery pack current, single cell voltage, battery pack power, battery module temperature, etc. In addition, in the data collection step, the vehicle status related data is also recorded at a certain frequency (for example, 1Hz to 10Hz, etc.) during the operation of the vehicle, wherein the vehicle status related data includes vehicle status, mileage, etc.

[0032] In the feature extraction step, features are extracted from the battery pack state related data 202 (e.g., at each time stamp). Features may include, for example, the kurtosis 204-1 and skewness 204-2 of the voltage distribution of all single cells in the battery pack, etc. Figure 3 and Figure 4 Detailed description.

[0033] The voltage distribution of all cells in the battery pack is affected by some factors, which may include, for example, the size and direction of the battery pack current, the battery pack charge, the battery pack health status, the battery module temperature, the vehicle status, the mileage, etc. Therefore, in the feature grouping step, the features are divided into multiple groups 206-1 to 206-n according to a combination of one or more of the factors affecting the voltage distribution of all cells in the battery pack. When a factor affecting the voltage distribution is not available, other available equivalent factors can be used instead. For example, the battery pack health status can be replaced by the mileage.

[0034] An example of feature grouping is given below:

[0035] Table 1 <Battery Pack Capacity Group 206-1>

[0036] Group Number scope 0 [0,25％) 1 [25％,50％) 2 [50％,75％) 3 [75％,100％) 4 100％

[0037] Table 2 <Vehicle Status Group 206-2>

[0038] Group Number state 0 rest 1 Charge 2 drive

[0039] Table 3 <Driving Mileage Group 206-3>

[0040] Group Number Range (km) 0 [0,10k) 1 [10k,20k) 2 [20k,30k) … … n [n*10k,(n+1)*10k)

[0041] Table 4 <Battery module temperature group 206-3>

[0042]

[0043]

[0044] According to the example in the table above, the characteristics of a vehicle can be divided into 5 groups according to the battery pack power, 3 groups according to the vehicle status, n+1 groups according to the mileage, and 5 groups according to the battery module temperature. Therefore, the characteristics of a vehicle can be divided into 5*3*(n+1)*5 groups at most.

[0045] In addition, the range of each group can be uniform (as shown in Table 1 and Table 3). Optionally, the range of each group can also be non-uniform. When anomalies are more likely to occur within certain ranges, a finer range can be set, for example, a battery module low temperature range (as shown in Table 4), or a low battery pack power and a high battery pack power (as shown in Table 1').

[0046] Table 1'<Battery Pack Capacity Group 206-1'>

[0047] Group Number scope 0 [0,10％) 1 [10％,50％) 2 [50％,90％) 3 [90％,100％) 4 100％

[0048] After the feature grouping, in the model training step, the machine learning model is trained by using the historical data, including training or learning the data distribution and normal data characteristics 208. The historical data includes the mean and / or maximum value of the features or features in each feature grouping, for example. If the historical data has labels about abnormal single cells, the machine learning model can be a supervised model, such as Xgboost and neural network, etc. Otherwise, the machine learning model can be an unsupervised model, such as isolation forest, covariance estimation, etc.).

[0049] In the model application step, after the machine learning model is trained, the machine learning model can be applied to new battery pack status-related data to detect the differences between each single cell in each group, thereby identifying abnormal single cells.

[0050] In the output step, the machine learning model outputs abnormal single cells 210 in the analyzed battery pack.

[0051] Figure 3 A schematic diagram of extracting the kurtosis of voltage distribution of all single cells in a battery pack as a feature according to an embodiment of the present invention is shown.

[0052] like Figure 3 As shown, the voltage distribution of all the single cells in the battery pack can be divided into: peak distribution 302, normal peak distribution 304, and low peak distribution 306 according to the peak degree.

[0053] An example of a calculation formula for extracting the kurtosis of the voltage distribution of all cells in a battery pack is given below:

[0054]

[0055] in,

[0056] ω bolt Represents the kurtosis of the voltage distribution of all cells in the battery pack;

[0057] n represents the number of single cells in the battery pack;

[0058] v i Represents the voltage of the i-th single cell;

[0059] Indicates the average value of the single cell voltage;

[0060] v s Indicates the standard deviation of the single cell voltage.

[0061] Figure 4 A schematic diagram of extracting the skewness of the voltage distribution of all single cells in a battery pack as a feature according to an embodiment of the present invention is shown.

[0062] like Figure 4 As shown, the voltage distribution of all the single cells in the battery pack can be divided into: negative bias (also called left bias) distribution 402, symmetrical distribution 404, and positive bias (also called right bias) distribution 406 according to the skewness.

[0063] An example of a calculation formula for extracting the skewness of the voltage distribution of all cells in a battery pack is given below:

[0064]

[0065] in,

[0066] g volt Indicates the skewness of the voltage distribution of all cells in the battery pack;

[0067] n represents the number of single cells in the battery pack;

[0068] v i Represents the voltage of the i-th single cell;

[0069] Indicates the average value of the single cell voltage;

[0070] v s Indicates the standard deviation of the single cell voltage.

[0071] According to one embodiment, the present invention further provides a non-transitory machine-readable storage medium (not shown). The non-transitory machine-readable storage medium may have machine-executable instructions, and when the machine-executable instructions are executed by one or more processors, the one or more processors execute the above combined methods in various embodiments of the present application. Figure 1-Figure 4 Describes the various operations and functions.

[0072] According to one embodiment, the present invention provides a computer program product (not shown). The computer program product may include processor executable instructions, and when the processor executable instructions are executed by one or more processors, the one or more processors execute the above combined methods in various embodiments of the present application. Figure 1-4 Describes the various operations and functions.

[0073] It should be understood that all operations in the method described above are merely exemplary, and the present disclosure is not limited to any operation in the method or the order of these operations, but should cover all other equivalent transformations under the same or similar concept.

[0074] It should also be understood that all steps or modules in the above-described device can be implemented in various ways. These steps or modules can be implemented as hardware, software or a combination thereof. In addition, any step or module in these steps or modules can be further divided into sub-steps or modules or combined together functionally.

[0075] Processors have been described in conjunction with various devices and methods. These processors can be implemented using electronic hardware, computer software or any combination thereof. Whether these processors are implemented as hardware or software will depend on specific application and the overall design constraints imposed on the system. As an example, the processor provided in the present disclosure, any part of the processor or any combination of processors can be implemented as a microprocessor, a microcontroller, a digital signal processor (DSP), a field programmable gate array (FPGA), a programmable logic device (PLD), a state machine, a gate logic, a discrete hardware circuit, and other suitable processing components configured to perform the various functions described in the present disclosure. The function of the processor provided in the present disclosure, any part of the processor or any combination of processors can be implemented as software performed by a microprocessor, a microcontroller, a DSP or other suitable platforms.

[0076] The specific embodiments described above in conjunction with the accompanying drawings describe exemplary embodiments, but do not represent all embodiments that can be implemented or fall within the scope of protection of the claims. The term "example" used throughout this specification may mean "used as an example, instance or illustration" and does not mean "preferred" or "having advantages" over other embodiments. For the purpose of providing an understanding of the described technology, the specific embodiments include specific details. However, these technologies can be implemented without these specific details. In some instances, in order to avoid making the concepts of the described embodiments difficult to understand, well-known structures and devices are shown in block diagram form.

[0077] The above description is provided to enable any person skilled in the art to implement the various aspects described herein. Various modifications and variations of these aspects are apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects. Therefore, the claims are not intended to be limited to the aspects shown herein. All structural and functional equivalents of the elements of the various aspects described in this disclosure that are known or to be known to those skilled in the art will be expressly included herein by reference and are intended to be covered by the claims. The present disclosure is not limited to the examples and designs described herein, but is consistent with the widest range of principles and novel features disclosed herein.

Claims

1. A method for detecting abnormal single cells in a battery pack, comprising: Data collection steps: obtain data related to battery pack status; Feature extraction step: extracting features from the battery pack state related data, wherein the features include the kurtosis and skewness of the voltage distribution of all single cells in the battery pack; Feature grouping step: grouping the features according to a combination of one or more of the factors affecting the voltage distribution.

2. The method according to claim 1, wherein: The feature grouping step further comprises: a step for calculating the mean and / or maximum value of the feature in each group. 3 . The method according to claim 1 , wherein in the feature grouping step, the group range is set uniformly or non-uniformly.

4. The method according to claim 2 further includes a model training step, wherein the machine learning model is trained by utilizing the feature or the mean and / or maximum value of the feature.

5. The method according to claim 4 further comprises a model application step, wherein: After training the machine learning model, the machine learning model is applied to new battery pack status related data to detect the differences of each single cell in each group, thereby identifying abnormal single cells.

6. The method according to claim 5, further comprising an output step, wherein: The machine learning model outputs abnormal single cell detection results.

7. The method according to claim 1, wherein: The battery pack status related data includes: timestamp, battery pack current, single cell voltage, battery pack power, and battery module temperature.

8. The method according to claim 1, wherein: The data collection step also includes: acquiring vehicle status related data, wherein the vehicle status related data includes vehicle status and mileage.

9. The method according to claim 1, wherein: The feature extraction step includes determining the feature at each timestamp based on each timestamp, the difference between each single cell voltage and the average single cell voltage, the standard deviation of the single cell voltage, and the number of single cells in the battery pack.

10. The method according to claim 1, wherein: The influencing factors of the voltage distribution include the battery pack current size and direction, battery pack power, battery pack health status, battery module temperature, vehicle status, and mileage.

11. A device for detecting abnormal single cells in a battery pack, comprising: Data acquisition module, used to obtain data related to battery pack status; A feature extraction module, used to extract features from the battery pack state related data, wherein the features include the kurtosis and skewness of the voltage distribution of all single cells in the battery pack; The feature grouping module is used to group the features according to a combination of one or more of the factors affecting the voltage distribution.

12. The device according to claim 11, wherein The feature grouping module further comprises: a module for calculating the mean value and / or maximum value of the feature in each group. 13 . The apparatus according to claim 11 , wherein in the feature grouping module, the group range is set uniformly or non-uniformly.

14. The apparatus according to claim 12 further comprises a model training module for training a machine learning model by utilizing the feature or the mean and / or maximum value of the feature.

15. The apparatus according to claim 14, further comprising a model application module, wherein: After training the machine learning model, the machine learning model is applied to new battery pack status related data to detect the differences of each single cell in each group, thereby identifying abnormal single cells.

16. The device according to claim 15 further comprises an output module, configured to output abnormal single cell detection results by the machine learning model.

17. The device according to claim 11, wherein: The battery pack status related data includes: timestamp, battery pack current, single cell voltage, battery pack power, and battery module temperature.

18. The device according to claim 11, wherein The data acquisition module is also used to obtain vehicle status related data, and the vehicle status related data includes vehicle status and mileage.

19. The device according to claim 11, wherein: The feature extraction module includes: determining the feature at each timestamp based on each timestamp, the difference between each single cell voltage and the average single cell voltage, the standard deviation of the single cell voltage, and the number of single cells in the battery pack.

20. The device according to claim 11, wherein The influencing factors of the voltage distribution include the battery pack current size and direction, battery pack power, battery pack health status, battery module temperature, vehicle status, and mileage.

21. A non-transitory machine-readable storage medium storing machine-executable instructions which, when executed by one or more processors, cause the one or more processors to perform the method according to any one of claims 1 to 10.

22. A computer program product comprising processor-executable instructions which, when executed by one or more processors, cause the one or more processors to perform the method according to any one of claims 1 to 10.

Images