An energy storage unit intelligent sorting method, system, storage medium and device

Abstract

This invention discloses an intelligent sorting method, system, storage medium, and device for energy storage cells, belonging to the field of lithium battery screening technology. It includes defining key parameters and identifying data sources for energy storage cells to determine key parameters for sorting; collecting key process parameter data of energy storage cells through an intelligent data source scheduling algorithm; and filtering the collected key process parameter data using a pre-established multi-parameter sorting model to obtain qualified energy storage cells. This invention achieves dynamic acquisition, intelligent screening, and visual display of cell parameters, and improves screening accuracy and speed through algorithm optimization and model training, thereby meeting the high standards of consistency and process requirements in lithium battery module production.

Description

Technical Field

[0001] This invention relates to an intelligent sorting method, system, storage medium, and device for energy storage units, belonging to the field of lithium battery screening technology. Background Technology

[0002] Lithium-ion batteries, as high-energy-density and long-life energy storage devices, have been widely used in new energy vehicles, energy storage systems, and consumer electronics. During the assembly of lithium battery modules, the performance consistency of individual cells plays a crucial role in the safety, stability, and lifespan of the final product. To meet the "parallel and series" process requirements of modules, manufacturers need to rigorously screen the produced cells to ensure the uniformity and matching of various parameters. However, because lithium battery production involves multiple stages and parameter data comes from diverse sources, traditional screening methods have the following limitations: 1. Complex data sources and inconvenient processing: Existing production systems usually cannot uniformly manage data sources from different production stages. For example, parameters such as open circuit voltage (OCV), DC internal resistance (DCIR), thickness, and cell capacity are generated by different testing equipment or processes. The data storage formats are diverse, making it difficult to integrate and retrieve efficiently.

[0003] 2. Insufficient data real-time performance and accuracy: In the current factory environment, data collected by some equipment may be missing or abnormal, making it difficult to automatically recursively supplement data from other data sources. For high-volume production lines, this hinders rapid response to screening needs, impacting screening efficiency.

[0004] 3. Screening relies on manual labor, resulting in low efficiency: Traditional cell parameter screening often relies on manual statistics and judgment, which is labor-intensive and prone to omissions or errors, affecting module consistency and assembly efficiency.

[0005] 4. Lack of intelligent screening mechanism: In existing technologies, screening criteria and strategies are usually predefined fixed rules that cannot be dynamically adjusted according to different time periods, production processes or incoming material characteristics, making it difficult for screening results to accurately adapt to actual production needs.

[0006] 5. Lack of data reverse tracing and algorithm optimization: The current sorting system lacks a continuous feedback and optimization mechanism for the screening effect, and cannot perform statistical analysis on indicators such as hit rate and missing rate of different data sources, which makes it difficult to continuously improve the adaptability and accuracy of the algorithm model.

[0007] 6. Insufficient visualization and anomaly response: Traditional methods are relatively simple in terms of parameter display and result output, failing to achieve dynamic visualization, labeling, and hierarchical processing of data anomalies. For example, for battery cells exceeding the screening threshold, there is a lack of integrated response mechanisms for automatic early warning, robotic waste removal, or manual handling.

[0008] In the sorting process of lithium battery production, several technologies have been used to screen cells using different parameters to meet the consistency requirements, achieving certain results. However, many limitations still exist, as detailed below: 1. Single-parameter screening, lacking multi-factor comprehensive analysis: The invention patent CN118275895, "A Dynamic Screening Method for Lithium-ion Batteries," proposes using the K value as a single indicator for cell screening, combined with statistical methods (such as quartile method and standard deviation calculation) to dynamically eliminate outliers, thereby achieving flexible cell screening. However, this method relies solely on the K value as the screening basis and does not comprehensively analyze key parameters such as OCV (open-circuit voltage), DCIR (DC internal resistance), capacity, and thickness of the cell. This fails to meet the requirements for consistency of multi-dimensional parameters during module assembly, resulting in insufficient screening accuracy.

[0009] 2. Scattered data sources and lack of intelligent recursive mechanism: The patent CN118294832, "A method for screening the self-discharge rate of lithium iron phosphate cells," calculates the self-discharge rate of cells through multiple rounds of resting and voltage measurement, thereby screening out defective cells with high self-discharge rates. Although this method effectively controls the cell self-discharge parameters, its data acquisition is limited to a single process and lacks a unified processing and intelligent retrieval mechanism for multiple processes and multiple data sources. If a data source is abnormal or missing, the system cannot automatically recursively supplement other data sources, resulting in incomplete or inaccurate screening data.

[0010] 3. Fixed threshold algorithms struggle to adapt to process changes: Traditional sorting methods typically rely on preset parameter thresholds for judgment, lacking dynamic adjustment and self-learning capabilities. For example, in production, changes in cell condition, time period, and process conditions can cause parameter distribution to drift, and fixed threshold screening methods struggle to adapt to these changes in a timely manner, leading to some normally functioning cells being mistakenly judged as unqualified.

[0011] 4. Lack of training and continuous optimization mechanisms for algorithm models: The current sorting system lacks a mechanism for historical data backtracking and model training, making it impossible to train adaptive algorithm models based on different process states, cell characteristics, and other factors. For example, when faced with new batches of cells or new production processes, existing algorithms cannot adapt quickly, affecting the accuracy and efficiency of screening results. Furthermore, the lack of statistical analysis of data source hit rates, missing data rates, and other indicators makes it difficult to continuously optimize algorithm logic and data acquisition paths.

[0012] 5. Insufficient data result display and anomaly response: Existing sorting systems mostly present data results in simple text or report formats, failing to achieve dynamic visualization, marking, and tracking of abnormal data. Furthermore, the data results are not linked to hardware devices (such as alarms and robotic arms), lacking an automated anomaly response mechanism. Manual intervention is still required to select defective cells, increasing labor costs and the risk of errors.

[0013] 6. To address the above issues, there is an urgent need for an intelligent sorting method that combines software processing and hardware response to achieve dynamic acquisition, intelligent screening, and visual display of cell parameters. Furthermore, the method should improve screening accuracy and speed through algorithm optimization and model training, thereby meeting the high standards of consistency and process requirements in lithium battery module production. Summary of the Invention

[0014] The purpose of this invention is to overcome the shortcomings of the prior art and provide an intelligent sorting method, system, storage medium and device for energy storage units, so as to realize the dynamic acquisition, intelligent screening and visualization of cell parameters, and improve the screening accuracy and speed through algorithm optimization and model training, thereby meeting the high standard requirements of consistency and process requirements in lithium battery module production.

[0015] To achieve the above objectives / to solve the above technical problems, the present invention is implemented using the following technical solution: First aspect: A method for intelligent sorting of energy storage units, the method comprising: Define the key parameters and identify the data sources for energy storage units to determine the key parameters for energy storage unit sorting; Key process parameter data of energy storage units are collected through intelligent data source scheduling algorithms; By using a pre-established multi-parameter sorting model to filter the collected key process parameter data, qualified energy storage units are obtained.

[0016] The above technical solution can use multiple process parameters of the battery cell to screen and judge the consistency of the battery cell, rather than just using the K value.

[0017] Optionally, the intelligent scheduling algorithm for the data source includes: The system records whether each parameter was successfully matched with the data source, and the response time of the data returned each time a match was successfully made with the data source. As a result of historical requests; If one of the data sources fails to be acquired or the data is abnormal, a recursive algorithm is used to automatically recursively move to the next data source in order. The results of historical requests are used to calculate the priority of sending requests for key process parameters from different data sources, and the calculation results are used as the collected data.

[0018] The above technical solution records the content of the data extracted from the data source for subsequent intelligent scheduling of the data source.

[0019] Optionally, the recursive algorithm includes: Let D i For the i-th data source, V i For the corresponding parameter value, if V1 is null or has an abnormal value, recursion is executed:

[0020] .

[0021] The above technical solution uses historical request records to calculate a sequence of data response times from the data source, which is used as a reference for data scheduling.

[0022] Optionally, the step of calculating the priority of sending requests for key process parameters from different data sources using historical request results, and using the calculation results as collected data, includes: For each request for key process parameters of the battery cell, a data source is randomly selected from the existing data sources through a multi-source load protection method. If the request fails, another data source is randomly requested, and the request result and response time are recorded. By using historical request results within a preset time period, and grouping them by parameter name and data source, the total number of successful hits for each parameter in each data source is obtained. Number of failures Based on the parameter name, data source, and whether the entry with a hit rate of 1 is matched, the aggregated algorithm calculates the average response time for each data source when retrieving each parameter. ; Using the number of successful hits Number of failures The hit rate H of each parameter in each data source is calculated: ; For each parameter and each data source, filter out all data with a hit rate greater than [percentage missing]. The entries, and according to the average response time The data is sorted in ascending order to obtain a matching sequence. The energy storage unit parameters are then recursively requested to obtain the data according to this matching sequence.

[0023] The above technical solution further utilizes historical request records to calculate a sequence of data source request response times, which is used as a reference for data scheduling to improve response speed.

[0024] Optionally, the multi-source load protection method includes: For a given data source, if the last N requests are empty or the response time is greater than T... ulim Count the number of times the count is greater than or equal to N. lagIf this happens, the current data source is skipped, and the next data source in the sequence is selected for the data request.

[0025] Although the above technical solution can request data from a certain data source quickly, if data is not retrieved multiple times recently, it indicates that the data source may be under high load or experiencing network fluctuations. Therefore, the data source is skipped for load protection.

[0026] Optionally, the multi-source load protection method further includes: For each data source, set up a request bucket, with a maximum number of requests in the bucket being C. max At the same time, C can be replenished into the bucket every 1 second. add Each time a request occurs, 1 is decremented from the request bucket. When a request is scheduled to a data source but the number of requests in the request bucket of that data source is 0, the next data source is selected according to the sequence order to perform the data request operation.

[0027] The above technical solution limits the request traffic of a certain data source to prevent high request load on that data source.

[0028] Optionally, the multi-parameter sorting model set includes upper and lower limit threshold judgment, Z-Score method, and integrated scoring to determine the qualification of energy storage units.

[0029] The above technical solution combines traditional upper and lower limit thresholds with the statistical Z-Score scheme, integrating the scoring of multi-dimensional parameters of the battery cell to improve the effectiveness of consistency judgment.

[0030] Optionally, the upper and lower threshold determination includes: Preset upper limits for each parameter and lower limit value When the parameter value X satisfies: ; Return results It is 1 if it is true, otherwise it is 0; The Z-Score determination includes: Data mean: ; Data standard deviation: ; Z-Score: ; in The selected parameter sample set contains sample values, where N is the data size of the sample set, and X is the sample data for the current Z-Score calculation. After the Z-value of the cell parameter X used in this calculation is calculated based on the parameter sample set, the result is returned according to the set Z-value threshold. .

[0031] The above technical solution combines traditional upper and lower thresholds with the statistical Z-Score scheme to provide more consistency judgment methods.

[0032] Optionally, the integrated scoring includes: After obtaining the upper and lower limit threshold judgment results for each parameter Comparison with Z-Score results Then, based on production needs, different parameters are handled using either AND or OR operations. and The result of determining whether the parameter is qualified is obtained. The AND operation requires and Both equal to 1, the "OR" operation requires and Any parameter is set to 1; then different integrated scoring weights are assigned to each parameter. And set a lower limit for the integrated scoring of qualified battery cells. ; The results were obtained through integrated calculations. : ; Where K represents the number of cell parameters. The scoring weights for a certain battery cell parameter are set. The result of judging a certain cell parameter, when The energy storage unit is judged to be qualified at that time.

[0033] The above technical solution integrates the scoring of multi-dimensional parameters of the battery cell, improving the effectiveness of consistency judgment.

[0034] Optionally, the method further includes: marking the detected abnormal data in real time, triggering an alarm system, and discharging unqualified energy storage units.

[0035] The above technical solution provides real-time early warning and visualizes the warning results, facilitating waste discharge or downgrading operations.

[0036] Second aspect: A smart sorting system for energy storage units, characterized in that the system comprises: The parameter definition module is configured to define key parameters of energy storage units and identify data sources, thereby determining the key parameters for energy storage unit sorting. The data acquisition module is configured to collect key process parameter data of the energy storage unit through a data source intelligent scheduling algorithm; The multi-dimensional parameter screening module is configured to screen the collected key process parameter data through a pre-established multi-parameter sorting model to obtain qualified energy storage units.

[0037] Optionally, the data acquisition module includes a data source intelligent scheduling module, configured to: The system records whether each parameter was successfully matched with the data source, and the response time of the data returned each time a match was successfully made with the data source. As a result of historical requests; If one of the data sources fails to be acquired or the data is abnormal, a recursive algorithm is used to automatically recursively move to the next data source in order. The results of historical requests are used to calculate the priority of sending requests for key process parameters from different data sources, and the calculation results are used as the collected data.

[0038] The above technical solution dynamically acquires data from multiple data sources, accelerating data acquisition speed and improving data acquisition robustness.

[0039] Third aspect: A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent sorting method for energy storage units as described in any of the first aspects.

[0040] Fourth aspect: A device comprising: Memory, used to store instructions; A processor for executing the instructions to cause the device to perform operations implementing the intelligent sorting method for energy storage units as described in any of the first aspects.

[0041] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: This invention identifies key parameters for energy storage unit sorting by defining key parameters and identifying data sources. It then uses a pre-established multi-parameter sorting model to filter the collected key process parameter data, thereby selecting qualified energy storage units. This enables dynamic acquisition, intelligent screening, and visualization of cell parameters. Furthermore, it improves screening accuracy through algorithm optimization and model training, thus meeting the high standards of consistency and process requirements in lithium battery module production. This invention accelerates data acquisition speed and improves data acquisition robustness by dynamically acquiring data from multiple data sources. This invention uses intelligent filtering rules to improve filtering accuracy. Attached Figure Description

[0042] Figure 1 The diagram shown is a schematic of the sorting process of the energy storage unit of this invention. Figure 2 The diagram shown is a schematic representation of the data source of this invention; Figure 3 The diagram shown is a schematic diagram of the data recursion of this invention. Detailed Implementation

[0043] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0044] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are used only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0045] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; 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; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0046] Example 1:

[0047] like Figure 1 As shown, a smart sorting method for energy storage units is disclosed, the method comprising:

[0048] Step S1: Define the key parameters and data sources of the energy storage unit to determine the key parameters for energy storage unit sorting;

[0049] Step S2: Collect key process parameter data of the energy storage unit through the intelligent scheduling algorithm of the data source;

[0050] Step S3: The collected key process parameter data are screened using a pre-established multi-parameter sorting model to obtain qualified energy storage units.

[0051] The specific details of steps S1-S3 in this embodiment are as follows:

[0052] S1: Definition of key cell parameters and identification of data sources, determining the key process parameters for cell sorting, including but not limited to: OCV (open circuit voltage), capacity, DCIR (DC internal resistance), K value (voltage drop per unit time in static state), thickness, and capacity.

[0053] S2: Data source identification: Data sources include MES (Manufacturing Execution System), databases, and host computer interfaces of test equipment, etc.

[0054] S3. Parameter corresponding data source mark: Mark the priority and redundant path of each data source to form a data source mapping relationship.

[0055] S4. Data Acquisition Module: Designed to interact with multiple data sources such as the MES system, database, and testing equipment to acquire key parameter data. This includes interfaces for connecting to multiple data sources to dynamically acquire key process parameters of the battery cells, such as OCV, DCIR, capacity, thickness, and K-value. All key process parameters of the battery cells to be sorted are ultimately obtained from the data acquisition module, such as… Figure 2 As shown.

[0056] S5, Intelligent Data Source Scheduling Module: Records request-response behavior results, which are the results of data collection from different data sources for different key process parameters. It integrates recursive acquisition logic and a data source scheduling sequence generation algorithm. The main purpose of this module is to design a multi-data intelligent scheduling model for switching target data sources and quickly and accurately collecting the required key process parameter data.

[0057] S6. Record whether each parameter of the data record was successfully matched from the data source, and the response time of the data returned each time a data match is successfully matched from the data source. The data recorded here is shown in Table 1.

[0058] Table 1. Data Source Hit Record Table for Battery Cell Parameters

[0059]

[0060] In the "Hit or Not" column, 0 indicates a failed hit, and 1 indicates a successful hit.

[0061] S7. Data Recursion Mechanism: If data source A fails to retrieve data or the data is abnormal, the system automatically recursively proceeds to the next next data source for data retrieval. Figure 3 As shown;

[0062] Recursive algorithm logic:

[0063] Let D i For the i-th data source, V i This corresponds to the parameter value. If V1 is null or has an abnormal value, recursion is executed:

[0064]

[0065] S8. Data source scheduling: Utilize the historical request results of cell parameters in step S6 to calculate the priority of sending requests for key process parameters from different data sources, and provide the calculation results to the data acquisition module for data acquisition.

[0066] I. Function initialization: For each request for key process parameters of the battery cell, a data source is randomly selected from the existing data sources for request. If the request fails, another data source is randomly requested, and the request result and response time are recorded.

[0067] II. Scheduling Sequence Generation: Utilizing historical request results from the past 2 hours, grouping by parameter name and data source, and aggregating the counts, we obtain the number of successful hits for each parameter across each data source. Number of failures Based on the parameter name, data source, and whether the entry with a hit rate of 1 is matched, the aggregated algorithm calculates the average response time for each data source when retrieving each parameter. Utilizing the number of successful hits Number of failures The hit rate H of each parameter in each data source is calculated:

[0068] ;

[0069] The results after calculation are shown in Table 2:

[0070] Table 2. Data Source Hit Rate and Average Response Time of Cell Parameters

[0071]

[0072] For each parameter and each data source, filter out all those with a hit rate greater than [specified value]. The entries, and according to the average response time The data is sorted in ascending order, and the sorted result is the data acquisition matching sequence. Then, the cell parameters are recursively requested according to this matching sequence. In a practical example of this invention, this scheduling sequence is updated every 3 minutes. Set to 75%.

[0073] III. Multi-Source Load Balancing: After obtaining the data source scheduling sequence from Step I, the target data source is selected in a balanced manner, considering both the core stability of the data source and request traffic limits. In actual production environments, request concurrency is high, and top-ranked data sources receive a large number of requests, leading to high loads on the target data source. To address the potential for delayed data request responses due to high server load and to balance the request load across data sources, the following two methods are used for load balancing:

[0074] 1. Core stability calculation: For a given data source, if the last N requests are empty or the response time is greater than T... ulim Count the number of times the count is greater than or equal to N. lag In this case, the current data source is skipped, and the next data source in the sequence is selected for the data request. In a practical example of this invention, N is 100, N lag Take 8, T ulim Take 200ms.

[0075] 2. Request traffic limiting: For each data source, set up a request bucket, with a maximum number of requests per bucket of C. max At the same time, C can be replenished into the bucket every 1 second. add Each time a request occurs, 1 is decremented from the request bucket. When a request is scheduled to a data source but the request bucket for that data source has 0 requests, the next data source is selected based on the sequence order to perform the data request operation. In a practical example of this invention, C max Take 500, C add Take 50.

[0076] Once the target data source for this request is identified, the relevant key process parameters can be collected through the S4 data acquisition module and used for subsequent calculations.

[0077] S9. Multi-dimensional Parameter Screening Module: Based on the cell grading standards, this module defines the upper and lower limits of each key process parameter, establishes a multi-parameter sorting model, and dynamically filters the data. The multi-parameter sorting model combines upper and lower limit threshold judgment, the Z-Score method, and an integrated scoring approach to determine the cell's qualification.

[0078] I. Upper and lower limit threshold judgment: These are the upper limits of each parameter set manually based on the engineer's experience (see Tables 3 and 4). and lower limit value When the parameter value X satisfies:

[0079] ;

[0080] Return results It is 1 if it is true, otherwise it is 0;

[0081] II. Z-Score Judgment: Based on the position of the data relative to the population mean, it represents the difference between the sample value and the mean.

[0082] Data mean:

[0083] ;

[0084] Data standard deviation:

[0085] ;

[0086] Z-Score:

[0087] ;

[0088] in, Let N be the sample values ​​in the selected parameter sample set, N be the data size of the sample set, and X be the sample data for calculating the Z-Score. For each parameter, the selection of the parameter sample set can be divided into two types: one is data from all battery cells produced in the same batch, and the other is data from battery cells produced in the same batch within a certain period. Different sample sets can be selected based on the actual battery cell production situation. In a practical example of this invention, data from qualified battery cells produced in the same batch within the previous week (from Monday to Sunday of the previous week) are selected as the parameter sample set to calculate the Z-Score for different grades. as well as Once the Z-value of the cell parameter X used in this calculation is obtained from the parameter sample set, the result is returned based on the set Z-value threshold. In a practical example of this invention, the Z-value threshold is set to 3, that is, when hour, ,otherwise .

[0089] III. Integrated scoring algorithm: Based on the upper and lower thresholds of each parameter, a judgment result is obtained. Comparison with Z-Score results Then, depending on actual production needs, either an AND operation or an OR operation can be used for different parameters. and The result of determining whether the parameter is qualified is obtained. The AND operation requires and Both equal to 1, the "OR" operation requires and Any parameter is set to 1; then different integrated scoring weights are assigned to each parameter. And set a lower limit for the integrated scoring of qualified battery cells. .

[0090] The results were obtained through integrated calculations. :

[0091] ;

[0092] Where n is the number of cell parameters, The scoring weights for a certain battery cell parameter are set. This is the result of judging a certain cell parameter. When The battery cell is judged to be qualified at that time.

[0093] In a practical example of this invention, the number of cell parameters n involved in the scoring is 5. The cell parameters are calculated using an AND operation, and the scoring weights of each cell parameter are determined. All are 1, the lower limit of the rating. The weight is 5; in summary, the integrated scoring weight of each cell parameter is consistent, and the cell can only be judged as a qualified graded cell if all 5 cell parameters pass the judgment result.

[0094] Tables 3 and 4 show the cell parameter data and aggregate scoring results of 10 cells in one example of the present invention:

[0095] Table 3. Cell Parameter Data

[0096]

[0097] Table 4. Cell Assembly Scoring Judgment

[0098]

[0099] S10 Waste Discharge Module: Based on the calculation results of the cell screening module designed in S9, the detected abnormal data is marked in real time, triggering the alarm system and sending instructions to the robot to complete the cell waste discharge work.

[0100] S11. Visualization Module: This module dynamically displays the filtering results through the software interface, including cell metadata, cell parameters, requested data sources, and abnormal parameter markers. It also provides an export function, allowing the filtered cell parameters and results to be displayed as charts or dynamic graphs, providing real-time feedback on parameter distribution. Users can perform backtracking analysis on abnormal data and adjust the filtering strategy.

[0101] The above embodiments propose several technological innovations in the field of intelligent sorting methods in lithium battery production, specifically including the following:

[0102] Dynamic Acquisition and Recursive Mechanism from Multiple Data Sources: A mechanism for dynamically acquiring parameters from multiple data sources is proposed. When a data source fails or data is missing, the system can automatically recursively acquire data from other available data sources to supplement the parameters, ensuring the integrity and accuracy of the cell parameters.

[0103] Intelligent screening algorithm and dynamic rule adjustment: By introducing intelligent algorithms based on historical data analysis, screening rules are dynamically generated to adapt to the production environment and process requirements, enabling the screening process to flexibly respond to changes in different time periods, equipment status and incoming material characteristics.

[0104] Algorithm encapsulation and data source optimization feedback mechanism: The screening algorithm is encapsulated as a software module, which can record indicators such as the acquisition speed, accuracy, hit rate and missing rate of the data source during each screening process. Based on this feedback, the data source selection strategy and algorithm performance are continuously optimized.

[0105] Visualization and anomaly detection of screening results: It innovatively provides a combination of static and dynamic anomaly data visualization function, intuitively marks abnormal parameters in the screening results, and provides real-time alarms for abnormal data, and completes automatic waste removal or prompting by driving hardware devices such as alarms and robotic arms.

[0106] Model training and parameter prediction capabilities: Based on historical data from multiple screening operations, a parameter matching model is trained for different production periods, process conditions, and cell states. The model predicts the parameter distribution of new incoming materials, improving the intelligence and accuracy of the screening process.

[0107] Closed-loop optimization of parameter screening results: By evaluating the module matching performance of the screened cells and checking the matching degree of the algorithm, the screening logic and model parameters are further optimized to form a complete closed-loop optimization mechanism.

[0108] System design that integrates hardware and software: Deeply integrates software processing and hardware response to achieve full-process automation from data acquisition to screening, anomaly handling and waste disposal, thereby improving production efficiency and reducing manual intervention.

[0109] Modular system architecture adaptable to multiple operating conditions: The system adopts a modular design, which can be flexibly configured and expanded according to the actual needs of the factory. It is suitable for different lithium battery production lines and has good adaptability and scalability.

[0110] Example 2 discloses an intelligent sorting system for energy storage units, the system comprising:

[0111] The parameter definition module is configured to define key parameters of energy storage units and identify data sources, thereby determining the key parameters for energy storage unit sorting.

[0112] The data acquisition module is configured to collect key process parameter data of the energy storage unit through a data source intelligent scheduling algorithm;

[0113] The multi-dimensional parameter screening module is configured to screen the collected key process parameter data through a pre-established multi-parameter sorting model to obtain qualified energy storage units.

[0114] In this embodiment, the data acquisition module includes a data source intelligent scheduling module, which is configured to:

[0115] The system records whether each parameter was successfully matched with the data source, and the response time of the data returned each time a match was successfully made with the data source. As a result of historical requests;

[0116] If one of the data sources fails to be acquired or the data is abnormal, a recursive algorithm is used to automatically recursively move to the next data source in order. The results of historical requests are used to calculate the priority of sending requests for key process parameters from different data sources, and the calculation results are used as the collected data.

[0117] The above technical solution dynamically acquires data from multiple data sources, accelerating data acquisition speed and improving data acquisition robustness.

[0118] Dynamic acquisition from multiple data sources improves data integrity and reliability.

[0119] Intelligent filtering rules are generated to improve filtering accuracy.

[0120] Anomaly detection and real-time response reduce the cost of manual intervention.

[0121] The screening results are visualized to improve user experience and decision-making ability.

[0122] Closed-loop feedback optimization improves system adaptability and algorithm performance.

[0123] Example 3: A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent sorting method for energy storage units as described in any one of Examples 1.

[0124] Fourth aspect: A device comprising:

[0125] Memory, used to store instructions;

[0126] A processor is configured to execute the instructions, causing the device to perform operations that implement the intelligent sorting method for energy storage units as described in any of Embodiment 1.

[0127] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0129] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0130] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0131] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for intelligent sorting of energy storage units, characterized in that, The method includes: Define the key parameters and identify the data sources for energy storage units to determine the key parameters for energy storage unit sorting; Key process parameter data of energy storage units are collected through intelligent data source scheduling algorithms; By using a pre-established multi-parameter sorting model to filter the collected key process parameter data, qualified energy storage units are obtained.

2. The intelligent sorting method for energy storage units according to claim 1, characterized in that, The intelligent scheduling algorithm for the data source includes: The system records whether each parameter was successfully matched with the data source, and the response time of the data returned each time a match was successfully made with the data source. As a result of historical requests; If one of the data sources fails to be acquired or the data is abnormal, a recursive algorithm is used to automatically recursively move to the next data source in order. The results of historical requests are used to calculate the priority of sending requests for key process parameters from different data sources, and the calculation results are used as the collected data.

3. The intelligent sorting method for energy storage units according to claim 2, characterized in that, The recursive algorithm includes: Let D i be the ith data source, V i be the corresponding parameter value, when V1 is null or numerical anomaly, execute recursion: 。 4. The intelligent sorting method for energy storage units according to claim 2, characterized in that, The calculation of the priority of sending requests for key process parameters from different data sources using historical request results, and the use of the calculation results as collected data, includes: For each request for key process parameters of the battery cell, a data source is randomly selected from the existing data sources through a multi-source load protection method. If the request fails, another data source is randomly requested, and the request result and response time are recorded. By using historical request results within a preset time period, and grouping them by parameter name and data source, the total number of successful hits for each parameter in each data source is obtained. Number of failures Based on the parameter name, data source, and whether the entry with a hit rate of 1 is matched, the aggregated algorithm calculates the average response time for each data source when retrieving each parameter. ; Using the number of successful hits Number of failures The hit rate H of each parameter in each data source is calculated: ; For each parameter and each data source, filter out all data with a hit rate greater than [percentage missing]. The entries, and according to the average response time The data is sorted in ascending order to obtain a matching sequence. The energy storage unit parameters are then recursively requested to obtain the data according to this matching sequence.

5. The intelligent sorting method for energy storage units according to claim 4, characterized in that, The multi-source load protection method includes: The request of a data source is empty or the response time is greater than T for nearly N times ulim Counting is performed, and when the counting number is greater than or equal to N lag times, the data source is skipped, and the data source in the next sequence order is selected for data request behavior.

6. The intelligent sorting method for energy storage units according to claim 5, characterized in that, The multi-source load protection method further includes: For each data source, set up a request bucket, with a maximum number of requests in the bucket being C. max At the same time, C can be replenished into the bucket every 1 second. add Each time a request occurs, 1 is decremented from the request bucket. When the number of requests in the bucket is 0, the next data source is selected according to the sequence order to perform the data request operation.

7. The intelligent sorting method for energy storage units according to claim 1, characterized in that, The multi-parameter sorting model combines upper and lower limit threshold judgment, Z-Score method, and integrated scoring to determine the qualification of energy storage units.

8. The intelligent sorting method for energy storage units according to claim 7, characterized in that, The upper and lower threshold determination includes: Preset upper limits for each parameter and lower limit value When the parameter value X satisfies: ； Return results It is 1 if it is true, otherwise it is 0; The Z-Score determination includes: Data mean: ; Data standard deviation: ; Z-Score: ; in For the sample values ​​in the selected parameter sample set, This represents a specific parameter value for a single cell in the sample. N is the data size of the sample set, and X is the current Z-Score sample data for calculation. After the Z-value of the cell parameter X used in this calculation is calculated based on the parameter sample set, the result is returned according to the set Z-value threshold. ; This represents a score indicating the integration performance of the battery cell. Another score in the cell integration rating is the calculated Z-Score, which is equal to 1 when the Z-Score is less than the set Z value threshold and equal to 0 otherwise.

9. The intelligent sorting method for energy storage units according to claim 8, characterized in that, The integrated scoring includes: After obtaining the upper and lower limit threshold judgment results for each parameter Comparison with Z-Score results Then, based on production needs, either an AND or OR operation is used for different parameters. and The result of determining whether the parameter is qualified is obtained. The "AND" operation requires and Both equal to 1, the "OR" operation requires and After any parameter is equal to 1, different integrated scoring weights are set for each parameter. And set a lower limit for the integrated scoring of qualified battery cells. ; The results were obtained through integrated calculations. : ; Where K represents the number of cell parameters. The scoring weights for a certain battery cell parameter are set. The result of judging a certain cell parameter, when When determining whether an energy storage unit is qualified, This refers to a specific parameter within a battery cell.

10. The intelligent sorting method for energy storage units according to claim 1, characterized in that, The method also includes: marking the detected abnormal data in real time, triggering the alarm system, and discharging the unqualified energy storage units.

11. An intelligent sorting system for energy storage units, characterized in that, The system includes: The parameter definition module is configured to define key parameters of energy storage units and identify data sources, thereby determining the key parameters for energy storage unit sorting. The data acquisition module is configured to collect key process parameter data of the energy storage unit through a data source intelligent scheduling algorithm; The multi-dimensional parameter screening module is configured to screen the collected key process parameter data through a pre-established multi-parameter sorting model to obtain qualified energy storage units.

12. The intelligent sorting system for energy storage units according to claim 11, characterized in that, The data acquisition module includes a data source intelligent scheduling module, configured to: The system records whether each parameter was successfully matched with the data source, and the response time of the data returned each time a match was successfully made with the data source. As a result of historical requests; If one of the data sources fails to be acquired or the data is abnormal, a recursive algorithm is used to automatically recursively move to the next data source in order. The results of historical requests are used to calculate the priority of sending requests for key process parameters from different data sources, and the calculation results are used as the collected data.

13. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the intelligent sorting method for energy storage units as described in any one of claims 1-10.

14. A device, characterized in that, include: Memory, used to store instructions; A processor is configured to execute the instructions, causing the device to perform operations that implement the intelligent sorting method for energy storage units as described in any one of claims 1-10.

Images