Big data consistency intelligent detection system and method based on battery pack electrical performance

Abstract

The invention provides an intelligent detection system and method based on battery pack electrical performance big data consistency, and belongs to the field of battery pack electrical performance test analysis, and the system comprises a data access module, a standard configuration module, and an intelligent algorithm library module which comprises a standard algorithm library, an electrical performance algorithm model library, and a chart analysis algorithm model library. The standard algorithm library is configured to integrate various algorithm models, the electrical performance algorithm model library is configured to match corresponding algorithms for electrical performance detection items from the standard algorithm library according to historical electrical measurement data and electrical measurement production lines, procedures and equipment, and the chart analysis algorithm model library is configured to match corresponding algorithms for electrical performance detection items from the standard algorithm library according to the historical electrical measurement data and the electrical measurement production lines, procedures and equipment. Matching different analysis chart types for the electrical performance detection items; the electrical performance detection module is used for respectively configuring data names needing to be extracted for different projects, equipment manufacturers and electrical testing steps and carrying out highlight labeling; a data analysis module; huge electrical testing data are processed, and the efficiency and accuracy of electrical performance consistency detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery pack electrical performance testing and analysis, and in particular to an intelligent detection system and method based on battery pack electrical performance big data consistency. Background Art

[0002] At present, the consistency analysis of battery pack electrical performance mainly relies on offline manual analysis of huge electrical test files, manual operation to obtain electrical test files on the device side, and then offline data analysis through cumbersome file export process. The electrical test data generated by each battery pack is as high as 20 million, which makes the manual screening process extremely time-consuming and consumes a lot of manpower costs. More importantly, there is a significant risk of missed detection and misjudgment in manual screening of electrical performance, which may cause battery packs with potential defects to be mistakenly used for loading, thus posing a potential safety hazard. In addition, the current offline electrical performance test results are only presented in the form of pure data, lacking intuitive graphical display, making early warning difficult. Finally, the manually uploaded data is often not fully displayed in the shipment report, further reducing the comprehensiveness and accuracy of the analysis.

[0003] In the prior art, the Chinese invention patent application "Intelligent Comprehensive Computation Method, System, Equipment and Medium for Lithium Battery Test Data" with publication number CN118410224A includes: S1. Extracting multi-category data from the server database, filtering and processing to obtain result data; S2. Integrating result data and data in related systems to obtain integrated data, and storing them in clickhouse database and Mysql database for unified management to obtain integrated data; S3. Analyzing and processing the integrated data, adapting corresponding task parameters, forming work task modules of different data analysis methods, and obtaining work task analysis data results; S4. Based on the analysis data results of different work tasks, combined with lithium battery application environment factors, building a Pack electrical performance simulation model; S5. Visualizing the analysis data results; S6. Storing the analysis data results and providing query data download. In the above method, the steps of data extraction, data budgeting and data analysis adopt simple conceptual descriptions, without specific technical analysis descriptions and steps. At the same time, the method does not provide how to perform consistency analysis on different types of electrical performance data, which is not conducive to improving the efficiency of battery pack electrical performance consistency detection. Summary of the invention

[0004] The technical problem to be solved by the present invention is how to improve the efficiency and accuracy of battery pack electrical performance consistency detection.

[0005] The present invention solves the above technical problems through the following technical solutions: an intelligent detection system based on battery pack electrical performance big data consistency, comprising:

[0006] Data access module, used to extract historical electrical measurement data and configuration parameters, the historical electrical measurement data includes dynamic temperature difference, charge and discharge dynamic pressure difference, charge and discharge static pressure difference, charge and discharge capacity, charge and discharge energy, charge and discharge DCR, current accuracy, SOC, ICC, discharge temperature rise, process module temperature difference, and pre-measurement temperature;

[0007] Standard configuration module, used to configure and manage electrical test items, test item ranges, test steps, and display chart types;

[0008] An intelligent algorithm library module is used to adapt to the electrical performance detection data analysis of multiple types of products, equipment and electrical measurement files. The intelligent algorithm library module includes a standard algorithm library, an electrical performance algorithm model library, and a chart analysis algorithm model library. The standard algorithm library is configured to integrate multiple algorithm models. The electrical performance algorithm model library is configured to match the electrical performance detection items to corresponding algorithms from the standard algorithm library based on historical electrical measurement data and electrical measurement production lines, processes, and equipment. The chart analysis algorithm model library is configured to match different analysis chart types for electrical performance detection items based on historical electrical measurement data and electrical measurement production lines, processes, and equipment.

[0009] The electrical performance detection module is used to configure the data names to be extracted for different projects, equipment manufacturers, and electrical measurement steps, and select the corresponding chart type. The background automatically highlights the maximum, minimum, and second minimum values ​​of each chart.

[0010] The data analysis module is used to query the analysis results of historical electrical measurement data.

[0011] Beneficial effect: The present invention constructs a specific data structure and a matching efficient parsing algorithm in the intelligent algorithm library module. The standard algorithm library is configured to integrate multiple algorithm models, which can perform consistency analysis on different types of electrical measurement data. The electrical performance algorithm model library is configured to match the electrical performance detection items to the corresponding algorithms from the standard algorithm library according to historical electrical measurement data and electrical measurement production lines, processes, and equipment. It can process huge electrical measurement data files, solving the problems of huge data, time-consuming manual screening, high labor costs, and insufficient consistency and accuracy in traditional methods.

[0012] Preferably, the standard algorithm library includes a slice-type difference algorithm, a slice-type accumulation algorithm, data extraction, interval maximum value divide-and-conquer calculation, interval minimum value divide-and-conquer calculation, interval extreme value summation calculation, interval extreme value difference calculation, horizontal derivative difference calculation, vertical derivative difference calculation, slice-type difference-quotient operation, row threshold positioning difference calculation, inter-module voltage trend algorithm, and linear regression.

[0013] Beneficial effects: By integrating multiple efficient algorithms in the standard algorithm library, a unified statistical algorithm is constructed for each small table data (the same group), which can process each small table in a loop and calculate the maximum and minimum vertical values. The concept of horizontal difference calculation is introduced. By calculating the differences of all columns, the maximum and minimum horizontal differences are calculated, achieving more comprehensive and in-depth data analysis.

[0014] Preferably, the calculation process of the slice difference algorithm is: take the value of step A at the Tth second as X AT and the value of the T+n(n>0) second is X A(T+n) , respectively calculate the difference X i =X AT -X A(T+n) , extract X i The value is displayed and graphed; the calculation process of the slice accumulation algorithm is: take the value of step A at the Tth second as X AT , take the value of the Tth second of the A+n (n≥0) step as X (A+n)T , do cumulative calculations Extract X i The values ​​are displayed and graphed.

[0015] Beneficial effects: By configuring a slice-type difference algorithm, the relevant pressure difference data of the battery pack can be extracted and analyzed, and the pressure difference performance of the battery pack during the charge and discharge process can be controlled according to the process threshold. By configuring a slice-type accumulation algorithm, the battery discharge energy index can be extracted and analyzed, and the energy and capacity performance of the battery pack during the charge and discharge process can be controlled according to the process threshold.

[0016] Preferably, the process of calculating the interval maximum value by divide-and-conquer is as follows: take the value of [α1, α2, ..., α2] in the interval [T, T+n] seconds of step A corresponding to the value of [α1, α2, ..., α2] in the interval [Y, Y+n] columns. i ] Maximum value α Max The process of calculating the minimum value of the interval by dividing and conquering is as follows: take the value of [α1,α2,...,α in the interval [T,t+n] seconds of step A corresponding to the value of [α1,α2,...,α in the interval [Y,y+n] i ]Minimum value α Min , where T represents time and Y represents the number of columns.

[0017] Beneficial effect: By configuring the interval maximum value divide-and-conquer calculation and the interval minimum value divide-and-conquer calculation, the data extraction and analysis of the maximum and minimum values ​​of the pressure difference indicator during the charging and discharging process are realized, the difference between the data is intuitively displayed, and data support is provided for the pressure difference performance evaluation of the battery pack.

[0018] Preferably, the process of calculating the interval extreme value sum is: take the A step [T, T+n] seconds interval corresponding to the [Y, Y+n] column interval [α1, α2, .., α i ] Maximum value αMax and the minimum value α Min , do α Max +α Min Sum calculation; the process of calculating the interval extreme value difference is: take the A step [T, T+n] seconds interval corresponding to the [Y, Y+n] column interval [α1, α2, .., α i ] Maximum value α Max and the minimum value α Min , do α Max -α Min Difference calculation, T represents time, Y represents the number of columns.

[0019] Beneficial effect: By configuring the interval extreme value sum calculation and the interval extreme value difference calculation, the actual pressure difference of the measured voltage data can be extracted, providing data support for the pressure difference performance evaluation of the battery pack.

[0020] Preferably, the process of calculating the lateral derivative difference is as follows: taking the value of the [i, j, ..., k, ...] column at the Tth second of step A, and calculating the derivative value X Max =Max[(T ai -T aj -T ak -...),(T bi -T bj -T bk -..),(T ci -T cj -T ck -..)...], derived value X Min =Min[(T ai -T aj -T ak -...),(T bi -T bj -T bk -..),(T ci -T cj -T ck -..)...], T represents time, subscripts a, b, c... represent row numbers, and subscripts i, j, k... represent column numbers; the process of calculating the longitudinal derivative difference is: take the value of the [A1, A2, ..., An]th column in any process step, and calculate the maximum value Y of the derivative value difference Max =Max[(A1a-A1b-A1c...),(A2a-A2b-A2c...),...,(Ana-Anb-Anc...)], Y Min =Min[(A1a-a1b-A1c...),(A2a-A2b-A2c...),...,(Ana-Anb-Anc...)].

[0021] Beneficial effect: By configuring the lateral and longitudinal derivative difference calculations, the current accuracy (Hall charge and discharge current, shunt charge and discharge current) can be automatically analyzed, thereby achieving quality control of the current detection items.

[0022] Preferably, it also includes:

[0023] Data management module, used to record operation errors and warning prompts;

[0024] The data comparison module is used to compare the analysis results of historical electrical measurement data with offline manual statistical data, compare and analyze battery packs of the same dimension, verify the accuracy of the data, and display summary information of system data;

[0025] The system management module includes part management, user management, and role management, which are used to authorize and group user accounts.

[0026] The present invention also provides an intelligent detection method based on battery pack electrical performance big data consistency, comprising the following steps:

[0027] Step 1: After the battery pack test is completed, the electric test file is automatically generated, and the data access module scans the electric test file at set intervals to obtain the electric test file data;

[0028] Step 2: After configuring the electrical test items, test item ranges, test steps, and display chart types in the standard configuration module, the model library in the intelligent algorithm library module is called to perform real-time analysis on the electrical test file data;

[0029] Step 3: The electrical performance detection module highlights the analysis results and sends them to emails, intercepting defective batteries whose electrical performance indicators exceed the process threshold.

[0030] Beneficial effects: By setting the dynamic signal column of the electrical test file, identifying and reading the precise position information of each signal column in the complex table, using precise algorithms for intelligent segmentation and reorganization, ensuring that every piece of data as the input feature of the algorithm is accurate, using advanced model data structure analysis engine, automatically and efficiently mapping to the huge system algorithm library, accurately selecting the most suitable algorithm combination for the current task, the data analysis time of a single electrical test file can be shortened from 12 minutes of manual operation to 7.2 seconds of automatic system processing, and the efficiency is significantly improved by up to 100 times. At the same time, the system has also reached a high level of 99.5% in terms of analysis accuracy, ensuring the accuracy and reliability of detection and analysis.

[0031] Preferably, the step 1 comprises:

[0032] 1.1. Introduce the OpenXML library, use the C# file I / O function to open and read the electrical test file, and use the SpreadsheetDocument class in the OpenXML library to read the csv file content when reading the content of the electrical test file. Use the SpreadsheetDocument.Open method to open the file and obtain the WorkbookPart and WorksheetPart components.

[0033] 1.2. Traverse the Sheets element in WorkbookPart, obtain the names and IDs of all worksheets, select the worksheet to be parsed according to the user's selection or default settings, find the corresponding WorksheetPart by the worksheet ID, parse the SheetData element, which contains all the row and column information in the worksheet, traverse the Row element in SheetData and the Cell element in each Row, determine the row number and column number of each cell, and construct the two-dimensional table structure of the csv file;

[0034] 1.3. According to user needs, set the rules or conditions for extracting cells. When traversing the two-dimensional table structure of the CSV file, perform logical judgment on each cell, determine the cells that meet the extraction requirements according to the set rules, and mark them. According to the marked cell position, extract the content of the target cell from the two-dimensional table structure, remove spaces and convert data types, and save them to the custom data structure according to the predetermined format.

[0035] Preferably, the step 2 comprises:

[0036] 2.1. Enter a unique process number to identify the csv file or data set currently being processed. The user selects the table row area to be processed, locates the table row area, provides a date selector and a time input box, and the user enters or selects a date range to filter rows that meet the time conditions. A drop-down list is provided to list all column names in the csv file. The user can select one or more column names to determine the column area.

[0037] 2.2. Select an algorithm from the standard algorithm library in the electrical performance algorithm model library, combine the selected row area, column area and algorithm, and construct a custom model data structure that meets your needs;

[0038] 2.3. Based on the configuration information, the background system obtains the algorithm in the standard algorithm library and automatically maps and matches it. It extracts the table content based on the selected row and column areas and passes it to the algorithm for calculation. The calculation results are displayed to the user in the form of charts and reports. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1An architecture diagram of an intelligent detection system for battery pack electrical performance big data consistency provided by an embodiment of the present invention;

[0040] Figure 2 An example diagram of calculation based on the intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0041] Figure 3 A diagram showing calculation results of an intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0042] Figure 4 An example diagram of calculation of lateral derivative difference in an intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0043] Figure 5 A diagram showing the calculation results of the lateral derivative difference in the intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0044] Figure 6 A diagram showing the calculation results of the longitudinal derivative difference in the intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0045] Figure 7 A calculation example diagram of a slice difference-quotient operation in a battery pack electrical performance big data consistency intelligent detection system provided by an embodiment of the present invention;

[0046] Figure 8 An example diagram of calculating the difference in threshold positioning value in the intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0047] Fig. 9 An example diagram of calculation of an inter-module voltage trend algorithm in an intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0048] Fig.10 An example diagram of calculation of linear regression in an intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0049] Fig.11 A schematic diagram of an electric performance algorithm model library in an intelligent detection system for consistency of battery pack electric performance big data provided by an embodiment of the present invention;

[0050] Fig.12 A schematic diagram of a chart analysis algorithm model library in an intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0051] Fig.13A discrete graph of the analysis of static discharge data based on the intelligent detection system for consistency of battery pack electrical performance big data provided by an embodiment of the present invention;

[0052] Fig.14 A DCR data analysis line chart based on a battery pack electrical performance big data consistency intelligent detection system provided by an embodiment of the present invention;

[0053] Fig.15 A combined discrete graph for analyzing dynamic voltage data of charge and discharge based on a battery pack electrical performance big data consistency intelligent detection system provided by an embodiment of the present invention;

[0054] Fig.16 A flow chart of an intelligent detection method for consistency of battery pack electrical performance big data provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the technical solution of the present invention is clearly and completely described below in combination with specific embodiments and with reference to the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] like Figure 1 As shown, this embodiment provides an intelligent detection system based on battery pack electrical performance big data consistency, including:

[0057] The data access module is used to extract historical electrical measurement data and configuration parameters of electrical measurement production lines, processes, equipment, etc., and verify the process step number of the electrical measurement file; the historical electrical measurement data includes key indicators such as dynamic temperature difference, dynamic pressure difference of charge and discharge, static pressure difference of charge and discharge, charge and discharge capacity, charge and discharge energy, charge and discharge DCR, current accuracy, SOC, ICC, discharge temperature rise, process module temperature difference, and pre-measurement temperature;

[0058] Standard configuration module, used to configure and manage electrical test items, test item ranges, test steps, and display chart types;

[0059] The intelligent algorithm library module is used to adapt to the electrical performance detection data analysis of various types of products, equipment and electrical measurement files, including the standard algorithm library, the electrical performance algorithm model library, and the chart analysis algorithm model library. The standard algorithm library is configured to integrate multiple algorithm models, including slice difference algorithm, slice accumulation algorithm, data extraction, interval maximum value divide-and-conquer calculation, interval minimum value divide-and-conquer calculation, interval extreme value summation calculation, interval extreme value difference calculation, horizontal derivative difference calculation, vertical derivative difference calculation, slice difference_quotient operation, row threshold positioning difference calculation, inter-module voltage trend algorithm, linear regression, and recursive algorithm, which can perform consistency analysis on different types of electrical measurement data.

[0060] The calculation process of the slice difference algorithm is as follows: take the value of step A at the Tth second as X AT and the value of the T+n(n>0) second is X A(T+n) , respectively calculate the difference X i =X AT -X A(T+n) , extract X i The values ​​can be displayed and graphed for the extraction and automated analysis of historical electrical measurement data, and the electrical performance data can be controlled according to process thresholds.

[0061] like Figure 2 As shown, taking voltage as an example, the value of the first column in the last second of step 11 is A1, and the value of the first second of step 12 is A2. The difference of the voltage in the first column is X1=A1-A2. Similarly, the difference of the voltage in the second column, the third column, ..., the i-th column is calculated as X2, X3, ..., X i , take [X1,...,X i ] values ​​to draw slice difference line charts and scatter plots. Figure 3 As shown, the slice difference algorithm is used to extract and automatically analyze the pressure difference data related to the battery pack, and control the pressure difference performance of the battery pack during the charging and discharging process according to the process threshold.

[0062] The calculation process of the slice accumulation algorithm is: take the value of step A at the Tth second as X AT , take the value of the Tth second of the A+n (n≥0) step as X (A+n)t , do cumulative calculations Extract X i The values ​​are displayed and graphed.

[0063] like Figure 2 As shown, taking voltage as an example, the value of the first column in the last second of step 11 is A1, and the value of the first second of step 12 is A2, and the accumulation of voltage 1 is X1=A1+A2; take [X1,...,X i ] values ​​to draw slice-type cumulative line charts and scatter plots. Figure 3As shown, the slice difference algorithm is used to extract and automatically analyze the pressure difference data related to the battery pack, and control the pressure difference performance of the battery pack during the charging and discharging process according to the process threshold.

[0064] The process of data extraction is: take the Tth second of step A Column values ​​are displayed and graphed.

[0065] like Figure 2 In the configuration shown, the value of the voltage in the first column at the last second of step 12 is A1, and the value of the voltage in the second column at the last second of step 12 is A2. The values ​​of [A1, A2, ..., Ai] are used to draw the data extraction line graph and scatter graph. Figure 3 As shown, by configuring data extraction, data extraction and automated analysis of battery SOC indicators are performed, and the SOC performance of the battery pack charge and discharge process is controlled according to the process threshold.

[0066] The process of calculating the interval maximum value by divide-and-conquer is as follows: take the value of [α1,α2,...,α in the interval [T,T+n] seconds of step A corresponding to the value of [α1,α2,...,α in the interval [Y,Y+n] columns. i ] Maximum value α Max . Where T represents time and Y represents the number of columns, such as Figure 3 As shown, through the divide-and-conquer calculation of the maximum value between configurations, the maximum value of the pressure difference indicator during the charging and discharging process is used for data extraction and automatic analysis, which intuitively displays the differences between the data and provides data support for the pressure difference performance evaluation of the battery pack.

[0067] The process of interval minimum divide-and-conquer calculation is as follows: take the value of [α1,α2,...,α in the interval [T,T+n] seconds of step A corresponding to the value of [α1,α2,...,α in the interval [Y,Y+n] columns. i ]Minimum value α Min . Where T represents time and Y represents the number of columns, such as Figure 3 As shown, through the divide-and-conquer calculation of the minimum value between configurations, the minimum value of the pressure difference indicator during the charging and discharging process is used for data extraction and automatic analysis, which intuitively displays the differences between the data and provides data support for the pressure difference performance evaluation of the battery pack.

[0068] The process of calculating the interval extreme value sum is as follows: take the interval [T,T+n] seconds of step A and the interval [α1,α2,..,α i ] Maximum value α Max and the minimum value α Min , do α Max +α Min Sum calculation. (T represents time, Y represents the number of columns). Figure 2As shown, the value corresponding to the voltage in the first column in the last second of step 11 is α1, and the value corresponding to the voltage in the second column is α2. The value corresponding to the voltage in the first column in the first second of step 12 is α3, and the value corresponding to the voltage in the second column is α4. The maximum value among [α1, α2, α3, α4] is α Max , take the minimum value among [α1, α2, α3, α4] as α Min , α Max +α Min The value of is the interval extreme value and.

[0069] The process of calculating the interval extreme value difference is as follows: take the interval [T,T+n] seconds of step A corresponding to the interval [Y,Y+n] columns [α1,α2,..,α i ] Maximum value α Max and the minimum value α Min , do α Max -α Min Calculate the difference. (T represents time, Y represents the number of columns). Figure 2 As shown, the value corresponding to the voltage in the first column in the last second of step 11 is α1, and the value corresponding to the voltage in the second column is α2. The value corresponding to the voltage in the first column in the first second of step 12 is α3, and the value corresponding to the voltage in the second column is α4. The maximum value among [α1, α2, α3, α4] is α Max , take the minimum value among [α1, α2, α3, α4] as α Min , α Max -α Min The value of is the interval extreme difference.

[0070] Through the algorithm configuration of interval extreme value summation and interval extreme value difference calculation, the measured voltage data is extracted, and the actual pressure difference data is extracted from the interval extreme value calculation, providing data support for the pressure difference performance evaluation of the battery pack.

[0071] The process of calculating the lateral derivative difference is as follows: take the value of the [i,j,...,k,...] column at the Tth second of step A and calculate the derivative value X Max =Max[(T ai -T aj -T ak -...),(T bi -T bj -T bk -..),(T ci -T cj -T ck -..)...], derived value X Min =Min[(T ai -T aj -T ak -...),(T bi -T bj -T bk-..),(T ci -T cj -T ck -..)...], T represents time, subscripts a, b, c... represent row numbers, and subscripts i, j, k... represent column numbers. Figure 4 As shown, the value of column i in the first second of step 12 is A1, the value of column j is B2, and a column is expanded to |A1-B2|=α1. The value of column A in the nth second of step 12 is An, the value of column B is Bn, and a column is expanded to |An-Bn|=αn. All values ​​of the difference column [α1,..,αn] are derived, and the maximum value Max[α1,...,αn] or Min[α1,...,αn] is taken. Figure 5 As shown, the current accuracy (Hall charge and discharge current, shunt charge and discharge current) is automatically analyzed through lateral derivative difference calculation, and the quality control of the current detection items is carried out according to the process threshold.

[0072] The process of calculating the longitudinal derivative difference is as follows: take the value of the [A1, A2, ..., An] column in any process step and calculate the maximum value Y of the derivative difference Max =Max[(A1a-A1b-A1c...),(A2a-A2b-A2c...),...,(Ana-Anb-Anc...)], Y Min =Min[(A1a-A1b-A1c...),(A2a-A2b-A2c...),...,(Ana-Anb-Anc...)], a, b, c... represent the row numbers, such as Figure 6 As shown, the value of the ath row of column A1 in step H is A1a, and the value of the bth row is A1b, and a behavior |A1a-A1b|=Y1 is extended. The value of the ath row of column An in step H is Ana, and the value of the bth row is Anb, and a behavior |Ana-Anb|=Yn is extended. All values ​​of the derivative difference column [Y1,...,Yn] are taken, and then the maximum value Max[Y1,...,Yn] or the minimum value Min[Y1,...,Yn] is taken.

[0073] The process of slice difference quotient calculation is: take the value of any process step at the Tth second as A, the value of the Tth + nth (n>0)th second as B, slice difference quotient value X = (AB) / C (C is a dynamic value), extract the X value for display and make a chart. Figure 7 As shown, by configuring the slice difference-quotient algorithm, the ICC is automatically analyzed and the quality of the current detection items is controlled according to the process threshold.

[0074] The process of calculating the row threshold positioning difference is: Min{|Xi-C|}, C is a dynamic variable, Xi is the value in a certain column under a certain process step, and the row where the value Xi reaches 2.85V or is closest to 2.85V under the "lowest single cell voltage" column is located is selected, and the "single cell voltage difference" column of the row is captured. Figure 8 As shown, by configuring the algorithm, all data in any process step can be compared with the standard, and the pressure difference of the battery pack can be determined according to the process regulations to determine whether it meets the specifications, thereby solving quality problems caused by non-compliant pressure differences and large differences in battery cells.

[0075] The calculation process of the inter-module voltage trend algorithm is as follows:

[0076] The average value of all battery cell values ​​in the module is calculated using the following formula:

[0077] Where n represents the number of modules, i represents the number and position of cells under the module, and X i Indicates the indicator value of the battery cell under this module, U n Indicates the average value of all battery cells in this module.

[0078] Then, the average values ​​of all the battery cell values ​​under adjacent modules are arranged and combined to obtain the difference and get the voltage trend between modules.

[0079] Permutations and combinations:

[0080]

[0081] When m=n:

[0082]

[0083] A represents the arrangement method, n represents the number of modules, and m represents the number of modules to be compared. C: Combination method: n represents the number of modules, and m represents the number of modules to be compared. For a battery pack with multiple modules, the cells in the scattered points are divided according to the number of cells bound to the modules, and the average value U of all the cell values ​​under each module is calculated. n , take the average value of adjacent modules, perform permutations and combinations to find the difference, such as Fig. 9 shown.

[0084] The calculation process of linear regression is:

[0085] y=a+bx

[0086] Among them, y represents the index value of the battery cell, x represents the position of the battery cell in the battery pack, b represents the slope, and a is the parameter.

[0087]

[0088] Where n represents the number of cells in the battery pack. Represents the average value of the number of cells. Indicates the average value of battery cell indicators.

[0089] Linear regression can be used to predict the index value of any battery cell, and the degree of dispersion between the index value and the target value can be displayed, so as to predict the change trend of the battery cell voltage, such as Fig.10 As shown, potential voltage anomalies, such as overcharging and over-discharging, can be discovered in time, so that measures can be taken in advance to intervene and avoid damage to battery performance caused by voltage anomalies.

[0090] The recursive algorithm can realize batch processing of files. The execution process of the algorithm includes:

[0091] (1) Design a recursive function: Write a recursive function that can traverse the specified folder and all its subfolders and collect all qualified files to be processed. For each folder, first check whether it contains subfolders. If so, recursively call the function for each subfolder and then check whether the current folder contains the file to be processed. If so, add it to the file list.

[0092] (2) File filtering: During the recursive traversal process, the files to be processed are filtered based on their extensions, name patterns, or other features.

[0093] (3) Unified processing: After all the files to be processed are collected, the batch processing module processes these files one by one. The processing process includes reading the file content, applying the regular expression matching algorithm to determine the position of the header column, performing vertical segmentation operations, etc. The recursive algorithm can automatically retrieve all the files to be processed in the folder and perform unified processing, which greatly improves the analysis efficiency.

[0094] Before executing each algorithm, the present invention uses a regular expression to determine the table header position, and the specific process includes:

[0095] (1) Matching header features: Use regular expressions to match specific patterns or keywords in the header. For example, the header may contain specific tags, formats, or content. Regular expressions can accurately identify these features.

[0096] (2) Locating the header column: Once the features of the header are matched, the position of the header column in the table can be determined. By parsing the regular expression matching results, the start and end positions of the header column can be determined.

[0097] (3) Identify delimiters: Regular expressions identify delimiters in tables, such as blank lines, specific characters, or strings, which are used to identify the boundaries between different parts or rows.

[0098] (4) Define segmentation rules: Define segmentation rules based on the position and characteristics of the delimiter. For example, when encountering multiple consecutive blank lines or specific strings, the table is segmented into multiple small tables.

[0099] (5) Perform splitting operation: After matching the delimiter with a regular expression, perform splitting operation according to the splitting rule, traverse and split the table content, and generate multiple small table data.

[0100] (6) Data preprocessing: Preprocess the tabular data using regular expressions to remove irrelevant information and standardize the format, providing a clean and accurate data basis for subsequent statistical analysis.

[0101] (7) Automated processing: Split and preprocess table data using regular expressions.

[0102] (8) After splitting the large table into multiple small tables, more detailed and in-depth analysis of different small tables can help discover patterns and trends in the data and improve the accuracy and efficiency of statistical analysis.

[0103] The regular expression matching algorithm is used to automatically determine the position information of the header column and implement accurate intelligent automatic vertical segmentation, splitting the large table into multiple small table data, thereby greatly improving the efficiency of subsequent statistical analysis.

[0104] The electrical performance algorithm model library is configured to match the electrical performance detection items to the corresponding algorithms from the standard algorithm library based on historical electrical measurement data and electrical measurement production lines, processes, and equipment, including adding modules, editing modules, deleting modules, and page display area modules, such as Fig.11 As shown, in the newly added module, you can set the project name, equipment manufacturer, electrical test step, file name, line number, test item, title name, algorithm, unit conversion, title name, check item order, advanced mode and other configuration items to configure the module and algorithm for the newly added electrical performance test items. The specific method is as follows:

[0105] Project name, equipment manufacturer, electrical measurement step selection test item configuration; File name: step layer electrical measurement file is the step layer, record layer electrical measurement file is the record layer; Row number: configure the data position of the record layer electrical measurement parameters, fill in "1" in the first row and "-1" in the last row; Test item: configure standard test items and determine the upper and lower limit range values; Title name: electrical performance analysis parameter name; Algorithm: electrical performance analysis source file containing multiple parameter values, configure the standard algorithm library model; Unit conversion: if the value parameter unit and the later formed electrical performance analysis parameter unit are inconsistent, perform unit conversion; Check item order: the order of electrical performance analysis parameters in the subsequent electrical performance analysis; Advanced mode: configure when formulating multi-column and multi-row data calculations.

[0106] The chart analysis algorithm model library is configured to match different analysis chart types for electrical performance test items based on historical electrical measurement data and electrical measurement production lines, processes, and equipment, including add modules, add-special modules, edit modules, delete modules, and page display area modules, such as Fig.12 As shown, in the newly added module, set configuration items such as project name, equipment manufacturer, electrical measurement step, step number, chart title, chart type, lowest and second lowest difference, chart order, column name, and configuration column information.

[0107] The electrical performance detection module is used to configure the data names to be extracted for different projects, equipment manufacturers, and electrical measurement steps, and select the corresponding chart type (scatter chart or line chart). The background automatically highlights the maximum, minimum, and second minimum values ​​of each chart, which is convenient for intuitive data display and analysis, and realizes online big data analysis of electrical measurement files, and finally obtains electrical performance parameter analysis and related discrete analysis charts. Figure 13-15 As shown, Fig.13 For the analysis of the scatter plot of the discharge static data, Fig.14 Analyze the line chart for DCR data. Fig.15 Scatter plot for charge and discharge dynamic voltage data analysis.

[0108] Data management module, used to record operation errors, warnings and other prompts;

[0109] The data analysis module is used to query the analysis results of historical electrical measurement data; query or export the project name, equipment manufacturer, electrical measurement date, electrical measurement process, judgment result, battery pack code, filter the data analyzed by the system, and display all bad data battery packs in yellow on the centralized interface, and display abnormal data of the detection items in red.

[0110] The data comparison module is used to compare the analysis results of historical electrical measurement data with offline manual statistical data, perform comparative analysis on battery packs of the same dimension, verify the accuracy of the data, and display summary information of system data.

[0111] The system management module includes functions such as part management, user management, and role management, which are used to authorize and group user accounts.

[0112] The present invention constructs a specific data structure and a matching efficient parsing algorithm in the intelligent algorithm library module. The standard algorithm library is configured to integrate multiple algorithm models, which can perform consistency analysis on different types of electrical measurement data. The electrical performance algorithm model library is configured to match the electrical performance detection items to the corresponding algorithms from the standard algorithm library according to historical electrical measurement data and electrical measurement production lines, processes, and equipment. It can process huge electrical measurement data files, solving the problems of huge data, time-consuming manual screening, high labor costs, and insufficient consistency accuracy in traditional methods. By integrating more than a dozen efficient algorithms, a unified statistical algorithm is constructed for each small table data (the same group), which can process each small table in a loop and calculate the maximum and minimum values ​​in the vertical direction. The concept of horizontal difference calculation is introduced. By calculating the differences of all columns, the maximum and minimum values ​​of the horizontal differences are calculated, achieving more comprehensive and in-depth data analysis.

[0113] This system shortens the data analysis time of a single electrical test file from 12 minutes of manual operation to 7.2 seconds of automatic system processing, significantly improving efficiency by up to 100 times. At the same time, the system has also achieved a high level of 99.5% in terms of analysis accuracy, ensuring the accuracy and reliability of detection and analysis.

[0114] Through system algorithm configuration and timed file scanning, automated analysis is performed, the statistical maximum and minimum values ​​are compared with the preset thresholds, and the data that exceeds the threshold range is automatically colored and highlighted, making abnormal data clear at a glance, and automatically intercepting bad battery packs according to the preset process standard thresholds, which is convenient for users to quickly identify and process. At the same time, the data is uploaded to the MES system immediately after the automated data analysis is completed, and unqualified products are intercepted and controlled in the next process, effectively preventing bad products from entering the market and ensuring product quality and customer satisfaction. Through this series of technical innovations and optimizations, the present invention not only improves the efficiency and accuracy of battery pack electrical performance detection and analysis, but also makes important contributions to the improvement of the intelligence and automation level of the battery pack manufacturing industry.

[0115] Continue reading Figure 1 The basic service layer of the intelligent detection system for battery pack performance big data consistency of the present invention adopts distributed high-speed cache, distributed message middleware, and WCF service. The distributed high-speed cache adopts Redis non-relational memory database storage. Redis mainly uses memory to store data. For cached data that is frequently accessed (such as statistical results, intermediate calculation results, etc.), Redis non-relational memory database storage is adopted, so that the system has a very fast reading and writing speed, improves data access speed, reduces database load, and improves the overall performance of the system.

[0116] The distributed message middleware uses RabbitMQ distributed message middleware to realize the asynchronous execution of tasks such as real-time processing of data streams, concurrent processing, batch processing of data, and complex calculations.

[0117] WCF services are used to implement communication and data sharing between different services, such as data interaction between the front-end and the back-end, service calls between different modules, and asynchronous communication between the front-end and the back-end, improving the response speed and interactive experience of the user interface.

[0118] Example 2

[0119] like Fig.16 As shown, this embodiment provides an intelligent detection method based on battery pack electrical performance big data consistency, including the following steps:

[0120] Step 1: After the battery pack test is completed, the electric test file is automatically generated, and the data access module scans the electric test file at set intervals to obtain the electric test file data;

[0121] On-site personnel scan the code to enter the battery pack code and other information. After the test is completed, the CSV electrical measurement file is automatically generated to the specified path. The data access module performs automatic file scanning every 30 minutes to obtain the electrical measurement file data.

[0122] Step 1 specifically includes the following process:

[0123] 1.1. Introduce the OpenXML library and use the C# file I / O function to open and read the electrical test file. When reading the content of the electrical test file, use the SpreadsheetDocument class in the OpenXML library to read the csv file content, open the file through the SpreadsheetDocument.Open method, and obtain components such as WorkbookPart and WorksheetPart.

[0124] 1.2. Traverse the Sheets element in WorkbookPart, obtain the names and IDs of all worksheets, select the worksheet to be parsed according to the user's selection or default settings, find the corresponding WorksheetPart through the worksheet ID, parse the SheetData element, which contains all the row and column information in the worksheet, traverse the Row element in SheetData and the Cell element in each Row, determine the position of each cell (row number and column number), and construct the two-dimensional table structure of the csv file.

[0125] 1.3. According to user needs, set rules or conditions for extracting cells. These rules are based on the value, format, position or other attributes of the cell. When traversing the two-dimensional table structure of the CSV file, perform logical judgment on each cell, determine which cells meet the extraction requirements according to the set rules, and mark them. According to the marked cell position, extract the content of the target cell from the two-dimensional table structure, perform necessary processing (such as removing spaces, converting data types, etc.), and save it to a custom data structure in a predetermined format.

[0126] Step 2: After configuring the electrical test items, test item ranges, test steps, and display chart types in the standard configuration module, the model library in the intelligent algorithm library module is called to perform real-time analysis on the electrical test file data;

[0127] Step 2 specifically includes the following processes:

[0128] 2.1. The user enters a unique process number on the interface to identify the currently processed CSV file or data set. Visual tools (such as scroll bars, row number input boxes, etc.) are provided. The user selects the table row area to be processed and locates the table row area. A date selector and time input box are provided. The user enters or selects a date range to filter rows that meet the time conditions. A drop-down list is provided to list all column names in the CSV file. The user can select one or more column names to determine the column area. Specific text filtering, value range filtering, and format filtering are also supported.

[0129] 2.2. Select an algorithm from the standard algorithm library in the electrical performance algorithm model library, combine the selected row area, column area and algorithm, and build a custom model data structure that meets your needs. The system provides visualization tools to help users complete the construction.

[0130] 2.3. The backend system receives the configuration information submitted by the user. Based on the configuration information, the backend system obtains the algorithm in the standard algorithm library and automatically maps and matches it. It extracts the table content based on the selected row and column areas and passes it to the algorithm for calculation. The system provides visualization functions and displays the calculation results to the user in the form of charts, reports, etc.

[0131] Step 3: The electrical performance detection module highlights the analysis results and sends them to emails, intercepting defective batteries whose electrical performance indicators exceed the process threshold.

[0132] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. Based on the big data consistency intelligent detection system of battery pack performance, the characteristics are: include: Data access module, used to extract historical electrical measurement data and configuration parameters, the historical electrical measurement data includes dynamic temperature difference, charge and discharge dynamic pressure difference, charge and discharge static pressure difference, charge and discharge capacity, charge and discharge energy, charge and discharge DCR, current accuracy, SOC, ICC, discharge temperature rise, process module temperature difference, and pre-measurement temperature; Standard configuration module, used to configure and manage electrical test items, test item ranges, test steps, and display chart types; An intelligent algorithm library module is used to adapt to the electrical performance detection data analysis of multiple types of products, equipment and electrical measurement files. The intelligent algorithm library module includes a standard algorithm library, an electrical performance algorithm model library, and a chart analysis algorithm model library. The standard algorithm library is configured to integrate multiple algorithm models. The electrical performance algorithm model library is configured to match the electrical performance detection items to corresponding algorithms from the standard algorithm library based on historical electrical measurement data and electrical measurement production lines, processes, and equipment. The chart analysis algorithm model library is configured to match different analysis chart types for electrical performance detection items based on historical electrical measurement data and electrical measurement production lines, processes, and equipment. The electrical performance detection module is used to configure the data names to be extracted for different projects, equipment manufacturers, and electrical measurement steps, and select the corresponding chart type. The background automatically highlights the maximum, minimum, and second minimum values ​​of each chart. The data analysis module is used to query the analysis results of historical electrical measurement data.

2. The intelligent detection system based on battery pack electrical performance big data consistency according to claim 1 is characterized by: The standard algorithm library includes slice difference algorithm, slice accumulation algorithm, data extraction, interval maximum value divide and conquer calculation, interval minimum value divide and conquer calculation, interval extreme value sum calculation, interval extreme value difference calculation, horizontal derivative difference calculation, vertical derivative difference calculation, slice difference quotient operation, row threshold positioning difference calculation, inter-module voltage trend algorithm, and linear regression.

3. The intelligent detection system based on battery pack electrical performance big data consistency according to claim 2 is characterized by: The calculation process of the slice difference algorithm is: take the value of step A at the Tth second as X AT and the value of the T+n(n>0) second is X A(T+n) , respectively do the difference calculation X i =X AT -X A(T+n) , extract X i The value is displayed and graphed; the calculation process of the slice accumulation algorithm is: take the value of step A at the Tth second as X AT , take the value of the Tth second of the A+n (n≥0) step as X (A+n)T , do cumulative calculations Extract X i The values ​​are displayed and graphed.

4. The intelligent detection system based on battery pack electrical performance big data consistency according to claim 2 is characterized by: The process of dividing and conquering the interval maximum value is as follows: take the value of [α1, α2, ..., α2] in the interval [T, T+n] seconds of the A step corresponding to the interval [Y, Y+n] columns. i ] Maximum value α Max The process of calculating the minimum value of the interval by dividing and conquering is as follows: take the value of [α1, α2, ..., α2 in the interval of [Y, Y+n] seconds corresponding to ... i ]Minimum value α Min , where f represents time and Y represents the number of columns.

5. The intelligent detection system based on battery pack electrical performance big data consistency according to claim 2 is characterized by: The process of calculating the interval extreme value sum is as follows: take the interval [T, T+n] seconds of the A step corresponding to the interval [Y, Y+n] columns [α1, α2, .., α i ] Maximum value α Max and the minimum value α Min , do α Max +α Min Sum calculation; the process of calculating the interval extreme value difference is: take the A step [T, T+n] seconds interval corresponding to the [Y, Y+n] column interval [α1, α2, .., α i ] Maximum value α Max and the minimum value α Min , do α Max -α Min Difference calculation, T represents time, Y represents the number of columns.

6. The intelligent detection system based on battery pack electrical performance big data consistency according to claim 2 is characterized by: The process of calculating the lateral derivative difference is as follows: take the value of the column [i, j, ..., k, ...] at the Tth second of step A, and calculate the derivative value X Max =Max[(T ai -T aj -T ak -...), (T bi -T bj -T bk -..), (T ci -T cj -T ck -..)...], derived value X Min =Min[(T ai -T aj -T ak -...), (T bi -T bj -T bk -..), (T ci -T cj -T ck -..)...], T represents time, subscripts a, b, c... represent row numbers, and subscripts i, j, k... represent column numbers; the process of calculating the longitudinal derivative difference is: take the value of the [A1, A2, ..., An]th column in any process step, and calculate the maximum value Y of the derivative value difference Max =Max[(A1a-A1b-A1c...), (A2a-A2b-A2c...),..., (Ana-Anb-Anc...)], Y Min =Min[(A1a-A1b-A1c...), (A2a-A2b-A2c...),..., (Ana-Anb-Anc...)].

7. The intelligent detection system based on battery pack electrical performance big data consistency according to claim 2 is characterized by: Also includes: Data management module, used to record operation errors and warning prompts; The data comparison module is used to compare the analysis results of historical electrical measurement data with offline manual statistical data, compare and analyze battery packs of the same dimension, verify the accuracy of the data, and display summary information of system data; The system management module includes part management, user management, and role management, which are used to authorize and group user accounts.

8. An intelligent detection method based on the consistency of battery pack performance big data, characterized by: The following steps are involved: Step 1: After the battery pack test is completed, the electric test file is automatically generated, and the data access module scans the electric test file at a set interval to obtain the electric test file data; Step 2: After configuring the electrical test items, test item ranges, test steps, and display chart types in the standard configuration module, the model library in the intelligent algorithm library module is called to perform real-time analysis on the electrical test file data; Step 3: The electrical performance detection module highlights the analysis results and sends them to emails, intercepting defective batteries whose electrical performance indicators exceed the process threshold.

9. The intelligent detection method based on the consistency of battery pack electrical performance big data according to claim 8 is characterized by: The step one comprises: 1.

1. Introduce the OpenXML library, use the C# file I / O function to open and read the electrical test file, and use the SpreadsheetDocument class in the OpenXML library to read the csV file content when reading the content of the electrical test file. Use the SpreadsheetDocument.Open method to open the file and obtain the WorkbookPart and WorksheetPart components. 1.

2. Traverse the Sheets element in WorkbookPart, obtain the names and IDs of all worksheets, select the worksheet to be parsed according to the user's selection or default settings, find the corresponding WorksheetPart by the worksheet ID, parse the SheetData element, which contains all the row and column information in the worksheet, traverse the Row element in SheetData and the Cell element in each Row, determine the row number and column number of each cell, and construct the two-dimensional table structure of the csv file; 1.

3. According to user needs, set the rules or conditions for extracting cells. When traversing the two-dimensional table structure of the CSV file, perform logical judgment on each cell, determine the cells that meet the extraction requirements according to the set rules, and mark them. According to the marked cell position, extract the content of the target cell from the two-dimensional table structure, remove spaces and convert data types, and save them to the custom data structure according to the predetermined format.

10. The intelligent detection method based on battery pack electrical performance big data consistency according to claim 8 is characterized by: The second step comprises: 2.

1. Enter a unique process number to identify the csv file or data set currently being processed. The user selects the table row area to be processed, locates the table row area, provides a date selector and a time input box, and the user enters or selects a date range to filter rows that meet the time conditions. A drop-down list is provided to list all column names in the csv file. The user can select one or more column names to determine the column area. 2.

2. Select an algorithm from the standard algorithm library in the electrical performance algorithm model library, combine the selected row area, column area and algorithm, and construct a custom model data structure that meets your needs; 2.

3. Based on the configuration information, the background system obtains the algorithm in the standard algorithm library and automatically maps and matches it. It extracts the table content based on the selected row and column areas and passes it to the algorithm for calculation. The calculation results are displayed to the user in the form of charts and reports.

Images