Battery cell direct current internal resistance sorting method and computer equipment
By acquiring data from the entire process of cell formation and capacity assessment, extracting multi-dimensional feature sets and constructing feature vectors, and using Euclidean distance to calculate cell similarity, the problem of insufficient accuracy in existing cell DC internal resistance sorting methods is solved, achieving efficient and low-cost cell sorting and improving battery pack performance and lifespan.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SUZHOU QINGTAO NEW ENERGY TECH CO LTD
- Filing Date
- 2026-02-05
- Publication Date
- 2026-04-14
AI Technical Summary
Existing cell DC internal resistance sorting methods are greatly affected by test temperature, pulse conditions and SOC point, and cannot comprehensively collect inconsistent DCR information of cells across the entire operating range. This results in insufficient sorting accuracy, making it difficult to accurately identify and group cells, thus affecting the performance and lifespan of battery modules or battery packs.
By acquiring data from the entire process of cell formation and capacity assessment, extracting multi-dimensional feature sets, constructing feature vectors, and using Euclidean distance to calculate the feature similarity between cells, more accurate DCR sorting is achieved, and the sorting process is automated using computer equipment.
It improves the accuracy and consistency of cell sorting, reduces costs, enhances the precision and efficiency of battery pack assembly, and is suitable for integration into existing production lines.
Smart Images
Figure CN121847488A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of battery technology, and in particular to a method for sorting battery cells by DC internal resistance and a computer device. Background Technology
[0002] With the rapid development of new energy vehicles and energy storage systems, lithium-ion batteries have been widely used as core energy storage devices. In battery manufacturing, cell sorting is a crucial process to ensure the consistency and safety of battery pack performance. The DC internal resistance (DCR) of a cell, as an important parameter reflecting the internal impedance characteristics of the battery, directly affects the battery's power performance, heat dissipation characteristics, and cycle life. Therefore, cell sorting technology based on DC internal resistance has significant engineering application value.
[0003] However, existing cell DC internal resistance sorting methods still have significant technical shortcomings. The DCR value is greatly affected by test temperature, pulse conditions, and SOC point. Existing methods mostly use test conditions at a single operating point, which can only reflect the characteristics of the cell under that specific operating condition. They cannot comprehensively and effectively collect inconsistent information on the DCR of the cell across the entire operating range, resulting in insufficient accuracy of sorting results. It is difficult to accurately identify and group cells with similar internal resistance characteristics. Especially when processing large batches of cells, the sorting accuracy often cannot meet the requirements of high-performance battery pack assembly, thus affecting the overall performance and lifespan of the battery module or battery pack. Summary of the Invention
[0004] Therefore, it is necessary to provide a method for sorting the DC internal resistance of battery cells and a computer device to address the above problems.
[0005] According to a first aspect of the embodiments of this application, a method for sorting the DC internal resistance of battery cells is provided, comprising: Obtain the formation data and capacity assessment data of each cell to be sorted to form the original dataset; Feature extraction is performed on the original dataset to determine the multidimensional feature set corresponding to each cell to be sorted. The multidimensional feature set characterizes the DC internal resistance characteristics of the corresponding cell to be sorted. The multi-dimensional feature set corresponding to each battery cell to be sorted is converted into the feature vector of the corresponding battery cell to be sorted. Calculate the feature similarity between the feature vectors of different cells to be sorted; Based on the feature similarity between the feature vectors of different cells to be sorted, the cells to be sorted are sorted by DC internal resistance level.
[0006] In one embodiment, the formation data includes the temperature, time, voltage, and current of the cell to be sorted during the formation stage; the capacity testing data includes the temperature, time, voltage, and current of the cell to be sorted during the capacity testing stage.
[0007] In one embodiment, the step of extracting features from the original dataset to determine the multidimensional feature set corresponding to each cell to be sorted includes: Based on the original dataset, the first relationship curve between voltage and differential capacity corresponding to the formation stage of the cells to be sorted and the second relationship curve between voltage and differential capacity corresponding to the capacity grading stage are obtained respectively. Identify the peaks and valleys in the first relationship curve and the second relationship curve, respectively; The peaks and valleys in the first and second relationship curves are optimized to obtain the optimized peak and valley parameters. Based on the optimized peak and valley parameters, feature extraction is performed to obtain a multi-dimensional feature set corresponding to the battery cell to be sorted. The multi-dimensional feature set includes the half-peak width, half-peak height, peak value, voltage value, and temperature value of each peak and valley of the battery cell to be sorted during the formation and capacity testing stages.
[0008] In one embodiment, the step of identifying the peaks and valleys in the first relationship curve and the second relationship curve respectively includes: Obtain the slope of the curves at each point on the first relationship curve and the second relationship curve; The critical point before the slope of the curve changes from positive to negative is defined as the peak, and the critical point before the slope of the curve changes from negative to positive is defined as the valley.
[0009] In one embodiment, the step of optimizing the peaks and valleys in the first relationship curve and the second relationship curve respectively to obtain optimized peak shape parameters and valley shape parameters includes: Each identified peak or valley is assumed to follow a normal distribution, and the coordinate data corresponding to the identified peak or valley is used as the initial parameters of the normal distribution. The initial parameters of each peak or valley are optimized using an optimization algorithm to obtain the optimized peak and valley parameters. The optimization algorithm is selected from one of the following: the Cuckoo Algorithm, Gradient Descent, Particle Swarm Optimization, or Genetic Algorithm.
[0010] In one embodiment, the step of converting the multidimensional feature set corresponding to each cell to be sorted into a feature vector of the corresponding cell includes: The feature values in the multi-dimensional feature set corresponding to each cell to be sorted are arranged in a predetermined order to form the feature vector of the corresponding cell to be sorted.
[0011] In one embodiment, the step of calculating the similarity between the feature vectors of different cells to be sorted includes: The Euclidean distance between the feature vectors of different cells to be sorted is calculated based on the Euclidean distance grading model. The Euclidean distance represents the feature similarity between different cells to be sorted.
[0012] In one embodiment, the step of sorting the cells by DC internal resistance level based on the feature similarity between the feature vectors of different cells to be sorted includes: Based on the Euclidean distance between the feature vectors of different cells to be sorted, each cell is classified into its corresponding DC internal resistance level.
[0013] In one embodiment, the Euclidean distance is a standardized Euclidean distance, calculated using the following formula: in, The Euclidean distance between cell A and cell B to be sorted. Let i be the i-th feature value in the feature vector of cell A to be sorted. Let be the i-th eigenvalue in the feature vector of cell B to be sorted, and n be the total dimension of the feature vector. This refers to the mean of the i-th feature value calculated from the training set during the training phase of the Euclidean distance hierarchical model. This is the standard deviation of the i-th feature value calculated from the training set during the training phase of the Euclidean distance hierarchical model.
[0014] According to a second aspect of the present application, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described cell DC internal resistance sorting method.
[0015] The cell DC internal resistance sorting method provided in this application utilizes data from the entire cell formation and capacity testing process, rather than data from a single test point. This allows for a more comprehensive understanding of cell characteristics and effectively overcomes the characterization limitations caused by traditional methods that rely on testing DC internal resistance (DCR) at a single operating point. Furthermore, DCR sorting is performed based on the similarity of characteristics between cells to be sorted, ensuring that cells in the same group exhibit highly consistent DCR characteristics in practical applications, significantly improving sorting accuracy. This sorting method is based on data from existing mandatory testing processes (formation and capacity testing) on the production line, eliminating the need for additional complex testing steps or equipment. It boasts a high degree of automation during online implementation, achieving accurate sorting while also offering advantages of low cost and high efficiency, and is easily integrated into existing production lines. Attached Figure Description
[0016] Figure 1 A flowchart illustrating a method for sorting battery cells based on DC internal resistance according to an embodiment of this application; Figure 2 This is a flowchart of step S300 in a cell DC internal resistance sorting method provided in an embodiment of this application; Figure 3 A flowchart of step S320 in a cell DC internal resistance sorting method provided in an embodiment of this application; Figure 4 This is a flowchart of step S330 in a cell DC internal resistance sorting method provided in an embodiment of this application; Figure 5 This is a schematic diagram of some of the original data from the cell's capacity grading and discharge. Figure 6 This is a schematic diagram of the relationship between voltage and differential capacity in a cell DC internal resistance sorting method provided in an embodiment of this application. Figure 7 This is a schematic diagram of the peak initially identified in the voltage-differential capacity relationship curve in the cell DC internal resistance sorting method provided in an embodiment of this application; Figure 8 This is a schematic diagram illustrating the optimization using a genetic algorithm in a cell DC internal resistance sorting method provided in an embodiment of this application. Figure 9 This is a schematic diagram of the optimized peak-splitting result in the cell DC internal resistance sorting method provided in an embodiment of this application. Detailed Implementation
[0017] To facilitate understanding of this application, a more complete description will be provided below with reference to the accompanying drawings. Preferred embodiments of this application are shown in the drawings. However, this application can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of this application.
[0018] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "joining," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise expressly limited. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.
[0019] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.
[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein in the specification of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.
[0021] As described in the background section, with the rapid development of new energy vehicles and energy storage systems, lithium-ion batteries have been widely used as core energy storage devices. In the battery manufacturing process, cell sorting is a crucial step to ensure the consistency and safety of battery pack performance. The DC internal resistance (DCR) of a cell, as an important parameter reflecting the internal impedance characteristics of the battery, directly affects the battery's power performance, heat dissipation characteristics, and cycle life. Therefore, cell sorting technology based on DC internal resistance has significant engineering application value.
[0022] However, existing cell DC internal resistance sorting methods still have significant technical shortcomings. First, the DCR value is greatly affected by test temperature, pulse conditions, and SOC (state of charge) point. Existing methods mostly use test conditions at a single operating point, which can only reflect the characteristics of the cell under that specific operating condition. They cannot comprehensively and effectively collect inconsistent information on the DCR of the cell across the entire operating range, resulting in insufficient accuracy of sorting results. It is difficult to accurately identify and group cells with similar internal resistance characteristics. Especially when processing large batches of cells, the sorting accuracy often cannot meet the requirements of high-performance battery pack assembly, thus affecting the overall performance and lifespan of the battery module or battery pack.
[0023] To address the aforementioned problems, this application provides a method for sorting battery cells based on their DC internal resistance, a computer device, and a computer-readable storage medium.
[0024] In one embodiment, a method for sorting DC internal resistance of battery cells is provided. This method overcomes the limitations of existing technologies that rely on a single test condition for DC internal resistance sorting. By comprehensively utilizing global data from the cell formation and capacity testing processes, and combining electrochemical principles and big data algorithms, multidimensional features that can reflect the internal state and DCR characteristics of the battery cells are extracted, thereby achieving more accurate and comprehensive DCR sorting of battery cells.
[0025] Reference Figure 1The cell DC internal resistance sorting method provided in this embodiment includes the following steps: Step S100: Obtain the formation data and capacity assessment data of each cell to be sorted to form the original dataset.
[0026] Battery formation refers to the process of charging a newly assembled, electrolyte-filled battery cell for the first time. This initial charge and discharge allows a stable SEI film to form inside the cell, activating the active materials; this is a functional activation process. In this embodiment, the formation data for the cells to be sorted may include the temperature, time, voltage, and current during the formation stage. Battery capacity testing refers to performing a complete charge and discharge test on the formed cells to accurately measure their actual release capacity. In this embodiment, the capacity testing data for the cells to be sorted may include the temperature, time, voltage, and current during the capacity testing stage. Figure 5 This is a schematic diagram of some of the raw data from the capacity testing and discharge of the battery cells. The raw dataset is constructed from the formation and capacity testing data of each cell to be sorted. This raw dataset provides a complete electrochemical information foundation for subsequent feature extraction. In this embodiment, the use of complete formation and capacity testing data, rather than traditional single-point test data, facilitates a more comprehensive and accurate acquisition of cell characteristics.
[0027] Step S300: Extract features from the original dataset to determine the multidimensional feature set corresponding to each cell to be sorted. The multidimensional feature set characterizes the DC internal resistance characteristics of the corresponding cell to be sorted.
[0028] Once the original dataset, namely the complete formation and capacity data of the cells to be sorted, is determined, a feature extraction pipeline can be constructed based on electrochemical management to process the original data and extract multi-dimensional features that can characterize the DCR characteristics of the cells to be sorted. Then, a corresponding multi-dimensional feature set is formed for each cell to be sorted.
[0029] The "multidimensional feature set" refers to the collective term for various categories of feature values extracted from the raw data of a single cell to be sorted, covering its different stages (formation and capacity testing) and different reaction sites (peaks and valleys). "Multidimensional" refers to the diversity of features, including but not limited to various physical or geometric attribute parameters such as half-peak width, half-peak height, peak value, voltage value, and temperature value. For example, its form could be: {"formation_peak1_half-peak width": 0.5, "formation_peak1_voltage": 3.65, ... "capacity testing_valley2_temperature": 32.1, ...}.
[0030] Step S500: Convert the multi-dimensional feature set corresponding to each cell to be sorted into the feature vector of the corresponding cell.
[0031] After determining the multidimensional feature set corresponding to each cell to be sorted, the multidimensional feature set of each cell is converted into a feature vector, which contains all the feature values of each peak and valley in the formation and capacity testing stages of the cell.
[0032] Step S700: Calculate the feature similarity between the feature vectors of different cells to be sorted.
[0033] The similarity of the feature vectors between different cells to be sorted can reflect the similarity of the DC internal resistance between different cells to be sorted. Therefore, in this embodiment, the cells to be sorted are classified according to the similarity of the feature vectors between different cells to be sorted.
[0034] Step S900: Based on the feature similarity between the feature vectors of different cells to be sorted, sort each cell by its DC internal resistance level.
[0035] Based on the feature similarity calculated in step S700, cluster analysis, distance threshold classification, or other classification algorithms can be used to automatically categorize the cells to be sorted into several preset DC internal resistance levels. For example, cells with an Euclidean distance less than a specific threshold can be grouped into the same level to ensure that the cells within the group have highly consistent DC internal resistance characteristics. This step ultimately outputs the DC internal resistance level to which each cell to be sorted belongs, completing the automatic decision-making from data to sorting results.
[0036] The cell DC internal resistance sorting method provided in this application utilizes data from the entire cell formation and capacity testing process, rather than data from a single test point. This allows for a more comprehensive understanding of cell characteristics and effectively overcomes the characterization limitations caused by traditional methods that rely on testing DC internal resistance (DCR) at a single operating point. Furthermore, DCR sorting is performed based on the similarity of characteristics between cells to be sorted, ensuring that cells in the same group exhibit highly consistent DCR characteristics in practical applications, significantly improving sorting accuracy. This sorting method is based on data from existing mandatory testing processes (formation and capacity testing) on the production line, eliminating the need for additional complex testing steps or equipment. It boasts a high degree of automation during online implementation, achieving accurate sorting while also offering advantages of low cost and high efficiency, and is easily integrated into existing production lines.
[0037] Reference Figure 2 In one embodiment, step S300, namely, the step of extracting features from the original dataset to determine the multidimensional feature set corresponding to each cell to be sorted, may further include: Step S310: Based on the original dataset, obtain the first relationship curve of voltage and differential capacity corresponding to the formation stage of the cells to be sorted and the second relationship curve of voltage and differential capacity corresponding to the capacity grading stage.
[0038] Differential capacity (dQ / dV) is the capacity change corresponding to a unit voltage change. It is calculated as follows: First, the capacity (Q) and voltage (V) data in the original dataset are processed. Then, the differential capacity value is calculated by numerical differentiation method (such as the central difference method), and the relationship curve between voltage and differential capacity is constructed.
[0039] Specifically, smoothing filtering algorithms, such as Savitzky-Golay filtering, simple moving average, weighted moving average, exponential moving average, or Gaussian filtering, can be used first to smooth the voltage-time curves corresponding to the formation and capacity division processes, reducing random noise and errors during testing. Then, based on the denoised voltage-time curves, the voltage-differential capacity curves (v-dqdv, refer to...) corresponding to the formation and capacity division processes can be calculated and generated. Figure 6 In this embodiment, the voltage-differential capacity relationship curve corresponding to the cell formation stage to be sorted is defined as the first relationship curve, and the voltage-differential capacity relationship curve corresponding to the cell capacity grading stage to be sorted is defined as the second relationship curve.
[0040] Step S320: Identify the peaks and valleys in the first relationship curve and the second relationship curve, respectively.
[0041] For the first relationship curve and the second relationship curve, the peaks and valleys in the first relationship curve and the second relationship curve can be identified based on the changes in the slope of the curves, respectively. Figure 7 The peaks that were initially identified are shown in the figure.
[0042] Specifically, refer to Figure 3 Step S320 may further include: Step S321: Obtain the slope of each point in the first relationship curve and the second relationship curve; Step S322: Determine the critical point before the curve slope changes from positive to negative as the peak, and determine the critical point before the curve slope changes from negative to positive as the valley.
[0043] That is, firstly, the slope of each point in the first and second relationship curves can be obtained. Then, the changing pattern of the slope can be identified. When the slope changes from positive to negative, the critical point before the change can be identified as the peak of the relationship curve. Similarly, when the slope changes from negative to positive, the critical point before the change can be identified as the valley of the relationship curve. Thus, the peaks and valleys in the first and second relationship curves can be determined, and the coordinate data corresponding to the initially detected peaks and valleys can be recorded, namely, the voltage value (V) and the differential capacitance value (dq / dv).
[0044] Step S330: Optimize the peaks and valleys in the first relationship curve and the second relationship curve respectively to obtain the optimized peak shape parameters and valley shape parameters.
[0045] In this embodiment, the coordinate data of each peak and valley in the initially identified first and second relationship curves can be optimized to make the features extracted from the full data more accurate, thereby making the sorting criteria more correlated with the actual DCR performance of the battery cell.
[0046] Specifically, refer to Figure 4 Step S330 may further include: Step S331: Assume that each identified peak or valley follows a normal distribution, and use the coordinate data corresponding to the identified peak or valley as the initial parameters of the normal distribution; Step S332: Optimize the initial parameters of each peak or valley using an optimization algorithm to obtain the optimized peak shape parameters and valley shape parameters.
[0047] That is, for each initially identified peak or valley, it is assumed that it follows a normal distribution. Theoretical research suggests that real-world data is composed of a series of superimposed normal distributions, with parameters μ and σ. The coordinate data of the initially identified peaks or valleys, i.e., voltage values and differential capacitance values, can be used as the initial values of the normal distribution parameters μ and amplitude, i.e., the initial parameters. The optimization space is set within the range of -10% to 10% of the initial values for each parameter. Any one of the following optimization algorithms can be used (preferably the genetic algorithm). Figure 8 The diagram illustrates the optimization process using a genetic algorithm to find the optimal solution in space. The objective is to minimize the error between the fitted synthetic voltage-differential capacity curve and the original voltage-differential capacity curve. The parameters (μ, σ, amplitude) of each peak and valley are optimized to obtain the optimized peak and valley parameters. Those skilled in the art can select a suitable optimization algorithm based on actual needs. Figure 9This is a schematic diagram of the peak separation results after parameter optimization.
[0048] Here, error refers to the root mean square error (RMSE) of the difference between the optimized synthetic curve and the original curve (the difference between each point in the curve).
[0049] Step S340: Based on the optimized peak and valley parameters, feature extraction is performed to obtain a multi-dimensional feature set corresponding to the cell to be sorted. The multi-dimensional feature set includes the half-peak width, half-peak height, peak value, voltage value, and temperature value of each peak and valley of the cell to be sorted during the formation and capacity testing stages.
[0050] For the optimized peak and valley parameters, feature extraction can be performed. Specifically, the half-peak width (WHM), half-peak height (HHH), peak value, voltage value, and temperature value during the testing process can be extracted for each peak and valley to form a multi-dimensional feature set for each cell to be sorted. The WHM refers to the voltage span corresponding to half the peak value on the optimized curve, reflecting the speed and concentration of the electrochemical reaction kinetics. The HHH refers to the vertical height from the peak (valley) value to the local baseline, directly characterizing the reaction intensity or active material mass. Both are precisely extracted from the fitted curve after parameter optimization, serving as key feature values for quantifying peak and valley morphology and characterizing the internal reaction characteristics of the cell, used to construct the cell's feature vector. The temperature value refers to the real-time temperature corresponding to the moment the peak or valley appears. All extracted features are values obtained through optimized remapping, thereby improving the accuracy of the features and making the correlation between the sorting criteria and the actual DCR characteristics of the cell higher.
[0051] In one embodiment, step S500, namely the step of converting the multidimensional feature set corresponding to each cell to be sorted into the feature vector of the corresponding cell, includes: arranging each feature value in the multidimensional feature set corresponding to each cell to be sorted in a predetermined order to form the feature vector of the corresponding cell.
[0052] After obtaining the corresponding multidimensional features for each cell to be sorted, the feature values corresponding to each feature in the multidimensional feature set of each cell are arranged in a predetermined order to form the feature vector of each cell. In a specific example, the feature vector of cell A to be sorted is [x1, x2, x3, ..., xn], which includes the formation of cell A, the half-peak width, half-peak height, peak value, corresponding voltage value of each peak / valley in the capacity testing stage, and the temperature data during the testing process. These feature values are arranged in a predetermined order to form the feature vector. Similarly, the feature vector of cell B to be sorted is [y1, y2, y3, ..., yn], which includes the formation of cell B to be sorted, the half-peak width, half-peak height, peak value, corresponding voltage value, and temperature data during the testing process, and is arranged in the same predetermined order as cell A. It is understood that the "predetermined order" is a fixed arrangement rule set in advance to ensure the comparability of all cell feature vectors. Its core is to ensure that the same dimension of all cell feature vectors corresponds to the same physical features. The specific order can be flexibly defined according to principles such as stage, feature type or numerical scale, but once determined, it serves as the unified arrangement order for constructing all cell feature vectors.
[0053] For example, assume that the predetermined order rule is defined as follows: first arrange all features of the "formation" stage, then arrange all features of the "capacity testing" stage. Within each of the "formation" and "capacity testing" stages, features are arranged from low to high according to the voltage values corresponding to the feature points (peaks or valleys). For each feature point, its five feature values are arranged in a fixed order as follows: half-peak width, half-peak height, peak value, voltage value, and temperature value.
[0054] According to the above rules, assuming that each cell to be sorted has 3 peaks identified during the formation stage and 2 valleys identified during the capacity testing stage, then the feature vector of each cell to be sorted will be a 25-dimensional vector ((3 peaks + 2 valleys) × 5 features = 25 values). Thus, the feature vectors of each cell to be sorted are aligned in dimension, with the same position representing the exact same physical meaning. This is useful for subsequently calculating the Euclidean distance between the feature vectors of each cell to be sorted.
[0055] Then, the Euclidean distance between the feature vectors of different cells to be sorted can be calculated based on the pre-trained Euclidean distance hierarchical model. The Euclidean distance between the feature vectors of different cells to be sorted can characterize the similarity between feature vectors of the same type.
[0056] In one embodiment, step S700, which calculates the feature similarity between the feature vectors of different cells to be sorted, includes: determining the Euclidean distance between the feature vectors of different cells to be sorted according to an Euclidean distance grading model. The Euclidean distance between the feature vectors of different cells to be sorted characterizes the feature similarity between the different cells. By calculating the Euclidean distance between the feature vectors, the degree of difference in DC internal resistance characteristics among different cells can be quantified.
[0057] In this embodiment, the Euclidean distance is a standardized Euclidean distance, which can be specifically determined by the following formula to determine the Euclidean distance between different cells to be sorted, that is, the feature similarity between different cells to be sorted: in, The Euclidean distance between cell A and cell B to be sorted. Let i be the i-th feature value in the feature vector of cell A to be sorted. Let be the i-th eigenvalue in the feature vector of cell B to be sorted, and n be the total dimension of the feature vector. This refers to the mean of the i-th feature value calculated from the training set during the training phase of the Euclidean distance hierarchical model. This is the standard deviation of the i-th feature value calculated from the training set during the training phase of the Euclidean distance hierarchical model.
[0058] In one embodiment, step S900, which is to sort each cell to be sorted according to the feature similarity between the feature vectors of different cells to be sorted, includes: classifying each cell to be sorted into the corresponding DC internal resistance level according to the value of the Euclidean distance between the feature vectors of different cells to be sorted.
[0059] Based on the Euclidean distance calculated in step S700, which characterizes the differences in characteristics between battery cells, cluster analysis or distance threshold algorithms can be used to automatically classify all battery cells to be sorted into several preset DC internal resistance levels (e.g., 5 levels) according to the accuracy requirements for consistency in actual production. The smaller the Euclidean distance, the more similar the DC internal resistance characteristics of the battery cells, and they will be classified into the same level, thereby ensuring that the battery cells in the same group perform with high consistency in actual applications, and ultimately achieving automated and high-precision decision-making from feature data to sorting results.
[0060] Based on the same inventive concept, in one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the above-described method for sorting the DC internal resistance of battery cells.
[0061] The computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores various types of data related to the DC internal resistance sorting method for battery cells. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a DC internal resistance sorting method for battery cells.
[0062] Based on the same inventive concept, in one embodiment, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the above-described method for sorting the DC internal resistance of battery cells.
[0063] Based on the same inventive concept, in one embodiment, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the above-described cell DC internal resistance sorting method.
[0064] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. 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), magnetic 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 take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.
[0065] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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, they should be considered to be within the scope of this specification.
[0066] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for sorting battery cells based on their DC internal resistance, characterized in that, include: Obtain the formation data and capacity assessment data of each cell to be sorted to form the original dataset; Feature extraction is performed on the original dataset to determine the multidimensional feature set corresponding to each cell to be sorted. The multidimensional feature set characterizes the DC internal resistance characteristics of the corresponding cell to be sorted. The multi-dimensional feature set corresponding to each battery cell to be sorted is converted into the feature vector of the corresponding battery cell to be sorted. Calculate the feature similarity between the feature vectors of different cells to be sorted; Based on the feature similarity between the feature vectors of different cells to be sorted, the cells to be sorted are sorted by DC internal resistance level.
2. The method for sorting battery cells by DC internal resistance according to claim 1, characterized in that, The formation data includes the temperature, time, voltage, and current of the cells to be sorted during the formation stage; the capacity testing data includes the temperature, time, voltage, and current of the cells to be sorted during the capacity testing stage.
3. The cell DC internal resistance sorting method according to claim 2, characterized in that, The step of extracting features from the original dataset to determine the multidimensional feature set corresponding to each battery cell to be sorted includes: Based on the original dataset, the first relationship curve between voltage and differential capacity corresponding to the formation stage of the cells to be sorted and the second relationship curve between voltage and differential capacity corresponding to the capacity grading stage are obtained respectively. Identify the peaks and valleys in the first relationship curve and the second relationship curve, respectively; The peaks and valleys in the first and second relationship curves are optimized to obtain the optimized peak and valley parameters. Based on the optimized peak and valley parameters, feature extraction is performed to obtain a multi-dimensional feature set corresponding to the battery cell to be sorted. The multi-dimensional feature set includes the half-peak width, half-peak height, peak value, voltage value, and temperature value of each peak and valley of the battery cell to be sorted during the formation and capacity testing stages.
4. The cell DC internal resistance sorting method according to claim 3, characterized in that, The steps of identifying the peaks and valleys in the first relationship curve and the second relationship curve respectively include: Obtain the slope of the curves at each point on the first relationship curve and the second relationship curve; The critical point before the slope of the curve changes from positive to negative is defined as the peak, and the critical point before the slope of the curve changes from negative to positive is defined as the valley.
5. The cell DC internal resistance sorting method according to claim 3, characterized in that, The step of optimizing the peaks and valleys in the first and second relationship curves respectively to obtain optimized peak shape parameters and valley shape parameters includes: Each identified peak or valley is assumed to follow a normal distribution, and the coordinate data corresponding to the identified peak or valley is used as the initial parameters of the normal distribution. The initial parameters of each peak or valley are optimized by using an optimization algorithm to obtain the optimized peak shape parameters and valley shape parameters. The optimization algorithm is selected from one of the following: Cuckoo Algorithm, Gradient Descent, Particle Swarm Optimization, and Genetic Algorithm.
6. The method for sorting battery cells by DC internal resistance according to claim 1, characterized in that, The step of converting the multi-dimensional feature set corresponding to each cell to be sorted into the feature vector of the corresponding cell includes: The feature values in the multi-dimensional feature set corresponding to each cell to be sorted are arranged in a predetermined order to form the feature vector of the corresponding cell to be sorted.
7. The method for sorting battery cells by DC internal resistance according to claim 1, characterized in that, The steps for calculating the similarity between the feature vectors of different cells to be sorted include: The Euclidean distance between the feature vectors of different cells to be sorted is calculated based on the Euclidean distance grading model. The Euclidean distance represents the feature similarity between different cells to be sorted.
8. The cell DC internal resistance sorting method according to claim 7, characterized in that, The step of sorting each battery cell by DC internal resistance level based on the feature similarity between the feature vectors of different cells to be sorted includes: Based on the Euclidean distance between the feature vectors of different cells to be sorted, each cell is classified into its corresponding DC internal resistance level.
9. The cell DC internal resistance sorting method according to claim 8, characterized in that, The Euclidean distance used is the standardized Euclidean distance, and the calculation formula is as follows: in, The Euclidean distance between cell A and cell B to be sorted. Let i be the i-th feature value in the feature vector of cell A to be sorted. Let be the i-th eigenvalue in the feature vector of cell B to be sorted, and n be the total dimension of the feature vector. This refers to the mean of the i-th feature value calculated from the training set during the training phase of the Euclidean distance hierarchical model. This is the standard deviation of the i-th feature value calculated from the training set during the training phase of the Euclidean distance hierarchical model.
10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the cell DC internal resistance sorting method according to any one of claims 1-9.