Method, apparatus, device, medium and product for detecting differences in production data of batteries

Abstract

The present application relates to a method, device, equipment, medium and product for detecting differences in production data of batteries. The method includes: obtaining probability distributions corresponding to two target production data sets in a production data set group of batteries, determining a target ratio according to the probability distributions corresponding to each target production data set in the production data set group, and determining the degree of difference between a first production data set and a second production data set according to the probability distributions corresponding to at least one target production data set and the target ratio. It can quantify the overlap degree and shape similarity of the interval ranges of production data, provide an in-depth data analysis perspective, and improve the accuracy of determining the degree of difference between two data sets.

Description

Technical Field

[0001] This application relates to the technical field of data detection, and particularly to a method, device, equipment, medium and product for detecting differences in production data of batteries. Background Art

[0002] Detecting differences in data in the production environment of batteries can effectively improve production efficiency and battery quality, and reduce uncertainties and risks in the production process.

[0003] Currently, when detecting differences between two data sets of batteries, the median consistency detection method is usually used to detect differences in data in the production environment to obtain the degree of difference between the two data sets, that is, to determine the median of each data set and judge whether the medians of the two data sets are consistent. If they are consistent, it is determined that the two data sets are basically the same. If there are differences in the medians of the two data sets, it is determined that the two data sets may be abnormal.

[0004] However, there is a problem that the accuracy of the determined degree of difference in current data difference detection is relatively low. Summary of the Invention

[0005] Based on this, it is necessary to provide a method, device, equipment, medium and product for detecting differences in production data of batteries that can improve the accuracy of the degree of difference in view of the above technical problems.

[0006] In a first aspect, this application provides a method for detecting differences in production data of batteries. The method includes:

[0007] Obtain the probability distributions corresponding to two target production data sets in a production data set group of batteries; the two target production data sets include a first production data set and a second production data set;

[0008] Determine a target ratio according to the probability distributions corresponding to the target production data sets in the production data set group;

[0009] Determine the degree of difference between the first production data set and the second production data set according to the probability distributions corresponding to at least one target production data set and the target ratio.

[0010] The method provided in this embodiment can quantify the overlap degree and shape similarity of the interval ranges of production data by obtaining the probability distributions corresponding to two target production data sets in the production data set group of batteries, determining the target ratio according to the probability distributions corresponding to the target production data sets in the production data set group, and determining the degree of difference between the first production data set and the second production data set according to the probability distributions corresponding to at least one target production data set and the target ratio. It provides an in-depth data analysis perspective and improves the accuracy of determining the degree of difference between the two data sets.

[0011] In one embodiment, obtaining the probability distributions corresponding to two target production data sets in the production data set group of the battery includes:

[0012] Obtaining the frequency sets corresponding to two target production data sets in the production data set group of the battery; the frequency sets are determined according to the frequencies of occurrence of the production data in the target production data sets;

[0013] Processing each frequency set to obtain the probability distribution corresponding to each target production data set.

[0014] The method provided in this embodiment balances the magnitudes of different features and reduces the improper influence of numerical magnitudes on the analysis results by obtaining the frequency sets corresponding to two target production data sets in the production data set group of the battery, processing each frequency set to obtain the probability distribution corresponding to each target production data set, thereby improving the accuracy of the obtained degree of difference.

[0015] In one embodiment, determining the target ratio according to the probability distributions corresponding to the target production data sets in the production data set group includes:

[0016] Determining a first ratio of the probability distribution of the first production data set to the probability distribution of the second production data set; the target ratio includes the first ratio.

[0017] The method provided in this embodiment lays a foundation for determining the degree of difference based on the target ratio and improves the accuracy of determining the degree of difference between two data sets by determining a first ratio of the probability distribution of the first production data set to the probability distribution of the second production data set and using the first ratio as the target ratio.

[0018] In one embodiment, determining the degree of difference according to the probability distributions corresponding to at least one target production data set and the target ratio includes:

[0019] Determining a first logarithm of the first ratio and determining a first product of the first logarithm and the probability distribution of the first production data set;

[0020] Performing integral processing on the first product within the production data interval corresponding to the target production data set to obtain a first integral result, and using the first integral result as the degree of difference.

[0021] The method provided in this embodiment improves the accuracy of determining the degree of difference between two data sets by determining a first logarithm of the first ratio, determining a first product of the first logarithm and the probability distribution of the first production data set, performing integral processing on the first product within the production data interval corresponding to the target production data set to obtain a first integral result, and using the first integral result as the degree of difference.

[0022] In one embodiment, determining a target ratio according to the probability distributions corresponding to the target production data sets in the production data set group includes:

[0023] Determining a target probability distribution according to the probability distributions corresponding to the target production data sets;

[0024] Determining a second ratio between the probability distribution of the first production data set and the target probability distribution, and a third ratio between the probability distribution of the second production data set and the target probability distribution; the target ratio includes the second ratio and the third ratio.

[0025] The method provided in this embodiment, by determining a second ratio between the probability distribution of the first production data set and the target probability distribution, and a third ratio between the probability distribution of the second production data set and the target probability distribution, and using the second ratio and the third ratio as the target ratio, thus lays a foundation for determining the degree of difference based on the target ratio and improves the accuracy of determining the degree of difference between the two data sets.

[0026] In one embodiment, determining the degree of difference according to the probability distributions corresponding to at least one production data set and the target ratio includes:

[0027] Determining a second logarithm of the second ratio, and determining a second product of the second logarithm and the probability distribution of the first production data set;

[0028] Determining a third logarithm of the third ratio, and determining a third product of the third logarithm and the probability distribution of the second production data set;

[0029] Performing integral processing on the second product within the production data interval range corresponding to the target production data set to obtain a second integral result, and performing integral processing on the third product within the production data interval range to obtain a third integral result;

[0030] Determining the degree of difference according to the second integral result and the third integral result.

[0031] The method provided in this embodiment, by determining the degree of difference according to the second integral result and the third integral result, this degree of difference can quantify the overlap degree and shape similarity of the production data interval range, thus improving the accuracy of determining the degree of difference between the two data sets.

[0032] In one embodiment, when the number of production data set groups is multiple, the method further includes:

[0033] Determining the anomaly detection result of each production data set group according to the degree of difference corresponding to each production data set group and the preset threshold corresponding to the production data set group.

[0034] The method provided in this embodiment determines the anomaly detection results of each production dataset group according to the difference degree corresponding to each production dataset group and the preset threshold corresponding to the production dataset group, so as to be able to analyze the change trend of the target production dataset in at least two production dataset groups in the time dimension and the consistency in the space dimension.

[0035] In one of the embodiments, the method further includes:

[0036] Obtain each initial production dataset of the battery;

[0037] If the initial production data in each initial production dataset is inconsistent, determine the union of the initial production data in each initial production dataset;

[0038] Update the to-be-updated initial production dataset according to the union to obtain the updated production dataset corresponding to the to-be-updated initial production dataset; the to-be-updated initial production dataset includes the initial production dataset whose initial production data is inconsistent with the union;

[0039] Determine each production dataset group according to each target production dataset; each target production dataset includes at least the updated production dataset.

[0040] The method provided in this embodiment, in the case where the initial production data in each initial production dataset is inconsistent, determines the union of the initial production data in each initial production dataset, and updates the to-be-updated initial production dataset according to the union, so as to make the data ranges of the production data in each target production dataset consistent, and further to determine the difference degree between two target production datasets in the production dataset group within the same data range.

[0041] In one of the embodiments, obtaining the frequency sets corresponding to two target production datasets in the production dataset group of the battery includes:

[0042] If the two target production datasets include the updated production dataset, fill the preset value into the target position in the initial frequency set corresponding to the updated production dataset to obtain the frequency set corresponding to the updated production dataset; the target position is the position in the initial frequency set corresponding to the newly added production data in the updated production dataset;

[0043] If the two target production datasets include the initial production dataset, use the initial frequency set corresponding to the initial production dataset as the frequency set corresponding to the initial production dataset.

[0044] For the method provided in this embodiment, if the updated production data set is included in the two target production data sets, the preset value is filled into the target position in the initial frequency set corresponding to the updated production data set to obtain the frequency set corresponding to the updated production data set, which can reduce the degree of analysis deviation caused by data missing, and further improve the accuracy of the obtained difference degree.

[0045] In one embodiment, each initial production data set includes initial production data sets corresponding to multiple different timestamps under the same production line;

[0046] Determining each production data set group according to each target production data set includes:

[0047] Taking the target production data set corresponding to the maximum timestamp as the first production data set in the production data set group, and taking the target production data set corresponding to a target timestamp as the second production data set in the production data set group, so as to obtain each production data set group;

[0048] Wherein, the target timestamp includes timestamps other than the maximum timestamp among all timestamps.

[0049] For the method provided in this embodiment, for the target production data set in the time dimension, by taking the target production data set corresponding to the maximum timestamp as the first production data set in the production data set group, and taking the target production data set corresponding to a target timestamp as the second production data set in the production data set group, so as to obtain each production data set group, it lays a foundation for further determining the change trend in the time dimension based on the difference degree of each production data set group.

[0050] In one embodiment, the preset threshold includes a first preset threshold. Determining the anomaly detection result of each production data set group according to the difference degree corresponding to each production data set group and the preset threshold corresponding to the production data set group includes:

[0051] Determining the median of each difference degree according to the difference degree corresponding to each production data set group;

[0052] If the median is greater than the first preset threshold, it is determined that the anomaly detection result is that the change trend of the production data of the production line over time is abnormal;

[0053] If the median is not greater than the first preset threshold, it is determined that the anomaly detection result is that the change trend of the production data of the production line over time is normal.

[0054] For the method provided in this embodiment, for the difference degree between the target production data sets in the time dimension, the change trend of the production data of the production line over time is judged through the median, so as to determine the abnormal changes or turning points in the production data, which is convenient for subsequent analysis of the reasons to improve production efficiency and battery quality.

[0055] In one embodiment, each initial production dataset includes the initial production datasets corresponding to the same battery on different production lines, or includes the initial production datasets corresponding to different machines of the same battery on the same production line;

[0056] Determining each production dataset group according to each target production dataset includes:

[0057] Taking any two target production datasets in each target production dataset as a production dataset group to obtain each production dataset group.

[0058] The method provided in this embodiment, for the target production datasets in the spatial dimension, takes any two target production datasets in each target production dataset as a production dataset group to obtain each production dataset group, laying a foundation for further determining the consistency of production data in the spatial dimension based on the difference degree of each production dataset group.

[0059] In one embodiment, the preset threshold includes a second preset threshold. Determining the anomaly detection result of each production dataset group according to the difference degree corresponding to each production dataset group and the preset threshold corresponding to the production dataset group includes:

[0060] Determining the maximum difference degree from the difference degrees corresponding to each production dataset group;

[0061] If the maximum difference degree is greater than the second preset threshold, it is determined that the production data of the battery is inconsistent in spatial distribution;

[0062] If the maximum difference degree is not greater than the second preset threshold, it is determined that the production data of the battery is consistent in spatial distribution.

[0063] The method provided in this embodiment, for the difference degree between the target production datasets in the spatial dimension, determines the consistency of production data in spatial distribution by the maximum difference degree, thereby determining the abnormal conditions in the production data, facilitating subsequent cause analysis to improve production efficiency and battery quality.

[0064] In one embodiment, the production data in the target production dataset includes at least one of the first liquid injection volume, the first helium leak detection rate, the self-discharge rate of the battery, the thickness of the battery, and the internal resistance of the battery.

[0065] The method provided in this embodiment can determine the difference degree of the target production dataset including at least one of the first liquid injection volume, the first helium leak detection rate, the self-discharge rate of the battery, the thickness of the battery, and the internal resistance of the battery, so as to realize more comprehensive difference detection of the production data of the battery and improve the accuracy of the obtained difference degree.

[0066] In a second aspect, the present application also provides a device for detecting the difference in production data of a battery. The device includes:

[0067] A first acquisition module, configured to acquire probability distributions corresponding to two target production data sets in a production data set group of a battery; the two target production data sets include a first production data set and a second production data set;

[0068] A first determination module, configured to determine a target ratio according to probability distributions corresponding to respective target production data sets in the production data set group;

[0069] A second determination module, configured to determine a difference degree between the first production data set and the second production data set according to probability distributions corresponding to at least one target production data set and the target ratio.

[0070] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0071] Acquire probability distributions corresponding to two target production data sets in a production data set group of a battery; the two target production data sets include a first production data set and a second production data set;

[0072] Determine a target ratio according to probability distributions corresponding to respective target production data sets in the production data set group;

[0073] Determine a difference degree between the first production data set and the second production data set according to probability distributions corresponding to at least one target production data set and the target ratio.

[0074] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0075] Acquire probability distributions corresponding to two target production data sets in a production data set group of a battery; the two target production data sets include a first production data set and a second production data set;

[0076] Determine a target ratio according to probability distributions corresponding to respective target production data sets in the production data set group;

[0077] Determine a difference degree between the first production data set and the second production data set according to probability distributions corresponding to at least one target production data set and the target ratio.

[0078] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0079] Obtain the probability distributions corresponding to two target production data sets in the production data set group of the battery; the two target production data sets include a first production data set and a second production data set;

[0080] Determine a target ratio according to the probability distributions corresponding to the target production data sets in the production data set group;

[0081] Determine the difference degree between the first production data set and the second production data set according to the probability distributions corresponding to at least one target production data set and the target ratio.

[0082] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following specifically lists the specific implementation manners of this application. Brief Description of the Drawings

[0083] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of this application. And in all the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0084] Figure 1 is the internal structure diagram of a computer device provided by an embodiment of this application;

[0085] Figure 2 is the flowchart of a method for detecting the difference in production data of a battery provided by an embodiment of this application;

[0086] Figure 3 is the flowchart of a method for determining a probability distribution provided by an embodiment of this application;

[0087] Figure 4 is one of the flowcharts of a method for determining the difference degree provided by an embodiment of this application;

[0088] Figure 5 is the flowchart of a method for determining a target ratio provided by an embodiment of this application;

[0089] Figure 6 is the second flowchart of a method for determining the difference degree provided by an embodiment of this application;

[0090] Figure 7 is the schematic diagram of the frequency distribution of production data provided by an embodiment of this application;

[0091] Figure 8 is the flowchart of a method for determining a production data set group provided by an embodiment of this application;

[0092] Figure 9 It is a schematic flowchart of a method for determining an abnormal detection result provided by an embodiment of the present application;

[0093] Figure 10 It is a schematic flowchart of another method for determining an abnormal detection result provided by an embodiment of the present application;

[0094] Figure 11 It is a schematic flowchart of another method for detecting differences in production data of a battery provided by an embodiment of the present application;

[0095] Figure 12 It is a structural block diagram of a device for detecting differences in production data of a battery provided by an embodiment of the present application;

[0096] Figure 13 It is a structural block diagram of a first acquisition module provided by an embodiment of the present application;

[0097] Figure 14 It is a structural block diagram of another device for detecting differences in production data of a battery provided by an embodiment of the present application;

[0098] Figure 15 It is a structural block diagram of a device for determining a production data set group provided by an embodiment of the present application. Detailed implementation manners

[0099] Next, embodiments of the technical solution of the present application will be described in detail with reference to the accompanying drawings. The following embodiments are only used to illustrate the technical solution of the present application more clearly, and therefore are only examples and cannot be used to limit the protection scope of the present application.

[0100] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above drawings are intended to cover non-exclusive inclusion.

[0101] In the description of the embodiments of the present application, technical terms such as "first" and "second" are only used to distinguish different objects and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity, specific order or primary-secondary relationship of the indicated technical features. In the description of the embodiments of the present application, "a plurality of" means more than two unless otherwise specifically defined.

[0102] Reference to "embodiment" in this document means that the specific features, structures, or characteristics described in connection with the embodiment may be included in at least one embodiment of the present application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein may be combined with other embodiments.

[0103] In the description of the embodiments of the present application, the term "and / or" is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this document generally represents an "or" relationship between the associated objects before and after.

[0104] In the description of the embodiments of the present application, the term "plurality" refers to two or more (including two). Similarly, "multiple groups" refers to two or more groups (including two groups), and "multiple pieces" refers to two or more pieces (including two pieces).

[0105] Performing differential detection on the data in the production environment can effectively improve production efficiency and battery quality, and reduce the uncertainty and risks in the production process.

[0106] Currently, when performing differential detection on two data sets, the median consistency detection method is usually used to perform differential detection on the data in the production environment to obtain the degree of difference, that is, to determine the median of each data set, and judge whether the medians of the two data sets are consistent. If they are consistent, it is determined that the two data sets are basically the same. If there is a difference in the medians of the two data sets, it is determined that the two data sets may be abnormal.

[0107] However, the median mainly focuses on the central position of the data, while ignoring the distribution shape and interval range of the data, resulting in a low accuracy of the degree of difference. For example, in semiconductor production data, in addition to the central tendency, the distribution shapes such as the peak value, width, and skewness of the data may also have an important impact on the stability of the production process and battery quality. Therefore, in actual situations, the medians of two sets of data may be the same, but their distribution shapes may be very different. For example, one is a spike distribution and the other is a flat distribution, which may mean different risks and problems in the production process.

[0108] To solve the above technical problems, the embodiments of the present application provide a method for differential detection of production data of batteries, which can be applied to, for example Figure 1The computer device shown can be a terminal. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for detecting differences in production data of a battery. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0109] Those skilled in the art can understand that Figure 1 the structure shown in is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different component layout.

[0110] Taking as an example, the embodiments of this application will be introduced more clearly. First, the calculation method of the degree of difference between two production data sets involved in the embodiments of this application will be introduced here.

[0111] First, we define a "difference metric" to quantify the degree of difference between two frequency distributions P and Q. This metric is based on a more fundamental metric, which we call the "basic difference metric". The basic difference metric focuses on the relative probabilities of P and Q on their respective events, and it measures the "information loss" from one frequency distribution to another. Specifically, for continuous random variables, we use the probability density functions and to describe the frequency distributions P and Q. Among them, the probability density function is also called the probability distribution, and the calculation formula of the basic difference metric is as follows:

[0112] Basic difference metric ( ) = (1)

[0113] This formula is actually calculating the information loss from the frequency distribution P to Q, which we call "one-way information loss". However, to ensure that our difference measure is symmetric, that is, the difference of P relative to Q is the same as the difference of Q relative to P, we need to calculate the "one-way information loss" from P to Q and from Q to P, and then take their average. The resulting symmetric difference measure is what we call the "symmetric difference measure", and its calculation formula is as follows:

[0114] Symmetric difference measure ( ) = (2)

[0115] where m(x) is the average probability density function of P and Q,

[0116] Although the "symmetric difference measure" can already well reflect the difference between two frequency distributions, its value may be large and not easy to intuitively understand. To obtain a more intuitive similarity index within the range of [0, 1], we introduce the "similarity distance". This distance is obtained by taking the square root of the symmetric difference measure:

[0117] Similarity distance ( ) = (3)

[0118] In this way, the "similarity distance" provides a standardized and intuitive similarity index. When two distributions are exactly the same, the "similarity distance" is 0; when two distributions are completely mutually exclusive, the "similarity distance" reaches the maximum value of 1. Through the above steps, not only can the difference between two probability distributions be quantified, but also their similarity degree can be understood in an intuitive way.

[0119] It should be noted that the calculation of the degree of difference involved in the embodiments of this application can adopt any one of the above formulas (1), (2), and (3), or other algorithms evolved based on the above formulas.

[0120] In one embodiment, as Figure 2 shown, Figure 2 is a schematic flowchart of a method for detecting the difference in production data of a battery provided by the embodiments of this application. Taking the application of this method to the Figure 1 computer device as an example for illustration, it includes the following steps S201 - S203:

[0121] S201, obtain the probability distributions corresponding to two target production data sets in the production data set group of the battery; the two target production data sets include the first production data set and the second production data set.

[0122] In a possible implementation, obtaining the probability distributions corresponding to two target production data sets in the production data set group of the battery may include the following steps:

[0123] Obtain the frequency sets corresponding to two target production data sets in the production data set group of the battery, and process each frequency set to obtain the probability distribution corresponding to each target production data set. The frequency set is determined according to the frequency of occurrence of each production data in the target production data set.

[0124] Among them, the production data in the target production data set may include production quality production data or production efficiency production data, etc. The frequency set corresponding to the target production data set refers to the data set composed of the number of occurrences of each production data in the target production data set.

[0125] Taking battery manufacturing as an example, the internal resistance of the battery is a key quality production data that directly affects the performance and safety of the battery. A machine on a certain production line produced a total of 100 batteries on the first day. The internal resistance of 30 batteries is 1, the internal resistance of 30 batteries is 2, and the internal resistance of 40 batteries is 3. Then the production data set 1 corresponding to the first day is [1, 2, 3], and the frequency set corresponding to this production data set 1 is [30, 30, 40]. This machine produced a total of 100 batteries on the second day. The internal resistance of 28 batteries is 1, the internal resistance of 32 batteries is 2, and the internal resistance of 40 batteries is 3. Then the production data set 2 corresponding to the second day is [1, 2, 3], and the frequency set corresponding to this production data set 2 is [28, 32, 40]. If the production data set 1 of the first day is used as the first production data set and the production data set 2 of the second day is used as the second production data set, then the frequency set corresponding to the first production data set is [30, 30, 40], and the frequency set corresponding to the second production data set is [28, 32, 40]. Then process the frequency set corresponding to the first production data set to obtain the probability distribution corresponding to the first production data set, and process the frequency set corresponding to the second production data set to obtain the probability distribution corresponding to the second production data set. For example, the probability distribution obtained by normalizing the frequency set corresponding to the first production data set may be [0.3, 0.3, 0.4], and the probability distribution obtained by normalizing the frequency set corresponding to the second production data set may be [0.28, 0.32, 0.4].

[0126] It should be noted that by dividing the frequency corresponding to each production data by the total frequency, the frequency corresponding to the production data is converted into a probability value, thereby forming a probability distribution. This conversion ensures that the sum of all data points is equal to 1, meeting the basic requirements of the probability distribution. This process not only balances the magnitudes of different features, reduces the improper influence of the numerical size on the analysis result, but also for frequency data, normalization essentially converts it into a probability distribution form, enabling data to be compared on a unified scale and improving the effectiveness of comparison.

[0127] In another possible implementation, obtaining the probability distributions corresponding to two target production data sets in the production data set group of the battery may include the following steps:

[0128] Obtain two initial production data sets and the initial frequency sets corresponding to the initial production data sets. If the initial production data in each initial production data set is inconsistent, determine the union of the initial production data in each initial production data set, determine the initial production data sets in which the initial production data is inconsistent with the union, and use the initial production data sets in which the initial production data is inconsistent with the union as the initial production data sets to be updated. For the initial production data sets to be updated, the initial production data sets to be updated need to be updated according to the union to obtain the target production data sets, and a preset value is filled into the target positions in the initial frequency sets corresponding to the target production data sets to obtain the frequency sets corresponding to the target production data sets; where the target positions are the positions in the initial frequency sets corresponding to the newly added production data in the target production data sets. For the initial production data sets in which the production data is consistent with the union, use the initial production data sets in which the production data is consistent with the union as the target production data sets, and use the initial frequency sets corresponding to the target production data sets as the frequency sets corresponding to the target production data sets.

[0129] After determining the frequency sets corresponding to the target production data sets through the above steps, if the frequency sets corresponding to the target production data sets are already probability distributions, directly use the frequency sets corresponding to the target production data sets as the probability distributions corresponding to the target production data sets. If the frequency sets corresponding to the target production data sets are not probability distributions, perform normalization processing on the frequency sets corresponding to the target production data sets to obtain the probability distributions corresponding to the target production data sets.

[0130] Exemplarily, the machine tool produced 1 battery on the third day, and the internal resistance of 1 battery was 1, so the initial production data set A on the third day was [1], and the initial frequency set A corresponding to the initial production data set A was [1]. The machine tool produced 1 battery on the fourth day, and the internal resistance of 1 battery was 2, so the initial production data set B on the fourth day was [2], and the initial frequency set B corresponding to the initial production data set B was [1]. Since the initial production data in the initial production data set A and the initial production data set B is inconsistent, it is determined that both the initial production data set A and the initial production data set B are the initial production data sets to be updated. Since the union of the two initial production data sets is [1, 2], both the initial production data set A and the initial production data set B are updated to [1, 2]. For the updated target production data set A [1, 2] of the initial production data set A, a preset value is filled into the target position in the initial frequency set A corresponding to the target production data set A to obtain the frequency set A [1, 0] corresponding to the target production data set A. The preset value is 0.

[0131] For the updated target production dataset B[1, 2] of the initial production dataset B, fill the preset value into the target position in the corresponding initial frequency set of the target production dataset B to obtain the frequency set B[0, 1] corresponding to the target production dataset B. The preset value is 0.

[0132] Since the frequency sets A[1, 0] and B[0, 1] are already probability distributions, directly use the frequency set A[1, 0] as the probability distribution corresponding to the target production dataset A, and use the frequency set B[1, 0] as the probability distribution corresponding to the target production dataset B.

[0133] S202. Determine the target ratio according to the probability distributions corresponding to the target production datasets in the production dataset group.

[0134] S203. Determine the difference degree between the first production dataset and the second production dataset according to the probability distributions corresponding to at least one target production dataset and the target ratio.

[0135] Either of the two target production datasets in the two target production datasets can be used as the first production dataset, and the other as the second production dataset. Denote the probability distribution of the first production dataset with and the probability distribution of the second production dataset with . Then, the difference degree between the first production dataset and the second production dataset can be determined through any of the following implementation manners.

[0136] In a possible implementation manner, determine the first ratio of the probability distribution of the first production dataset to the probability distribution of the second production dataset. The target ratio includes the first ratio. Based on the above formula (1), according to the probability distribution of the first production dataset and the first ratio, determine the basic difference measure between the first production dataset and the second production dataset, and use the basic difference measure as the difference degree between the first production dataset and the second production dataset.

[0137] In another possible implementation manner, determine the fourth ratio of the probability distribution of the second production dataset to the probability distribution of the first production dataset. The target ratio includes the fourth ratio. Based on the following formula (4), according to the probability distribution of the second production dataset and the fourth ratio, determine the basic difference measure between the first production dataset and the second production dataset, and use the basic difference measure as the difference degree between the first production dataset and the second production dataset.

[0138] Among them, the basic difference measure ( ) = (4)

[0139] In yet another possible implementation, the average probability distribution is determined according to the probability distributions corresponding to the production data sets, the second ratio of the probability distribution of the first production data set to the target probability distribution, and the third ratio of the probability distribution of the second production data set to the target probability distribution are determined; the target ratios include the second ratio and the third ratio; based on the above formulas (2) and (3), the similarity distance between the first production data set and the second production data set is determined, and the similarity distance is used as the degree of difference between the first production data set and the second production data set.

[0140] The method provided in this embodiment does not need to rely on the normal distribution assumption of data, so it can be widely applied to various data distribution forms, realizing difference detection under any data distribution conditions, and improving the generality and application scope of the algorithm.

[0141] The method provided in this embodiment can quantify the overlap degree and shape similarity of the interval ranges of production data by obtaining the probability distributions corresponding to two target production data sets in the production data set group of the battery, and determining the target ratios according to the probability distributions corresponding to each target production data set in the production data set group, and determining the degree of difference between the first production data set and the second production data set according to the probability distributions corresponding to at least one target production data set and the target ratios, providing an in-depth data analysis perspective and improving the accuracy of determining the degree of difference between two data sets.

[0142] In one embodiment, as Figure 3 shown, Figure 3 is a schematic flowchart of a method for determining a probability distribution provided by an embodiment of the present application. This embodiment relates to a possible implementation of how to obtain the probability distributions corresponding to two target production data sets in the production data set group of the battery. On the basis of the above embodiment, the method may include the following steps S301-S302:

[0143] S301, obtain the frequency sets corresponding to two target production data sets in the production data set group of the battery; the frequency sets are determined according to the frequencies of occurrence of each production data in the target production data set.

[0144] In this embodiment, two initial production data sets and the corresponding initial frequency sets of the initial production data sets can be obtained. If the initial production data in each initial production data set are inconsistent, the union of the initial production data in each initial production data set is determined, and the initial production data set in which the initial production data is inconsistent with the union is determined. The initial production data set in which the initial production data is inconsistent with the union is used as the initial production data set to be updated. For the initial production data set to be updated, the initial production data set to be updated needs to be updated according to the union to obtain the target production data set, and a preset value is filled into the target position in the corresponding initial frequency set of the target production data set to obtain the frequency set corresponding to the target production data set; where the target position is the position in the initial frequency set corresponding to the newly added production data in the target production data set. For the initial production data set in which the production data is consistent with the union, the initial production data set in which the production data is consistent with the union is used as the target production data set, and the corresponding initial frequency set of the target production data set is used as the frequency set corresponding to the target production data set.

[0145] S302. Process each frequency set to obtain the probability distribution corresponding to each target production data set.

[0146] If the frequency set corresponding to the target production data set is not a probability distribution, perform normalization processing on the frequency set corresponding to the target production data set to obtain the probability distribution corresponding to the target production data set.

[0147] The method provided in this embodiment obtains the frequency sets corresponding to two target production data sets in the production data set group of the battery, processes each frequency set to obtain the probability distribution corresponding to each target production data set, thereby balancing the magnitudes of different features, reducing the improper influence of the numerical size on the analysis result, and further improving the accuracy of the obtained degree of difference.

[0148] In one embodiment, for S202 above, according to the probability distribution corresponding to each production data set in the production data set group, determining the target ratio can be implemented in the following manner:

[0149] Determine the first ratio of the probability distribution of the first production data set to the probability distribution of the second production data set; the target ratio includes the first ratio.

[0150] If the probability distribution of the first production data set is represented by and the probability distribution of the second production data set is represented by , then the first ratio is equal to .

[0151] The method provided in this embodiment lays a foundation for determining the degree of difference based on the target ratio by determining the first ratio of the probability distribution of the first production data set to the probability distribution of the second production data set, and taking the first ratio as the target ratio, thereby improving the accuracy of determining the degree of difference between the two data sets.

[0152] In one embodiment, if the first ratio of the probability distribution of the first production data set to the probability distribution of the second production data set is determined, correspondingly, for the above S203, determining the degree of difference between the first production data set and the second production data set according to the probability distribution corresponding to at least one target production data set and the target ratio can be implemented in the following manner. Refer to Figure 4 , Figure 4 is one of the flow diagrams of the method for determining the degree of difference provided in the embodiments of the present application. The method may include the following steps S401 - S402:

[0153] S401, determine the first logarithm of the first ratio, and determine the first product of the first logarithm and the probability distribution of the first production data set.

[0154] S402, perform an integration process on the first product within the production data interval range corresponding to the target production data set to obtain a first integration result, and take the first integration result as the degree of difference.

[0155] Combined with the above example, take the production data set 1 on the first day as the first production data set, and the production data set 2 on the second day as the second production data set. Both the production data set 1 and the production data set 2 are [1, 2, 3]. Therefore, the production data interval range of the first production data set and the second production data set is [1, 3]. Perform an integration process on the first product within this production data interval range to obtain a first integration result, and take the first integration result as the degree of difference.

[0156] The method provided in this embodiment improves the accuracy of determining the degree of difference between the two data sets by determining the first logarithm of the first ratio, determining the first product of the first logarithm and the probability distribution of the first production data set, and performing an integration process on the first product within the production data interval range corresponding to the target production data set to obtain a first integration result and taking the first integration result as the degree of difference.

[0157] In one embodiment, refer to Figure 5 , Figure 5 is the flow diagram of a method for determining the target ratio provided in the embodiments of the present application. For the above S202, determining the target ratio according to the probability distribution corresponding to each target production data set in the production data set group may include the following steps S501 - S502:

[0158] S501. Determine the target probability distribution according to the probability distributions corresponding to the respective target production data sets.

[0159] In a possible implementation, the average probability distribution of the probability distribution of the first production data set and the probability distribution of the second production data set can be determined, and the average probability distribution is used as the target probability distribution.

[0160] In another possible implementation, if the two probability distributions have corresponding weights, the weighted average probability distribution of the probability distribution of the first production data set and the probability distribution of the second production data set can be calculated, and the weighted average probability distribution is used as the target probability distribution.

[0161] S502. Determine the second ratio of the probability distribution of the first production data set to the target probability distribution, and the third ratio of the probability distribution of the second production data set to the target probability distribution; the target ratios include the second ratio and the third ratio.

[0162] The method provided in this embodiment, by determining the second ratio of the probability distribution of the first production data set to the target probability distribution and the third ratio of the probability distribution of the second production data set to the target probability distribution, and using the second ratio and the third ratio as the target ratios, thus lays a foundation for determining the degree of difference based on the target ratios, and improves the accuracy of determining the degree of difference between the two data sets.

[0163] In one embodiment, if the target ratios including the second ratio and the third ratio are determined, then correspondingly, the above S203. Determine the degree of difference between the first production data set and the second production data set according to the probability distribution corresponding to at least one target production data set and the target ratios can be implemented in the following manner. Refer to Figure 6 , Figure 6 is the second flow chart of the method for determining the degree of difference provided by the embodiment of the present application. The method may include the following steps S601 - S604:

[0164] S601. Determine the second logarithm of the second ratio, and determine the second product of the second logarithm and the probability distribution of the first production data set.

[0165] S602. Determine the third logarithm of the third ratio, and determine the third product of the third logarithm and the probability distribution of the second production data set.

[0166] S603. Integrate the second product within the production data interval corresponding to the target production data set to obtain a second integration result, and integrate the third product within the production data interval to obtain a third integration result.

[0167] S604. Determine the degree of difference according to the second integration result and the third integration result.

[0168] In a possible implementation, the second integration result and the third integration result can be determined in combination with the above formula (2), and the average value of the second integration result and the third integration result can be determined. This average value is the symmetric difference measure, and this average value is used as the degree of difference.

[0169] In another possible implementation, the square root of this average value can be calculated in combination with the above formula (3). This square root is the "similarity distance" defined in formula (3), and the square root is used as the degree of difference. The similarity distance can quantify the overlap degree and shape similarity of the index values, including detailed features such as peaks, widths, and slopes. This can not only evaluate the time change trend macroscopically, but also identify the distribution differences in the spatial dimension at the microscopic level, thereby providing a more in-depth and comprehensive data analysis perspective and improving the accuracy and comprehensiveness of difference determination.

[0170] The "similarity distance" provides a standardized and intuitive similarity index. When two distributions are exactly the same, the "similarity distance" is 0; when two distributions are completely mutually exclusive, the "similarity distance" reaches the maximum value of 1. Through the above steps, not only can the difference between two probability distributions be quantified, but also their similarity degree can be understood in an intuitive way. As Figure 7 shown, Figure 7 is a schematic diagram of the frequency distribution of production data provided by an embodiment of the present application. Among them, the abscissa is the production data, and the ordinate is the frequency. Figure 7 The similarity distance between the two distributions shown in (a) and (b) in it is 0.6169. It can be observed that the difference between the two data distributions is very large at this value.

[0171] The method provided by this embodiment determines the degree of difference according to the second integration result and the third integration result. This degree of difference can quantify the overlap degree and shape similarity of the interval range of production data, thereby improving the accuracy of determining the degree of difference between two data sets.

[0172] In one embodiment, the number of production data set groups is multiple, and the method further includes:

[0173] According to the degree of difference corresponding to each production data set group and the preset threshold corresponding to the production data set group, determine the anomaly detection result of each production data set group.

[0174] The target production dataset in the production dataset group can be a target production dataset in the time dimension or a target production dataset in the space dimension. The target production dataset in the time dimension can include target production datasets corresponding to multiple different timestamps under the same production line. The target production dataset in the space dimension can include target production datasets corresponding to the same battery under different production lines, or can include target production datasets corresponding to different machines under the same production line for the same battery. Among them, the timestamp can be a timestamp such as day, week, month, etc.

[0175] It should be noted that since different batteries may have different definitions of anomalies for different production data, it is necessary to set a preset threshold after analyzing a large amount of historical production data and business requirements. However, generally, when the similarity distance exceeds 0.6, it means that there is a large deviation in the overall data and a large change in the data distribution shape. Therefore, the preset thresholds in the space-time dimension can both be set to 0.6, or the preset threshold corresponding to the production dataset group in the time dimension is different from the preset threshold corresponding to the production dataset group in the space dimension.

[0176] In a possible implementation manner, taking the target production dataset in the time dimension as an example, the target production dataset corresponding to the maximum timestamp can be combined with the target production datasets corresponding to any other timestamp to obtain multiple production dataset groups, and the median of the difference degrees corresponding to each production dataset group is determined. If the median is greater than the preset threshold corresponding to the production dataset group, it is determined that the anomaly detection result is that the change trend of the production data of the production line over time is abnormal. If the median is not greater than the preset threshold corresponding to the production dataset group, it is determined that the anomaly detection result is that the change trend of the production data of the production line over time is normal. Among them, for the convenience of subsequent exemplary description, the initial production datasets and target production datasets corresponding to multiple different timestamps under the same production line are hereinafter referred to as production datasets in the time dimension, and the preset threshold corresponding to the production dataset group in the time dimension is the first preset threshold.

[0177] Exemplarily, production line A corresponds to target production dataset 1 on January 1, 2024, target production dataset 2 on January 2, 2024, target production dataset 3 on January 3, 2024, …, target production dataset 10 on January 10, 2024. Then, the production dataset group can include 9. Among them, production dataset group 1 consists of target production dataset 1 and target production dataset 10, production dataset group 2 consists of target production dataset 2 and target production dataset 10, production dataset group 3 consists of target production dataset 3 and target production dataset 10, …, production dataset group 9 consists of target production dataset 9 and target production dataset 10. The median of the difference degrees corresponding to the 9 production dataset groups can be determined. If the median is greater than the first preset threshold, it is determined that the abnormal detection result is that the change trend of the production data of the production line over time is abnormal. If the median is not greater than the first preset threshold, it is determined that the abnormal detection result is that the change trend of the production data of the production line over time is normal.

[0178] In another possible implementation, if the target production dataset in the production dataset group includes the target production dataset in the spatial dimension, and the target production dataset includes the target production datasets corresponding to different machines of the same battery under the same production line, then any two target production datasets are combined into a production dataset group. The maximum difference degree can be determined from the difference degrees corresponding to each production dataset group. If the maximum difference degree is greater than the second preset threshold, it is determined that the production data of the battery is inconsistent in spatial distribution; if the maximum difference degree is not greater than the second preset threshold, it is determined that the production data of the battery is consistent in spatial distribution. Among them, the first preset threshold and the second preset threshold can be the same or different.

[0179] The method provided in this embodiment can determine the abnormal detection results of each production dataset group by according to the difference degrees corresponding to each production dataset group and the preset threshold corresponding to the production dataset group, so as to analyze the change trend of the target production dataset in at least two production dataset groups in the time dimension and the consistency in the spatial dimension.

[0180] In one embodiment, as Figure 8 shown, Figure 8 is a schematic flowchart of a method for determining a production dataset group provided by an embodiment of the present application. The method includes the following steps S801 - S804:

[0181] S801, obtain each initial production dataset of the battery.

[0182] Exemplarily, if the initial production dataset in the time dimension is obtained, a total of 6 initial production datasets are obtained, and the initial production data is the battery internal resistance. The 6 initial production datasets include the initial production dataset 1 corresponding to production line A on January 1, 2024, the initial production dataset 2 corresponding to production line A on January 2, 2024, the initial production dataset 3 corresponding to production line A on January 3, 2024, …, the initial production dataset 6 corresponding to production line A on January 6, 2024. Among them, the initial production dataset 1 is [1, 2, 3], the initial production dataset 2 is [1, 2, 3], the initial production dataset 3 is [1, 2, 3, 4], the initial production dataset 4 is [1, 2, 3], the initial production dataset 5 is [2, 3, 4], and the initial production dataset 6 is [1, 2, 3, 4].

[0183] S802, if the initial production data in each initial production dataset is inconsistent, then determine the union of the initial production data in each initial production dataset.

[0184] Combined with the example in S801 above, it can be seen that the initial production data in the 6 initial production datasets is inconsistent, and the union of the initial production data in the 6 initial production datasets is [1, 2, 3, 4].

[0185] S803, update the initial production dataset to be updated according to the union to obtain the updated production dataset corresponding to the initial production dataset to be updated; the initial production dataset to be updated includes the initial production dataset whose initial production data is inconsistent with the union.

[0186] The initial production dataset to be updated includes the initial production dataset 1, the initial production dataset 2, the initial production dataset 4, and the initial production dataset 5. After updating the initial production dataset 1, the initial production dataset 2, the initial production dataset 4, and the initial production dataset 5 according to the union, the updated production datasets corresponding to the 4 initial production datasets are all [1, 2, 3, 4], that is, the initial production data in the updated production dataset is consistent with the union. By filling the initial production dataset whose initial production data is inconsistent with the union, the consistency of the production data in each target production dataset is achieved.

[0187] S804, determine each production dataset group according to each target production dataset; each target production dataset includes at least the updated production dataset.

[0188] For an initial production data set with inconsistent initial production data and union set, after updating the initial production data set, the obtained target production data set includes the updated production data set, and may include the initial production data set with consistent initial production data and union set. As in the above example, the obtained target production data set includes the initial production data set with consistent initial production data and union set, that is, it includes the initial production data set 3 and the initial production data set 6.

[0189] For an initial production data set without inconsistent initial production data and union set, each initial production data set is used as each target production data set.

[0190] After obtaining each target production data set, each production data set group can be determined according to each target production data set.

[0191] The method provided in this embodiment determines the union set of the initial production data in each initial production data set when the initial production data in each initial production data set is inconsistent, and updates the initial production data set to be updated according to the union set, so as to make the data ranges of the production data in each target production data set consistent, and further realize determining the difference degree between two target production data sets in the production data set group within the same data range.

[0192] In one embodiment, S301 above, obtaining the frequency sets corresponding to two target production data sets in the production data set group of the battery can be implemented in the following manner:

[0193] If the two target production data sets include the updated production data set, fill the preset value into the target position in the initial frequency set corresponding to the updated production data set to obtain the frequency set corresponding to the updated production data set; the target position is the position in the initial frequency set corresponding to the newly added production data in the updated production data set;

[0194] If the two target production data sets include the initial production data set, use the initial frequency set corresponding to the initial production data set as the frequency set corresponding to the initial production data set.

[0195] Exemplarily, for the initial production dataset 1 being [1, 2, 3], if the initial frequency set corresponding to the initial production dataset 1 is [30, 30, 40], and the target production dataset corresponding to the initial production dataset 1 is [1, 2, 3, 4], the newly added production data in the target production dataset is 4 and is located at the last position in this dataset. Then, a preset value is added to the end of the initial frequency set [30, 30, 40]. If the preset value is 0, the frequency set corresponding to the obtained target production dataset [1, 2, 3, 4] is [30, 30, 40, 0]. Herein, the preset value can also be an extremely small value that is very close to 0. By filling data into the initial frequency set corresponding to the updated production dataset, the dimensional consistency of the frequency sets corresponding to each obtained target production dataset is achieved, such that the frequency sets corresponding to each target production dataset are on the same scale.

[0196] For the initial production dataset 6 [1, 2, 3, 4], the production data of the initial production dataset 6 is the same as the union. If the initial frequency set corresponding to the initial production dataset 6 is [25, 30, 40, 5], then this initial production dataset is used as the target production dataset, and the initial frequency set is used as the frequency set of the target production dataset.

[0197] For the method provided in this embodiment, if two target production datasets include the updated production dataset, the preset value is filled into the target position in the initial frequency set corresponding to the updated production dataset to obtain the frequency set corresponding to the updated production dataset, which can reduce the degree of analysis deviation caused by data missing, and further improve the accuracy of the obtained degree of difference.

[0198] In one embodiment, each initial production dataset includes initial production datasets corresponding to multiple different timestamps under the same production line; then, for the above S804, determining each production dataset group according to each target production dataset can be achieved in the following manner:

[0199] The target production dataset corresponding to the maximum timestamp is used as the first production dataset in the production dataset group, and the target production dataset corresponding to a target timestamp is used as the second production dataset in the production dataset group to obtain each production dataset group;

[0200] Wherein, the target timestamp includes timestamps other than the maximum timestamp among all timestamps.

[0201] As introduced in the above example, for example, production line A corresponds to the target production dataset 1 on January 1, 2024, production line A corresponds to the target production dataset 2 on January 2, 2024, production line A corresponds to the target production dataset 3 on January 3, 2024, …, production line A corresponds to the target production dataset 10 on January 10, 2024. The maximum timestamp is January 10, 2024, and the composed production dataset group can include 9. Among them, production dataset group 1 is composed of target production dataset 1 and target production dataset 10, production dataset group 2 is composed of target production dataset 2 and target production dataset 10, production dataset group 3 is composed of target production dataset 3 and target production dataset 10, …, production dataset group 9 is composed of target production dataset 9 and target production dataset 10.

[0202] The method provided in this embodiment, for the target production dataset in the time dimension, uses the target production dataset corresponding to the maximum timestamp as the first production dataset in the production dataset group, and the target production dataset corresponding to a target timestamp as the second production dataset in the production dataset group, so as to obtain each production dataset group, laying a foundation for further determining the change trend in the time dimension based on the difference degree of each production dataset group.

[0203] In one embodiment, as Figure 9 shown, Figure 9 is a schematic flowchart of a method for determining an anomaly detection result provided by an embodiment of the present application. The preset threshold in this embodiment includes a first preset threshold. The above-mentioned determining the anomaly detection result of each production dataset group according to the difference degree corresponding to each production dataset group and the preset threshold corresponding to the production dataset group may include the following steps S901 - S903:

[0204] S901, determine the median of each difference degree according to the difference degree corresponding to each production dataset group.

[0205] In this embodiment, in time series analysis, taking the median of all difference degrees as a comparison index means focusing on the "central tendency" of the differences between the data of all timestamps and the data of the maximum timestamp, rather than the average difference affected by the data of individual abnormal timestamps. If the median is small, it means that the production data of most timestamps has little difference from the production data of the maximum timestamp, indicating that the trend of production data changing with time is relatively stable. On the contrary, if the median is large, it means that the trend of production data changing with time is relatively significant, and there may be abnormal changes or turning points in the production data, and the reasons need to be further analyzed.

[0206] S902, if the median is greater than the first preset threshold, then determine that the anomaly detection result is that the change trend of the production data of the production line with time is abnormal.

[0207] S903, if the median is not greater than the first preset threshold, then determine that the abnormal detection result is that the changing trend of the production data of the production line over time is normal.

[0208] In the case where it is determined that the abnormal detection result is that the changing trend of the production data of the production line over time is abnormal, the system can generate an alarm and output the first abnormal information, and the first abnormal information can include at least one of a timestamp, an abnormal production data set group, an abnormal degree, etc., so that relevant personnel can analyze the cause of the abnormality based on the abnormal information, such as whether it is caused by an external event, to adjust the subsequent data collection strategy or model parameters.

[0209] The method provided in this embodiment, aiming at the difference degree between target production data sets in the time dimension, determines the changing trend of the production data of the production line over time through the median, so as to determine the abnormal changes or turning points in the production data, which is convenient for subsequent cause analysis to improve production efficiency and battery quality.

[0210] In one embodiment, each initial production data set includes the initial production data sets corresponding to the same battery on different production lines, or includes the initial production data sets corresponding to different machines of the same battery on the same production line. Then, for the above S904, determining each production data set group according to each target production data set can be implemented in the following manner:

[0211] Take any two target production data sets in each target production data set as a production data set group to obtain each production data set group.

[0212] In this embodiment, each initial production data set includes the initial production data sets corresponding to the same battery on different production lines, or includes the initial production data sets corresponding to different machines of the same battery on the same production line, that is, the initial production data set is the initial production data set in the space dimension. The initial production data set in the space dimension includes two types of initial production data sets in the space dimension. One is the initial production data set corresponding to the same battery on different production lines, and the other is the initial production data set corresponding to different machines of the same battery on the same production line. The initial production data sets of the same type in the space dimension are data sets within the same time. Exemplarily, each initial production data set includes the initial production data set of battery A corresponding to production line A on March 1, 2024, the initial production data set of battery A corresponding to production line B on March 1, 2024, and the initial production data set of battery A corresponding to production line C on March 1, 2024. The initial production data sets in the space dimension to be analyzed can be determined according to requirements. It is possible to analyze the initial production data sets of the two space dimensions, or analyze the initial production data sets of one of the space dimensions. For the production data in the space dimension, any two target production data sets in each target production data set can be taken as a production data set group to obtain each production data set group.

[0213] The method provided in this embodiment, for the target production dataset in the spatial dimension, takes any two target production datasets in each target production dataset as a production dataset group to obtain each production dataset group, laying a foundation for further determining the consistency of production data in the spatial dimension based on the difference degree of each production dataset group.

[0214] In one embodiment, as Figure 10 shown, Figure 10 FIG. is a schematic flowchart of another method for determining an anomaly detection result provided by an embodiment of the present application. The preset threshold in this embodiment includes a first preset threshold. The above-mentioned determining the anomaly detection result of each production dataset group according to the difference degree corresponding to each production dataset group and the preset threshold corresponding to the production dataset group may include the following steps S1001 - S1003:

[0215] S1001, determine the maximum difference degree from the difference degrees corresponding to each production dataset group.

[0216] S1002, if the maximum difference degree is greater than the second preset threshold, determine that the production data of the battery is inconsistent in spatial distribution.

[0217] S1003, if the maximum difference degree is not greater than the second preset threshold, determine that the production data of the battery is consistent in spatial distribution.

[0218] In this embodiment, by traversing all production dataset groups and calculating the difference degree, it is possible to find the production dataset group with the maximum difference degree. This means that among all possible production dataset groups, the production dataset group with the maximum difference degree shows the greatest inconsistency or difference in spatial distribution.

[0219] The maximum difference degree and its corresponding production dataset group provide key information about the consistency of the spatial distribution in the dataset. If the maximum difference degree is small, it indicates that the production data in each production dataset group is relatively consistent in spatial distribution; on the contrary, if the maximum difference degree is large, it indicates that there are significant spatial distribution differences in the production data of each production dataset group, which may imply the internal structural complexity or potential anomalies of the production dataset group.

[0220] For example, if the production dataset group includes dataset group 1 composed of the target production dataset corresponding to the same model of battery on production line A and the target production dataset corresponding to the same model of battery on production line B; dataset group 2 composed of the target production dataset corresponding to the same model of battery on production line A and the target production dataset corresponding to the same model of battery on production line C; dataset group 3 composed of the target production dataset corresponding to the same model of battery on production line B and the target production dataset corresponding to the same model of battery on production line C. If the degree of difference corresponding to dataset group 1 is the largest and this degree of difference is greater than the second preset threshold, it is considered that there are abnormal differences in the spatial distribution of the internal resistance of the battery on these three production lines, and the consistency of this model of battery is poor.

[0221] In the case of determining that the production data of the battery is inconsistent in spatial distribution, the system can generate an alarm and output the second abnormal information. The second abnormal information can include at least one of a timestamp, an abnormal production dataset group, an abnormal degree, etc., so that relevant personnel can analyze the cause of the abnormality based on the abnormal information. For example, check the data source and processing flow of the abnormal production dataset group, and it may be necessary to recalibrate the sensor or update the data processing logic, etc.

[0222] The method provided in this embodiment determines the consistency of production data in spatial distribution by the largest degree of difference for the degree of difference between target production datasets in the spatial dimension, thereby determining the abnormal conditions occurring in the production data, facilitating subsequent cause analysis to improve production efficiency and battery quality.

[0223] In the current related technologies, if the detection method ignores the spatio-temporal dimension, it may miss key abnormal data, affecting battery quality. Changes in equipment status or environment during the production cycle are not captured, easily leading to unqualified products. Differences between different production lines are not considered, resulting in inconsistent battery quality. In this embodiment, for the production data in the time dimension and the production data in the spatial dimension of the same production line, the embodiments provided in this application can be used to analyze the production data in the time dimension of the same production line and analyze the production data in the spatial dimension, thereby improving the comprehensiveness of data detection, being able to capture abnormal distributions of production data caused by changes in equipment status or environment during the production cycle, and the differences in production data distributions between different production lines, reducing the probability of missing key abnormal data due to one-sided detection methods, thereby improving the consistency of battery quality and production efficiency. It should be noted that, according to actual needs, the production data in a certain dimension can also be analyzed.

[0224] In one embodiment, the production data in the target production dataset includes at least one of the first injection volume, the first helium leak detection rate, the self-discharge rate of the battery, the thickness of the battery, and the internal resistance of the battery.

[0225] The primary helium leak rate refers to the amount of helium gas escaping from the battery under test per unit time measured during a primary helium leak detection process.

[0226] The method provided in this embodiment can determine the degree of difference of a target production data set including at least one production data such as the primary liquid injection volume, the primary helium leak rate, the self-discharge rate of the battery, the battery thickness, and the internal resistance of the battery, so as to achieve a more comprehensive difference detection of the production data of the battery and improve the accuracy of the obtained degree of difference.

[0227] Refer to Figure 11 , Figure 11 is a schematic flowchart of another method for detecting differences in production data of a battery provided in an embodiment of the present application. The method includes obtaining an initial production data set, filling data for the initial production data set and the corresponding initial frequency set of the initial production data set to obtain a target production data set and the corresponding frequency set of the target production data set, and performing a normalization process on the frequency set corresponding to the target production data set to obtain the probability distribution corresponding to the target frequency set. For the target production data set in the time dimension, calculate the degree of difference between the target production data sets corresponding to other timestamps and the target production data set corresponding to the maximum timestamp, determine the median among the degrees of difference, compare the median with a first preset threshold to obtain an anomaly detection result, and output a first anomaly message when the anomaly detection result indicates that the change trend of the production data over time is abnormal. For the target production data set in the spatial dimension, calculate the degree of difference between any two target production data sets, determine the maximum degree of difference, compare the maximum degree of difference with a second preset threshold to obtain an anomaly detection result, and output a second anomaly message when the anomaly detection result indicates that the production data is inconsistent in spatial distribution.

[0228] It should be understood that although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise clearly stated in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or steps in other steps.

[0229] Based on the same inventive concept, an embodiment of the present application further provides a production data difference detection device for a battery for implementing the production data difference detection method for the battery involved above. The solution provided by this device for solving problems is similar to the solution described in the above method. Therefore, the specific limitations in one or more embodiments of the production data difference detection device for a battery provided below can refer to the limitations on the production data difference detection method for a battery in the above text, and will not be repeated here.

[0230] In one embodiment, as Figure 12 shown, Figure 12 FIG. is a structural block diagram of a production data difference detection device for a battery provided by an embodiment of the present application. The device 1200 includes:

[0231] A first acquisition module 1201, configured to acquire probability distributions corresponding to two target production data sets in a production data set group of a battery; the two target production data sets include a first production data set and a second production data set;

[0232] A first determination module 1202, configured to determine a target ratio according to probability distributions corresponding to each target production data set in the production data set group;

[0233] A second determination module 1203, configured to determine the difference degree between the first production data set and the second production data set according to probability distributions corresponding to at least one target production data set and the target ratio.

[0234] In one embodiment, as Figure 13 shown, Figure 13 FIG. is a structural block diagram of a first acquisition module provided by an embodiment of the present application. The first acquisition module 1201 includes:

[0235] A first acquisition unit 1301, configured to acquire frequency sets corresponding to two target production data sets in a production data set group of a battery; the frequency sets are determined according to the frequencies of occurrence of each production data in the target production data set;

[0236] A processing unit 1302, configured to process each frequency set to obtain probability distributions corresponding to each target production data set.

[0237] In one embodiment, the first determination module 1202 is specifically configured to determine a first ratio of the probability distribution of the first production data set to the probability distribution of the second production data set; the target ratio includes the first ratio.

[0238] In one embodiment, the second determination module 1203 is specifically configured to determine the first logarithm of the first ratio, and determine the first product of the first logarithm and the probability distribution of the first production data set; perform an integration process on the first product within the production data interval corresponding to the target production data set to obtain a first integration result, and use the first integration result as the degree of difference.

[0239] In one embodiment, the first determination module 1202 is specifically configured to determine a target probability distribution according to the probability distributions corresponding to each target production data set; determine a second ratio of the probability distribution of the first production data set to the target probability distribution, and a third ratio of the probability distribution of the second production data set to the target probability distribution; the target ratios include the second ratio and the third ratio.

[0240] In one embodiment, the second determination module 1203 is specifically configured to determine the second logarithm of the second ratio, and determine the second product of the second logarithm and the probability distribution of the first production data set; determine the third logarithm of the third ratio, and determine the third product of the third logarithm and the probability distribution of the second production data set; perform an integration process on the second product within the production data interval corresponding to the target production data set to obtain a second integration result, and perform an integration process on the third product within the production data interval to obtain a third integration result; determine the degree of difference according to the second integration result and the third integration result.

[0241] In one embodiment, the number of production data set groups is multiple, as Figure 14 shown, Figure 14 is a structural block diagram of another production data difference detection device for a battery provided by an embodiment of the present application. The device 1400 includes:

[0242] A third determination module 1401, configured to determine an anomaly detection result for each production data set group according to the degree of difference corresponding to each production data set group and a preset threshold corresponding to the production data set group.

[0243] In one embodiment, as Figure 15 shown, Figure 15 is a structural block diagram of a production data set group determination device provided by an embodiment of the present application. The device 1500 includes:

[0244] A second acquisition module 1501, configured to acquire each initial production data set of the battery;

[0245] A fourth determination module 1502, configured to determine the union of the initial production data in each initial production data set if the initial production data in each initial production data set is inconsistent;

[0246] An update module 1503, configured to update the initial production dataset to be updated according to the union set to obtain an updated production dataset corresponding to the initial production dataset to be updated; the initial production dataset to be updated includes the initial production dataset that is inconsistent with the union set.

[0247] A fifth determination module 1504, configured to determine each production dataset group according to each target production dataset; each target production dataset includes at least the updated production dataset.

[0248] In one embodiment, the first acquisition module 1201 is specifically configured to, if the updated production dataset is included in two target production datasets, fill a preset value into the target position in the initial frequency set corresponding to the updated production dataset to obtain the frequency set corresponding to the updated production dataset; the target position is the position in the initial frequency set corresponding to the newly added production data in the updated production dataset; if the initial production dataset is included in two target production datasets, use the initial frequency set corresponding to the initial production dataset as the frequency set corresponding to the initial production dataset.

[0249] In one embodiment, each initial production dataset includes initial production datasets corresponding to multiple different timestamps under the same production line.

[0250] The fifth determination module 1504 is specifically configured to use the target production dataset corresponding to the maximum timestamp as the first production dataset in the production dataset group, and use the target production dataset corresponding to a target timestamp as the second production dataset in the production dataset group, so as to obtain each production dataset group.

[0251] Wherein, the target timestamp includes the timestamps other than the maximum timestamp among all timestamps.

[0252] In one embodiment, the preset threshold includes a first preset threshold. The third determination module 1401 is specifically configured to determine the median of each degree of difference according to the degree of difference corresponding to each production dataset group; if the median is greater than the first preset threshold, determine that the abnormal detection result is that the change trend of the production data of the production line over time is abnormal; if the median is not greater than the first preset threshold, determine that the abnormal detection result is that the change trend of the production data of the production line over time is normal.

[0253] In one embodiment, each initial production dataset includes initial production datasets corresponding to the same battery under different production lines, or includes initial production datasets corresponding to the same battery under different machines on the same production line; the fifth determination module 1504 is specifically configured to use any two target production datasets in each target production dataset as a production dataset group, so as to obtain each production dataset group.

[0254] In one embodiment, the preset threshold includes a second preset threshold. The third determination module 1401 is specifically configured to determine the maximum degree of difference from the degrees of difference corresponding to each production data set group. If the maximum degree of difference is greater than the second preset threshold, it is determined that the production data of the battery is inconsistent in spatial distribution. If the maximum degree of difference is not greater than the second preset threshold, it is determined that the production data of the battery is consistent in spatial distribution.

[0255] Each module in the above-mentioned production data difference detection device of the battery can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor in the computer device in hardware form or be independent of the processor, or can be stored in the memory in the computer device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0256] In one embodiment, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps of any of the above embodiments are implemented.

[0257] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the steps of any of the above embodiments are implemented.

[0258] In one embodiment, a computer program product is provided, including a computer program. When the computer program is executed by the processor, the steps of any of the above embodiments are implemented.

[0259] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0260] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0261] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0262] The above-described embodiments merely represent several implementation manners of the present application. Their descriptions are relatively specific and detailed, but they should not be construed as limiting the patent scope of the present application. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for detecting differences in production data of a battery, characterized in that: The method comprises: Obtaining probability distributions corresponding to two target production data sets in a production data set group of a battery; the two target production data sets include a first production data set and a second production data set; if the production data set group includes a target production data set in a time dimension, then the multiple production data set groups are obtained by combining the target production data set corresponding to the maximum timestamp among multiple different timestamps on the same production line with the target production data set corresponding to any other timestamp; if the production data set group includes a target production data set in a spatial dimension, then the multiple production data set groups are obtained by combining any two target production data sets on different production lines corresponding to the same time as one production data set group, or by combining any two target production data sets on different machines on the same production line corresponding to the same time as one production data set group; Determining a target ratio according to a probability distribution corresponding to each of the target production data sets in the production data set group; determining a degree of difference between the first production data set and the second production data set according to a probability distribution corresponding to at least one of the target production data sets and the target ratio; The anomaly detection result of each of the production data set groups is determined according to the difference degree corresponding to each of the production data set groups and the preset threshold value corresponding to the production data set group.

2. The method according to claim 1, characterized in that The obtaining of probability distributions corresponding to two target production data sets in the battery production data set group includes: Obtaining frequency sets corresponding to two target production data sets in the production data set group of the battery; the frequency sets are determined according to the frequency of occurrence of each production data in the target production data set; Each of the frequency sets is processed to obtain a probability distribution corresponding to each of the target production data sets.

3. The method according to claim 1, characterized in that The determining the target ratio according to the probability distribution corresponding to each target production data set in the production data set group includes: A first ratio of the probability distribution of the first production data set to the probability distribution of the second production data set is determined; the target ratio includes the first ratio.

4. The method according to claim 3, characterized in that The determining, according to the probability distribution corresponding to at least one of the target production data sets and the target ratio, the degree of difference between the first production data set and the second production data set comprises: determining a first logarithm of the first ratio, and determining a first product of the first logarithm and a probability distribution of the first production data set; The first product is integrated within a production data interval corresponding to the target production data set to obtain a first integration result, and the first integration result is used as the difference degree.

5. The method according to claim 1, characterized in that: The determining the target ratio according to the probability distribution corresponding to each target production data set in the production data set group includes: Determining a target probability distribution according to the probability distribution corresponding to each of the target production data sets; A second ratio of the probability distribution of the first production data set to the target probability distribution and a third ratio of the probability distribution of the second production data set to the target probability distribution are determined; the target ratio includes the second ratio and the third ratio.

6. The method according to claim 5, characterized in that The determining, according to the probability distribution corresponding to at least one of the production data sets and the target ratio, the degree of difference between the first production data set and the second production data set comprises: determining a second logarithm of the second ratio, and determining a second product of the second logarithm and the probability distribution of the first production data set; determining a third logarithm of the third ratio, and determining a third product of the third logarithm and the probability distribution of the second production data set; Integrating the second product within the production data interval corresponding to the target production data set to obtain a second integral result, and integrating the third product within the production data interval to obtain a third integral result; The degree of difference is determined according to the second integration result and the third integration result.

7. The method according to claim 1, characterized in that The method further comprises: Acquiring each initial production data set of the battery; If the initial production data in each of the initial production data sets are inconsistent, determining a union of the initial production data in each of the initial production data sets; The initial production data set to be updated is updated according to the union to obtain an updated production data set corresponding to the initial production data set to be updated; the initial production data set to be updated includes an initial production data set whose initial production data is inconsistent with the union; Each of the production data set groups is determined according to each of the target production data sets; each of the target production data sets at least includes the updated production data set.

8. The method according to claim 7, characterized in that The obtaining of the frequency sets corresponding to the two target production data sets in the production data set group of the battery includes: If the two target production data sets include the updated production data set, a preset value is filled into a target position in the initial frequency set corresponding to the updated production data set to obtain a frequency set corresponding to the updated production data set; the target position is a position in the initial frequency set corresponding to the newly added production data in the updated production data set; If the two target production data sets include the initial production data set, the initial frequency set corresponding to the initial production data set is used as the frequency set corresponding to the initial production data set.

9. The method according to claim 1, characterized in that: The preset threshold includes a first preset threshold. If the production data set group includes a target production data set in the time dimension; determining the abnormality detection result of each production data set group according to the difference degree corresponding to each production data set group and the preset threshold corresponding to the production data set group, includes: Determining the median of each degree of difference according to the degree of difference corresponding to each of the production data set groups; If the median is greater than the first preset threshold, determining that the abnormality detection result is that the change trend of the production data of the production line over time is abnormal; If the median is not greater than the first preset threshold, it is determined that the abnormal detection result is that the change trend of the production data of the production line over time is normal.

10. The method according to claim 1, characterized in that The preset threshold includes a second preset threshold, if the production data set group includes a target production data set in a spatial dimension; determining the abnormality detection result of each production data set group according to the difference degree corresponding to each production data set group and the preset threshold corresponding to the production data set group, includes: Determining the maximum degree of difference from the degrees of difference corresponding to each of the production data set groups; If the maximum difference is greater than the second preset threshold, it is determined that the production data of the battery is inconsistent in spatial distribution; If the maximum difference is not greater than the second preset threshold, it is determined that the production data of the battery are consistent in spatial distribution.

11. The method according to any one of claims 1 to 6, characterized in that: The production data in the target production data set includes at least one of a liquid injection volume, a helium leak detection rate, a battery self-discharge rate, a battery thickness and a battery internal resistance.

12. A battery production data difference detection device, characterized in that: The device comprises: A first acquisition module is used to obtain the probability distribution corresponding to two target production data sets in a production data set group of a battery; the two target production data sets include a first production data set and a second production data set; if the production data set group includes a target production data set in a time dimension, then the multiple production data set groups are obtained by combining the target production data set corresponding to the maximum timestamp among multiple different timestamps on the same production line with the target production data set corresponding to any other timestamp; if the production data set group includes a target production data set in a spatial dimension, then the multiple production data set groups are obtained by combining any two target production data sets on different production lines corresponding to the same time as one production data set group, or by combining any two target production data sets on different machines on the same production line corresponding to the same time as one production data set group; A first determination module, configured to determine a target ratio according to a probability distribution corresponding to each of the target production data sets in the production data set group; a second determination module, configured to determine a degree of difference between the first production data set and the second production data set according to a probability distribution corresponding to at least one of the target production data sets and the target ratio; The third determination module is used to determine the abnormality detection result of each of the production data set groups according to the difference degree corresponding to each of the production data set groups and the preset threshold value corresponding to the production data set group.

13. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 11 are implemented.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 11 are implemented.

15. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 11 are implemented.

Images