A method, system, product and storage medium for obtaining electrical performance of a lithium ion battery

Abstract

The application provides a lithium ion battery electrical performance acquisition method, system, product and storage medium, and relates to the technical field of battery data processing. Microscopic data is labeled according to conventional data at the same time to generate labeled data, the labeled data including the microscopic data and a corresponding label file, wherein the label file contains a target category and a bounding box coordinate, the target category being determined by a certain data in the conventional data falling into a corresponding sub-range, the continuous conventional data is discretized, so that the change of the microscopic structure can be associated with a specific performance index range, the change of a specific region inside the battery can be directly associated with an external performance index, the model can learn the relationship between the performance index and the microscopic structure, and the internal change of a single battery sample can be continuously monitored in the entire aging cycle without interrupting the test process or using multiple samples. Not only the demand for multiple battery samples is reduced, but also the time and resource consumption is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of battery data processing, and particularly relates to a lithium ion battery electrical performance acquisition method, system, product and storage medium. BACKGROUND

[0002] In the battery industry, aging test is a crucial step. The purpose of this process is to monitor the performance degradation of batteries after long-term use. Small changes on the surface of the battery, such as changes in color depth, slight swelling or other subtle changes, are important observation indicators. Through the use of precise imaging and analysis techniques, researchers can identify and record these small but critical changes. These data provide an important basis for optimizing and improving battery design. Battery design teams can continuously adjust and optimize the chemical composition and structural design of the battery based on these observations to extend the service life of the battery and improve its overall performance.

[0003] However, in the related art, to obtain the internal structure information of the battery, it is usually necessary to disassemble the battery. But for the aging test team, once the battery is disassembled, it cannot continue to be tested for aging. This means that each time the internal structure changes need to be understood, a new battery sample needs to be used, which not only increases the cost, but also significantly prolongs the entire test cycle. This situation leads to a dilemma: researchers need long-term aging data to evaluate battery performance, and also need to understand the changes in the internal structure at different stages. This requires simultaneous testing on multiple identical battery samples, with one part used for the complete aging cycle and the other part disassembled and analyzed at different time points. Although this method is feasible, it greatly increases the consumption of time and resources. SUMMARY

[0004] The present application provides a lithium ion battery electrical performance acquisition method, system, product and storage medium, which reduces the need for multiple battery samples and reduces the consumption of time and resources.

[0005] In a first aspect, a lithium ion battery electrical performance acquisition method includes:

[0006] The conventional data of the target battery at different times is acquired, the conventional data including discharge period, charge-discharge rate, temperature change, voltage and current waveform; the microscopic data of the target battery at different times is acquired, the microscopic data including microscopic structure image and crystal structure data of the target battery; the microscopic data is labeled according to the conventional data at the same time, and labeled data is generated; wherein the labeled data includes the microscopic data and a corresponding label file, the label file including a target category and a boundary box coordinate corresponding to the target category, the target category being a certain data in the conventional data falling into a corresponding sub-range, and the boundary box coordinate being determined by the microscopic data; the labeled data at different times is combined to form a data set; the data set is input into a preset battery performance prediction model for training; and the current conventional data is input into the battery performance prediction model to obtain corresponding current microscopic data.

[0007] By adopting the above technical solution, the time correlation between the external performance and the internal structure of the battery is established by acquiring the conventional data and the microscopic data of the target battery at different times, the microscopic data is labeled according to the conventional data at the same time, and the labeled data is generated, the labeled data including the microscopic data and a corresponding label file, wherein the label file includes a target category and a boundary box coordinate, the target category being determined by a certain data in the conventional data falling into a corresponding sub-range, the continuous conventional data is discretized, the change of the microscopic structure can be associated with a specific performance index range, and the change of a specific region in the battery can be directly associated with an external performance index. The labeled data at different times is combined to form a data set, and is used for training a battery performance prediction model, so that the model can learn the relationship between the performance index and the microscopic structure, researchers can continuously monitor the internal changes of a single battery sample during the entire aging period without interrupting the test process or using multiple samples. Not only the demand for multiple battery samples is reduced, but also the time and resource consumption is reduced.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of determining the boundary box coordinate from the microscopic data specifically includes: performing contour detection on the microscopic data to obtain a group of contours; performing screening on the contours according to a preset area threshold and shape feature to retain contours meeting the conditions; and calculating a minimum circumscribed rectangle of each retained contour, and taking the coordinates of the minimum circumscribed rectangle as the boundary box coordinate.

[0009] By adopting the above technical solution, the contours are obtained by performing contour detection on the microscopic data, the outstanding structures in the microscopic data can be automatically identified, the contours are screened according to the preset area threshold and shape feature, the noise and irrelevant structures are effectively excluded, the minimum circumscribed rectangle is calculated as the boundary box coordinate, and the accuracy of labeling is improved. This automatic boundary box determination method significantly reduces human intervention, improves the objectivity and consistency of labeling, and reduces errors that may be caused by manual labeling.

[0010] In some embodiments in combination with the first aspect, before the step of performing contour detection on the microstructure data to obtain a set of contours, the method further comprises: preprocessing the microstructure data to convert the microstructure data into a grayscale image; performing edge detection on the grayscale image to obtain an edge map; performing dilation and erosion operations on the edge map to obtain a processed image; performing binaryzation on the processed image; and applying connected component analysis on the binaryzated processed image to remove regions smaller than a predetermined area threshold.

[0011] By using the above technical solution, the microstructure data is converted into a grayscale image, which facilitates subsequent edge detection. The edge detection highlights the contours of the microstructure, and the dilation and erosion operations further enhance the clarity of these contours. The binaryzation converts the image into black and white, effectively separating the target structure from the background. Finally, the connected component analysis removes regions smaller than a predetermined area threshold, effectively filtering out noise and small irrelevant structures.

[0012] In some embodiments in combination with the first aspect, the step of obtaining microstructure data of the target battery at different times specifically comprises: obtaining a capacity decay curve of the same battery, wherein the same battery is a same type of battery as the target battery; obtaining feature points in the capacity decay curve; and obtaining microstructure data of the target battery at times corresponding to the feature points.

[0013] By using the above technical solution, the capacity decay curve of the same type of battery is obtained, providing a reference benchmark for the performance changes of the target battery. By analyzing the feature points in the capacity decay curve, key time nodes are determined, which usually represent moments when the battery performance changes significantly. Obtaining microstructure data of the target battery at times corresponding to these feature points ensures that the time points for microstructure analysis are representative and important. Therefore, the number of times of obtaining microstructure data is reduced, saving both time and the workload of professional technicians.

[0014] In some embodiments in combination with the first aspect, the step of obtaining feature points in the capacity decay curve specifically comprises: taking points on the capacity decay curve at a predetermined interval and calculating the curvature of each taken point; selecting points with curvature greater than a predetermined high curvature extreme value and points with curvature less than a predetermined low curvature extreme value as feature points; calculating the curvature change rate of each taken point according to the curvature; selecting taken points with curvature change rate higher than a predetermined curvature mutation threshold as feature points; dividing the capacity decay curve into multiple paragraphs; determining the number of taken points for each paragraph according to the length of the paragraph; and randomly taking feature points on each paragraph according to the number of taken points.

[0015] By adopting the technical scheme, the key change points are ensured to be captured by setting the curvature high-low extreme value and selecting the points with the most significant curve change as the feature points. The mutation points in the curve are further identified by calculating the curvature change rate and setting the mutation threshold, and these points often represent important turning points of the battery performance. The curve is segmented and random points are taken in each segment, which ensures the uniform distribution of the feature points on the entire curve. This multi-level and multi-angle feature point selection method not only captures the key changes of the curve, but also takes into account the overall trend, greatly improving the representativeness and comprehensiveness of the feature points.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after the step of obtaining the capacity attenuation curve of the same battery, the method further comprises: judging whether the capacity attenuation curve is continuous; if the capacity attenuation curve is not continuous, using an interpolation technique to supplement the missing data at the missing place of the capacity attenuation curve; and if the capacity attenuation curve is continuous, performing the step of obtaining the feature points in the capacity attenuation curve.

[0017] By adopting the technical scheme, the continuity of the capacity attenuation curve is judged, which can identify the possible missing or abnormal data in the data. For the discontinuous curve, the interpolation technique is used to supplement the missing data, ensuring the integrity and continuity of the curve.

[0018] In combination with some embodiments of the first aspect, in some embodiments, the step of labeling the micro data according to the conventional data at the same time to generate the labeled data specifically comprises: regarding the time within the same range as the same time, and determining the same range according to the capacity attenuation degree.

[0019] By adopting the technical scheme, the time within the same range is regarded as the same time, which solves the problem that the conventional data and the micro data do not strictly correspond in time, and also solves the problem that the conventional data and the micro data are not equal in quantity.

[0020] In the second aspect, the present application provides a lithium ion battery electrical performance acquisition system, which comprises one or more processors and a memory; the memory is coupled with the one or more processors, and is used to store computer program codes, the computer program codes comprising computer instructions; and the one or more processors invoke the computer instructions to enable the lithium ion battery electrical performance acquisition system to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0021] In the third aspect, the present application provides a computer program product comprising instructions, which, when the computer program product is run on the lithium ion battery electrical performance acquisition system, enable the lithium ion battery electrical performance acquisition system to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0022] In a fourth aspect, the present application provides a computer-readable storage medium comprising instructions that, when executed on a lithium-ion battery electrical performance acquisition system, cause the lithium-ion battery electrical performance acquisition system to perform the method as described in the first aspect and any possible implementation of the first aspect.

[0023] The one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0024] 1. By obtaining the regular data and micro data of the target battery at different times, the time correlation between the external performance and internal structure of the battery is established. The micro data is labeled according to the regular data at the same time, and the labeled data is generated, which includes the micro data and the corresponding label file. The target category and the bounding box coordinates are included in the label file, and the target category is determined by the fact that a certain data in the regular data falls into the corresponding sub-range. The continuous regular data is discretized, so that the change of the microstructure can be associated with a specific performance index range, and the change of a specific region in the battery can be directly associated with the external performance index. The labeled data at different times is combined into a data set, which is used to train the battery performance prediction model, so that the model can learn the relationship between the performance index and the microstructure, and researchers can continuously monitor the internal changes of a single battery sample throughout the aging cycle without interrupting the test process or using multiple samples. Not only does it reduce the need for multiple battery samples, but it also reduces time and resource consumption.

[0025] 2. The micro data is subjected to contour detection to obtain a group of contours, which can automatically identify prominent structures in the micro data. The contours are filtered through a pre-set area threshold and shape feature, effectively excluding noise and irrelevant structures. The minimum circumscribed rectangle is calculated as the bounding box coordinates, improving the accuracy of labeling. This automatic bounding box determination method significantly reduces human intervention, improves the objectivity and consistency of labeling, and reduces errors that may be caused by manual labeling.

[0026] 3. The capacity decay curve of the same type of battery is obtained, providing a reference benchmark for the performance change of the target battery. By analyzing the feature points in the capacity decay curve, key time nodes are determined, which usually represent the time when the battery performance changes significantly. The micro data of the target battery is obtained at the time corresponding to these feature points, ensuring that the time points for microanalysis are representative and important. Therefore, the number of times of obtaining micro data is reduced, saving time and reducing the workload of professional technicians. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 is a flowchart of the lithium-ion battery electrical performance acquisition method in the embodiments of the present application;

[0028] Figure 2 is Figure 1 a flowchart of step S103 in

[0029] Figure 3 is Figure 1 a flowchart of step S102 in

[0030] Figure 4 is Figure 3 a flowchart of step S302 in

[0031] Figure 5 is an exemplary hardware structure diagram of the lithium ion battery electrical performance acquisition system in the embodiments of the present application. DETAILED DESCRIPTION

[0032] The terms used in the following embodiments of the present application are only for the purpose of describing the specific embodiments and are not intended to be limiting of the present application. As used in the specification and the appended claims of the present application, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0033] Hereinafter, the terms "first" and "second" are used only for the purpose of description and should not be understood as implying or suggesting relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, the meaning of "a plurality of" is two or more, unless otherwise specified.

[0034] Please refer to Figure 1 , Figure 1 is a flowchart of the lithium ion battery electrical performance acquisition method in the embodiments of the present application;

[0035] In some embodiments, a lithium ion battery electrical performance acquisition method comprises:

[0036] S101, acquiring conventional data of a target battery at different times, the conventional data including discharge period, charge-discharge rate, temperature change, voltage and current waveform;

[0037] The target battery refers to a specific lithium-ion battery that needs to be analyzed and predicted in performance, the conventional data represent external performance indicators that can be directly measured or observed during the use of the battery, the discharge period refers to the time required for the battery to complete a complete discharge, the charge and discharge rate is used to represent the speed of battery charging and discharging. The temperature change refers to the temperature fluctuation of the battery during the working process, and the voltage and current waveform refers to the voltage and current value curve changing with time.

[0038] In some embodiments, one or more target batteries are selected, and the performance data of these batteries are recorded at different time points (possibly spanning several days, weeks or months). The duration of each charge and discharge cycle, the current size during charging and discharging, the change of the surface temperature of the battery, and the real-time change curve of the voltage and current during the whole process are recorded, and the data are sent to the terminal.

[0039] It should be noted that in this embodiment, only the acquisition of conventional data is involved, and the specific collection method or source of conventional data is not involved.

[0040] S102, micro data of the target battery at different times are acquired, and the micro data include microstructure images and crystal structure data of the target battery;

[0041] The micro data refers to data reflecting the internal structure and material characteristics of the battery, the microstructure image represents visual information of the morphology, distribution and structural characteristics of the internal materials of the battery, and the crystal structure data refers to data describing the arrangement of atoms or molecules of the battery materials, including lattice parameters, atomic positions and other information.

[0042] In some embodiments, the target battery is micro-analyzed at a predetermined time point. This may involve disassembling the battery or using non-invasive analysis techniques. The microstructure image is usually obtained by an electron microscope (such as a scanning electron microscope SEM or a transmission electron microscope TEM), which can show the morphology, particle size, pore distribution and other characteristics of the electrode material. The crystal structure data is usually obtained by X-ray diffraction (XRD) and other techniques, which can reflect the crystal structure changes of the material, such as phase transition, lattice expansion or contraction, and the like, and the data are sent to the terminal.

[0043] It should be noted that in this embodiment, only the acquisition of micro data is involved, and the specific collection method or source of micro data is not involved.

[0044] S103, the micro data are labeled according to the conventional data at the same time to generate labeled data; wherein the labeled data include the micro data and a corresponding label file, the label file includes a target category and a boundary box coordinate corresponding to the target category, the target category is a kind of data in the conventional data falling into a corresponding sub-range, and the boundary box coordinate is determined by the micro data.

[0045] In some embodiments, the indicators in the regular data, such as the capacity attenuation rate, are divided into several sub-ranges, each corresponding to a target category, for each time point of micro data, according to the regular data at the same time point, determine the target category to which it belongs, identify the key areas on the microstructure image, and record the boundary box coordinates of these areas, each micro data is assigned a target category label and the corresponding boundary box information, forming a complete labeled data.

[0046] In some specific embodiments, the total range of the indicators in the regular data is taken, and then the total range is equally divided into a corresponding number of sub-ranges according to the number of categories to be divided, which is not limited here.

[0047] In some specific embodiments, all capacity attenuation rate data is sorted, and according to the required number of categories, the corresponding quantile points are calculated, and these quantile points are used as the boundaries of the sub-ranges, which are not limited here.

[0048] In some specific embodiments, researchers themselves determine the sub-ranges according to their own experience, which is not limited here.

[0049] In some specific embodiments, researchers observe the image, identify the key areas that need to be labeled, such as specific crystal structures, defects, or micro features of interest, and use the drawing tools provided by the software (usually rectangular tools) to manually draw boundary boxes on the image to enclose the key areas. Assign a label to each boundary box to describe the characteristics or type of the area, and automatically record the coordinate information of each boundary box, which usually includes the x, y coordinates of the upper left corner and the lower right corner, which is not limited here.

[0050] S104, group the labeled data at different times into a data set;

[0051] S105, input the data set into a pre-set battery performance prediction model for training;

[0052] In some embodiments, the data set generated in the previous step is input as input into the pre-designed battery performance prediction model. The model will learn from these data and gradually adjust its internal parameters to minimize the error between the prediction results and the actual labeling. During the training process, the model will learn the complex relationship between the battery regular data and the micro data, so as to be able to predict the corresponding microstructure changes under the given regular data.

[0053] S106, input the current regular data to the battery performance prediction model to obtain the corresponding current micro data.

[0054] As can be seen, by acquiring conventional and microscopic data of the target battery at different times, a temporal correlation between the battery's external performance and internal structure was established. Microscopic data was annotated based on conventional data from the same time period, generating labeled data. The labeled data includes microscopic data and corresponding annotation files, where the annotation files contain target categories and bounding box coordinates. The target category is determined by a specific data type from the conventional data falling into a corresponding sub-range. Discretizing continuous conventional data allows changes in microstructure to be correlated with specific performance index ranges, and changes in specific regions within the battery to be directly correlated with external performance indicators. The labeled data from different times were compiled into a dataset and used to train a battery performance prediction model. This model learns the relationship between performance indicators and microstructure, enabling researchers to continuously monitor the internal changes of a single battery sample throughout the entire aging cycle without interrupting the testing process or using multiple samples. This not only reduces the need for multiple battery samples but also lowers time and resource consumption.

[0055] As mentioned in step S103 above, researchers manually draw bounding boxes on the image to enclose key areas based on experience and tools. However, in actual use, manually annotating microscopic data can be subjective, and different researchers may have different annotation results for the same image. The annotation process is time-consuming and prone to errors, especially when dealing with large amounts of data.

[0056] Please see Figure 2 , Figure 2 yes Figure 1 A flowchart of step S103.

[0057] Therefore, in some embodiments, the determination of the bounding box coordinates by microscopic data in step S103 specifically includes:

[0058] S201. Preprocess the microscopic data and convert it into a grayscale image;

[0059] In some embodiments, the original microscopic image data, which is typically in color image format, is read and converted into grayscale images using a specific algorithm. The converted grayscale images retain the structural information of the original images but reduce the amount of data, which is beneficial for subsequent processing.

[0060] In some specific embodiments, a weighted average method is used, assigning different weights to red, green, and blue colors based on the different sensitivities of the human eye to each color, to calculate the grayscale value. Optionally, the average value of the three RGB channels can be directly taken as the grayscale value; alternatively, only the value of a single color channel can be retained as the grayscale value. It is understood that other color space conversion methods can also be used to generate grayscale images, and this is not limited here.

[0061] S202, edge detection is performed on the grayscale image to obtain an edge map;

[0062] In some embodiments, the grayscale image data is read and an edge detection algorithm is applied, which typically identifies regions where pixel values change sharply by calculating the gradient of the image, which are likely to correspond to the edges or contours of objects. The result is usually a binary image where edge pixels are marked as white (or high intensity value) and non-edge pixels are marked as black (or low intensity value).

[0063] S203, an expansion operation and a corrosion operation are performed on the edge map to obtain a processed map;

[0064] In some embodiments, an expansion operation is performed on the edge map, which thickens the edges and fills small gaps that may exist in the edges, and then a corrosion operation is performed on the expanded image, which thins the edges and removes small structures that may be caused by noise, which can smooth the object contours, fill small holes, and eliminate narrow gaps.

[0065] S204, a binaryzation process is performed on the processed map;

[0066] In some embodiments, a threshold value is set and then each pixel in the processed map is traversed. For each pixel, if its grayscale value is greater than or equal to the threshold value, it is set to white (255); if it is less than the threshold value, it is set to black (0). This process simplifies the grayscale information in the image to only contain foreground (white) and background (black) information, which helps subsequent contour detection and analysis.

[0067] S205, a connected region analysis is applied to the binaryzation processed map to remove regions smaller than a predetermined area threshold.

[0068] In some embodiments, a connected region analysis is performed on the binary image to identify and label all connected white (or black) pixel regions. Then, the area (number of pixels) of each connected region is calculated. The area of each region is compared with a predetermined threshold, and if it is smaller than the threshold, all pixels in that region are set to the background color (usually black). This process can effectively remove interference caused by noise or small structures that are not important, so that only the main target structure is retained in the image.

[0069] It can be seen that converting the microscopic data into a grayscale image facilitates subsequent edge detection, which highlights the contours of the microscopic structures, and the expansion and corrosion operations further enhance the clarity of these contours. Binaryzation converts the image to black and white, effectively separating the target structure from the background. Finally, connected region analysis removes regions smaller than a predetermined area threshold, effectively filtering out noise and small irrelevant structures.

[0070] S206, contour detection is performed on the micro data to obtain a group of contours;

[0071] The micro data refers to the micro data processed through S201 to S205.

[0072] In some embodiments, the processed binary image is analyzed using a contour detection algorithm. The algorithm traverses the image, identifying the boundaries between the foreground (usually white) and the background (usually black). Each closed boundary is identified as a contour and represented by a series of coordinates of points.

[0073] In some specific embodiments, an edge tracking algorithm is used to start from an edge point and gradually track along the edge until it returns to the starting point, forming a complete contour, which is not limited here.

[0074] S207, screening the contours, retaining the contours meeting the preset area threshold and shape feature;

[0075] In some embodiments, the area of each contour is calculated and compared with the preset area threshold to remove contours that are too small or too large. Then, the shape features such as circularity, rectangularity, aspect ratio, etc. are calculated for the remaining contours, and further screening is performed according to the preset shape feature standard, which can effectively remove contours that do not meet the expected requirements, such as noise, background structure or incomplete targets, thereby retaining contours that are most likely to represent real target structures.

[0076] S208, calculating the minimum circumscribed rectangle of each retained contour, and taking the coordinates of the minimum circumscribed rectangle as the bounding box coordinates.

[0077] In some embodiments, the minimum circumscribed rectangle of each retained contour is calculated, which may be inclined, so further processing is needed to obtain a rectangle parallel to the image coordinate axis. Then, the top-left corner and bottom-right corner coordinates of this rectangle parallel to the coordinate axis are extracted, which constitute the bounding box coordinates. This process simplifies irregular contours into regular rectangular bounding boxes, facilitating subsequent analysis and processing.

[0078] It can be seen that contour detection is performed on the micro data to obtain a group of contours, which can automatically identify prominent structures in the micro data, and the contours are screened by preset area threshold and shape feature, effectively excluding noise and irrelevant structures, and the minimum circumscribed rectangle is calculated as the bounding box coordinates, improving the accuracy of labeling. This automatic bounding box determination method significantly reduces human intervention, improves the objectivity and consistency of labeling, and reduces errors that may be caused by manual labeling.

[0079] In actual use, the boundary box coordinate determination method mentioned in step S103 and the boundary box coordinate determination method mentioned in steps S201 to S208 can be combined to obtain more accurate and reliable results. A simple method is to further determine the boundary box coordinates obtained by steps S201 to S208 using the boundary box coordinate determination method mentioned in step S103.

[0080] In actual use, obtaining microstructure images and crystal structure data often requires expensive equipment and professional techniques, which may increase research costs, and frequent microanalysis may prolong the research period and affect experimental efficiency.

[0081] Please refer to Figure 3 , Figure 3 is Figure 1 a flowchart of step S102 in

[0082] Therefore, step S102 specifically includes:

[0083] S301, obtaining a capacity attenuation curve of the same battery, wherein the same battery is a battery of the same type as the target battery;

[0084] The capacity attenuation curve represents the graph of the change of the battery capacity with time or the number of charge and discharge cycles. The same battery refers to a battery with the same model, specification and manufacturing process as the target battery.

[0085] In some embodiments, researchers may select a batch of batteries of the same model and specification as the target battery, and conduct long-term charge and discharge cycle tests on them, and record the battery capacity after each cycle. Through these data, a curve of the change of the battery capacity with time or the number of cycles, i.e. the capacity attenuation curve, is drawn. This curve can reflect the overall trend and characteristics of the change of the battery capacity during use.

[0086] S302, obtaining feature points in the capacity attenuation curve;

[0087] In some embodiments, researchers may carefully analyze the capacity attenuation curve obtained in S301 to find important nodes that can represent the change of battery performance. The selection of these feature points may be based on mathematical characteristics (such as the change of the first or second derivative of the curve), or may be based on important thresholds in practical applications (such as the capacity falling to 80% of the rated capacity).

[0088] S303, obtaining microstructure data of the target battery at the time corresponding to the feature points.

[0089] It can be seen that the capacity attenuation curve of the same type of battery provides a reference benchmark for the performance change of the target battery. By analyzing the characteristic points in the capacity attenuation curve, the key time nodes are determined, which usually represent the time when the battery performance changes significantly. The micro data of the target battery is obtained at the time corresponding to these characteristic points, ensuring that the time points of micro analysis are representative and important. Therefore, the number of micro data acquisition is reduced, which not only saves the use time, but also reduces the work burden of professional and technical personnel.

[0090] In actual use, which of the characteristic points in the capacity attenuation curve best represents the important node of the battery performance change?

[0091] Please refer to Figure 4 , Figure 4 is Figure 3 a flowchart of step S302 in

[0092] Therefore, in some embodiments, step S302 specifically comprises:

[0093] S401, taking points on the capacity attenuation curve at a preset interval, and calculating the curvature of each taken point;

[0094] In some embodiments, a suitable point interval is first determined, which should be small enough to capture the detailed changes of the curve, but not too many data points. Then, starting from the beginning of the curve, points are taken on the entire curve at this fixed interval. For each taken point, the curvature of the point is calculated using mathematical methods. The calculation of curvature usually involves the first and second derivatives of the curve at that point.

[0095] However, in actual use, the capacity attenuation curve is obtained by measuring a finite number of sampling points. Since continuous and uninterrupted measurement of the battery is not possible, sampling can only be performed at specific time points. These discrete sampling points are connected by interpolation or fitting methods to draw a continuous curve. Therefore, most of the points on the curve are actually calculated, rather than directly measured. This may result in some points on the curve not accurately reflecting the actual capacity attenuation.

[0096] Therefore, in some embodiments, the sampling points used to draw the capacity attenuation curve can be directly selected as data points for subsequent processing to complete the next operation. By using actual measured data rather than interpolated points on the curve, the real situation of battery capacity attenuation can be more accurately reflected.

[0097] S402, selecting points with curvature greater than a preset high curvature extreme value and points with curvature less than a preset low curvature extreme value as characteristic points;

[0098] The preset high curvature extreme value refers to a large curvature threshold value preset for identifying points of sharp change in the curve. The preset low curvature extreme value refers to a small curvature threshold value preset for identifying points of slow change in the curve.

[0099] In some embodiments, a suitable high curvature extreme value and a low curvature extreme value are predetermined. Then, the curvature values of all the points calculated in S401 are traversed, and points whose curvatures exceed the high extreme value or are lower than the low extreme value are marked as feature points.

[0100] S403, calculating the curvature change rate of each selected point according to the curvature;

[0101] In some embodiments, the curvature change rate is calculated by calculating the difference in curvature between adjacent points and dividing by the time or position interval between the points. The calculation of the curvature change rate can reveal where the curve has undergone sharp changes, which may correspond to important turning points in the performance of the battery.

[0102] S404, selecting a selected point whose curvature change rate is higher than a preset curvature mutation threshold value as a feature point;

[0103] In some embodiments, a suitable curvature mutation threshold value is determined, and then the curvature change rates of all the points calculated in S403 are traversed, and points whose change rates exceed the preset threshold value are marked as feature points. These feature points generally represent key turning points in the capacity decay process and may correspond to significant changes in the internal structure or performance of the battery.

[0104] S405, dividing the capacity decay curve into multiple segments;

[0105] In some embodiments, the entire curve is divided into several segments according to the feature points identified previously or according to predefined time intervals or capacity change amplitudes. Each segment may represent a specific stage in the battery decay process, such as an initial rapid decay period, a stable decay period, an accelerated decay period, etc.

[0106] S406, determining the number of selected points for each segment according to the length of the segment;

[0107] In some embodiments, according to the length of each segment, a certain rule is used to determine how many feature points should be selected in the segment. This method can ensure that more points are selected in longer segments and fewer points are selected in shorter segments, thereby maintaining data representation while avoiding over-sampling.

[0108] S407, randomly selecting feature points on each segment according to the number of selected points.

[0109] It can be seen that by setting the curvature maximum and minimum, the most significant point of the curve is selected as the feature point, ensuring that the key change point is captured. The curvature change rate is calculated and the mutation threshold is set, further identifying the mutation points in the curve, which often represent important turning points in battery performance, segmenting the curve and randomly selecting points in each segment, ensuring the uniform distribution of feature points on the entire curve. This multi-level, multi-angle feature point selection method not only captures the key changes in the curve, but also takes into account the overall trend, greatly improving the representativeness and comprehensiveness of the feature points.

[0110] In actual use, the capacity decay curve may be incomplete. Since subsequent data processing and analysis are mainly based on the points selected from this curve, the incompleteness of the curve will have a significant impact on the entire analysis process.

[0111] In some embodiments, after step S301, the method further comprises:

[0112] S304, determining whether the capacity decay curve is continuous;

[0113] In some embodiments, the time or charge-discharge frequency interval between all adjacent data points is calculated; then a threshold is set, such as twice the normal sampling interval; finally, it is checked whether there is an interval that exceeds this threshold, and if there is, it is determined to be discontinuous, which is not limited here.

[0114] In some embodiments, the capacity data is first differentiated by first order, and the change rate between adjacent points is calculated; then a reasonable range of change rate is set; then it is checked whether there is a point with a change rate exceeding this range, and if there is, it may indicate that the curve is discontinuous, which is not limited here.

[0115] S305, if the capacity decay curve is discontinuous, using interpolation techniques to supplement the missing data at the missing part of the capacity decay curve;

[0116] In some embodiments, the specific location and range of missing data are first determined, and then an appropriate interpolation method is selected. Common interpolation methods include linear interpolation, polynomial interpolation, spline interpolation, etc. The interpolation process will consider the known data points around the missing point, and calculate a reasonable estimated value based on the values and distribution characteristics of these points. The interpolated data should be consistent with the original data and be able to reflect the overall trend of capacity decay.

[0117] In some specific embodiments, the start and end points of the missing data are determined; then the linear interpolation method is used, i.e. assuming the missing part is a straight line; finally, according to the values of the start and end points, the capacity values at each time point of the missing part are calculated in proportion, which is not limited here.

[0118] In some specific embodiments, several known points around the missing point are selected first; then a polynomial interpolation method such as Lagrange interpolation or Newton interpolation is used; and then the capacity value at the missing point is calculated according to the fitted polynomial function, which is not limited here.

[0119] If the capacity attenuation curve is continuous, step S302 is performed.

[0120] It can be seen that judging the continuity of the capacity attenuation curve can identify the possible missing or abnormal data in the data. For the discontinuous curve, the missing data is supplemented using the interpolation technique, ensuring the integrity and continuity of the curve.

[0121] In actual use, the regular data and the micro data should ideally correspond one by one in time and quantity, but there may be cases where the time and quantity do not completely correspond, for example, in the above steps S301 to S303, the sampling frequency and sampling time of the micro data are changed. In theory, one solution is to adjust the regular data accordingly, and another method is as follows:

[0122] The time within the same range is regarded as the same time, and the same range is determined according to the capacity attenuation degree.

[0123] In some embodiments, the capacity attenuation curve is first analyzed to identify time periods with similar capacity attenuation rates; then all time points within these time periods are regarded as the same time to simplify the data processing process; this processing method is based on the assumption that the performance and characteristics of the battery do not change much within a time period with similar capacity attenuation rates, so the data within this time period can be regarded as homogeneous.

[0124] In some specific embodiments, the slope of the capacity attenuation curve is first calculated to obtain the capacity attenuation rate; then a threshold range of the capacity attenuation rate is set; and finally, the time periods that are continuous and have a capacity attenuation rate within the threshold range are divided into the same range, which is not limited here.

[0125] In some specific embodiments, the capacity attenuation curve is first divided into several segments according to time; then the capacity attenuation percentage within each segment is calculated; and then adjacent time periods falling into the same interval are merged into the same range according to the pre-set capacity attenuation percentage interval, which is not limited here.

[0126] In some embodiments, there may be multiple micro data corresponding to one regular data, or vice versa, within each time period. This method effectively solves the quantity correspondence problem by regarding all data points within the same time period as data at the same time. The specific processing can be performed in the following way:

[0127] If there are multiple micro data corresponding to one regular data in a time period, the average of these micro data or other statistical methods (such as median) can be taken to represent the micro data of this time period. If there are multiple regular data corresponding to one micro data in a time period, the regular data can also be averaged or other statistical methods can be used.

[0128] The micro data and the regular data in each time period can also be matched arbitrarily.

[0129] It can be seen that the same range of time is regarded as the same time, which solves the problem that the regular data and the micro data do not strictly correspond in time, and also solves the problem that the regular data and the micro data are not equal in quantity.

[0130] The following introduces an exemplary lithium ion battery electrical performance acquisition system 500 provided by an embodiment of the present application. Figure 5 is an exemplary hardware structure schematic diagram of the lithium ion battery electrical performance acquisition system 500 provided by an embodiment of the present application.

[0131] In some embodiments, the lithium ion battery electrical performance acquisition system 500 is a computer device or includes a computer device in the lithium ion battery electrical performance acquisition system 500. The computer device includes a processor, a memory and a network interface connected through a system bus. 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, a computer program and a database. The internal memory provides an environment for the operating system and the computer program in the non-volatile storage medium to run. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with other terminals or servers outside through network connection. In some embodiments, the network interface can be a wired network interface, and in some embodiments, the network interface can also be a wireless network interface. The computer program is executed by the processor to implement the method in the embodiment of the present application.

[0132] Those skilled in the art can understand that, Figure 5 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0133] The above-described embodiments merely serve to illustrate the technical solutions of the present application, rather than limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those ordinarily skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or equivalently replace some of the technical features thereof; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0134] In the above-described embodiments, the term "when" can be interpreted as meaning "if" or "after" or "in response to determining" or "in response to detecting" according to the context. Similarly, the phrase "upon determining" or "if detecting (the stated condition or event)" can be interpreted as meaning "if determining" or "in response to determining" or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)" according to the context.

[0135] In the above-described embodiments, all or some of the flowcharts or functional modules can be realized in the form of a software program. The software program is stored in a volatile or non-volatile storage medium (for example, a RAM, a floppy disk, a USB flash disk, a ROM, or a DVD), and contains a number of instructions capable of being executed by one or more computer processors. The above-described embodiments can be realized by means of computer programs or pieces of program code that are executed on one or more computers or computer processors. The program code can be stored in any desired manner, for example, as a computer program product, on a data medium or a data storage medium, such as a floppy disk, a USB flash disk, a ROM, or a DVD, or can be transmitted via a data network, for example, the Internet or a local area network, between data processing devices, such as servers or data centers.

[0136] Those of ordinary skill in the art can understand that all or some of the flowcharts in the above-described embodiments can be implemented by a computer program instructing relevant hardware, and the program can be stored in a computer-readable storage medium and can include the flowcharts of the above-described embodiments when executed. The aforementioned storage medium includes a ROM or a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

Claims

1. A method for obtaining the electrical performance of a lithium-ion battery, characterized in that, include: Acquire routine data of the target battery at different times, including discharge cycle, charge / discharge rate, temperature change, voltage and current waveforms; The microscopic data of the target battery at different times are obtained, including microstructure images and crystal structure data of the target battery; The micro data is labeled based on the conventional data from the same time period to generate labeled data; wherein, the labeled data includes the micro data and the corresponding label file, the label file includes the target category and the bounding box coordinates corresponding to the target category, the target category is a certain type of data in the conventional data falling into the corresponding sub-range, and the bounding box coordinates are determined by the micro data; The labeled data from different times are combined into a dataset; The dataset is input into a pre-defined battery performance prediction model for training; Input the current conventional data into the battery performance prediction model to obtain the corresponding current micro data.

2. The method according to claim 1, characterized in that, The step of determining the bounding box coordinates from the microscopic data specifically includes: Contour detection is performed on the microscopic data to obtain a set of contours; The contours are filtered, and contours that meet the conditions are retained according to preset area thresholds and shape features; Calculate the minimum bounding rectangle for each retained contour, and use the coordinates of the minimum bounding rectangle as the bounding box coordinates.

3. The method according to claim 2, characterized in that, Before the step of performing contour detection on the microscopic data to obtain a set of contours, the method further includes: The microscopic data is preprocessed and converted into a grayscale image; Edge detection is performed on the grayscale image to obtain an edge map; The edge map is subjected to dilation and erosion operations to obtain a processed map; The processed image is binarized; Connectivity analysis is applied to the binarized graph to remove regions smaller than a predetermined area threshold.

4. The method according to claim 1, characterized in that, The step of acquiring the microscopic data of the target battery at different times specifically includes: Obtain the capacity decay curve of the same battery, wherein the same battery is a battery of the same type as the target battery; Obtain the feature points in the capacity decay curve; The microscopic data of the target battery are acquired at the time corresponding to the feature point.

5. The method according to claim 4, characterized in that, The step of obtaining the feature points in the capacity decay curve specifically includes: Points are taken on the capacity decay curve at preset intervals, and the curvature of each point is calculated. Points with curvature greater than a preset high curvature extreme value and points with curvature lower than a preset low curvature extreme value are selected as the feature points; Calculate the rate of change of curvature at each selected point based on the curvature; The points whose rate of curvature change is higher than the preset curvature abrupt change threshold are selected as the feature points; The capacity decay curve is divided into multiple segments; The number of points to be selected for each paragraph is determined according to the length of the paragraph. The feature points are randomly selected from each paragraph according to the specified number of points.

6. The method according to claim 4, characterized in that, After the step of obtaining the capacity decay curves of the same batteries, the method further includes: Determine whether the capacity decay curve is continuous; If the capacity decay curve is discontinuous, interpolation techniques are used to supplement the missing data in the capacity decay curve. If the capacity decay curve is continuous, then the step of obtaining the feature points in the capacity decay curve is performed.

7. The method according to claim 4, characterized in that, The step of annotating the micro data based on the conventional data from the same time period to generate annotated data specifically includes: The time within the same range is considered as the same time, and the same range is determined according to the degree of capacity decay.

8. A system for acquiring the electrical performance of a lithium-ion battery, characterized in that, The lithium-ion battery electrical performance acquisition system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the lithium-ion battery electrical performance acquisition system to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on the lithium-ion battery electrical performance acquisition system, the lithium-ion battery electrical performance acquisition system performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the lithium-ion battery electrical performance acquisition system, the lithium-ion battery electrical performance acquisition system performs the method as described in any one of claims 1-7.

Images