A method for processing battery test data
By calculating the parameter compensation values of the test channels and the overall average, the test data of lithium-ion batteries are corrected, which solves the measurement deviation problem between the test equipment and the channels and improves the accuracy of cell grading, grouping and anomaly screening.
Patent Information
- Application Number
- CN202611254976.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-18
- Publication Date
- 2026-11-10
AI Technical Summary
In existing technologies, there are systematic measurement deviations between different test channels of lithium-ion battery testing equipment, resulting in inaccurate cell grading, grouping, and anomaly screening results.
By acquiring reference data from multiple test channels, calculating the channel average and the overall average, determining the parameter compensation value, and correcting the measurement data of the cell to be processed, deviations between different test equipment and channels are eliminated.
It improves the consistency of voltage and internal resistance data, reduces misclassification of cells across grades and misjudgment of abnormal cells, and ensures the accuracy of cell grading, grouping, and abnormal screening.
Smart Images

Figure CN123066580A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of lithium-ion battery testing and data processing technology, and in particular to a battery testing data processing method. Background Technology
[0002] Lithium-ion batteries are widely used in power batteries and energy storage batteries. Battery modules typically consist of multiple individual cells. Due to performance differences among individual cells during manufacturing, their capacity, voltage, internal resistance, and self-discharge characteristics affect the consistency, capacity utilization, and cycle life of the battery module. Therefore, during cell production, it is usually necessary to classify, group, and screen for anomalies based on electrical performance test data.
[0003] Chinese invention patent application CN111786035A discloses a method for grouping lithium-ion batteries. This method first performs capacity screening and grading, then tests the open-circuit voltage and internal resistance of the cells, and combines voltage changes, internal resistance differences, and dynamic voltage drop before and after aging for further screening. Existing technologies already include methods that first screen and grade cells by capacity, then test their open-circuit voltage and internal resistance. Some methods also perform high-temperature and room-temperature aging on the cells, and further screen them based on voltage changes, internal resistance differences, and voltage drop parameters before and after aging. After completing static parameter screening, the cells can be connected in series and subjected to dynamic voltage drop consistency screening to improve the consistency of cell grouping.
[0004] Chinese invention patent application CN112993377A discloses a method for grouping lithium-ion power batteries. This method, based on capacity, voltage, and internal resistance screening, performs rapid charge-discharge on the cells and plots their current-voltage characteristic curves to select cells with similar dynamic characteristics for grouping. Other existing technologies, based on capacity, voltage, and internal resistance screening, also perform rapid charge-discharge on the cells and collect current and voltage data. By plotting the current-voltage characteristic curve of a single cell and comparing it with a standard current-voltage characteristic curve of the corresponding grade, cells with similar dynamic characteristics can be selected for grouping. This type of approach combines static electrical performance parameters with dynamic charge-discharge parameters to improve cell sorting and grouping results.
[0005] The existing solutions described above primarily target differences in cell capacity, voltage, internal resistance, self-discharge characteristics, and dynamic response. In actual production, to improve testing efficiency, battery testing equipment typically has multiple test channels, and multiple testing devices may jointly test the same batch of cells. Different testing devices, as well as different test channels within the same testing device, can be affected by probe contact status, circuit connection status, equipment calibration status, and inherent channel offsets. Even if the tested cell is within the normal range, the voltage or internal resistance data output by different test channels may still be generally higher or lower.
[0006] When unprocessed raw measurement data is used directly for cell grading and anomaly screening, systematic deviations in the test channels may be mistaken for individual cell performance differences. This can cause cells from the same test channel to be clustered into specific grades, or cause normal cells to be judged as abnormal, or prevent actual abnormal cells from being effectively identified. Existing capacity grading, static parameter screening, and dynamic performance screening methods cannot directly eliminate data offsets between different test devices or between different test channels within the same test device.
[0007] Therefore, how to reduce the systematic measurement deviation between different testing equipment and between different testing channels within the same testing equipment, improve the consistency of electrical performance data such as voltage and internal resistance, and reduce the impact of this deviation on cell grading, grouping, and abnormal screening results has become an urgent technical problem to be solved. Summary of the Invention
[0008] The main objective of this invention is to provide a battery test data processing method, which aims to reduce systematic measurement deviations between different test devices and between different test channels within the same test device, improve the consistency of electrical performance data such as voltage and internal resistance, and reduce the impact of such deviations on cell grading, grouping, and anomaly screening results.
[0009] To achieve the above objectives, the present invention proposes a battery test data processing method, comprising: Obtain reference data selected from at least one electrical performance parameter between voltage and internal resistance, obtained from testing multiple normal and qualified cells through multiple test channels respectively, with each piece of reference data corresponding to the test channel identifier actually performed; For the i-th test channel, calculate the channel mean of the reference data corresponding to that test channel. The overall mean G is determined based on all the reference data or the channel mean of each test channel. The overall mean G and the mean of each channel are used as the basis for comparison. The difference determines the parameter compensation value for each test channel. ,in, i is the index of the test channel; The original electrical performance parameters of the battery cell to be processed are obtained by measuring the cell through any test channel. And the corresponding test channel identifier, each original measurement data corresponds to the test channel identifier of the actual measurement; The original electrical performance parameter values Compensation value determined by the test channel identifier Add them together to obtain the corrected electrical performance parameter value. ,in, ; Based on the corrected electrical performance parameter values, the cells to be processed are subjected to at least one of the following processes: grading, grouping, and anomaly screening.
[0010] Preferably, the reference data is the electrical performance parameter data of the entire batch of normal and qualified battery cells in the previous batch, or the electrical performance parameter data of the first 1,000 normal and qualified battery cells in this batch.
[0011] Preferably, the channel average is calculated using reference data of normal, qualified battery cells tested by the same test channel within a single batch as the statistical unit.
[0012] Preferably, the overall mean is the arithmetic mean of all the reference data, or the arithmetic mean of the channel means of each test channel.
[0013] Preferably, the electrical performance parameters include voltage and internal resistance, and the average voltage value of each test channel is calculated separately. and the mean of internal resistance channels The voltage compensation value for each test channel was determined separately. and internal resistance compensation value .
[0014] Preferably, the first The voltage compensation value and internal resistance compensation value of the test channel are respectively and ,in, and These are the overall average voltage and the overall average internal resistance, respectively; the original voltage value measured by the i-th test channel. and the original internal resistance value According to and Perform correction. and These are the correction voltage value and the correction internal resistance value, respectively.
[0015] Preferably, the multiple test channels are set on a single test device or belong to multiple test devices, and the test channel index is used by the test channel identifier or a combination of the test device identifier and the test channel identifier, and the channel mean and parameter compensation value of each test channel are calculated respectively.
[0016] Preferably, the single or multiple test devices are open-circuit voltage (OCV) test devices used to measure cell voltage and internal resistance.
[0017] Preferably, the number of decimal places retained in the parameter compensation value ensures that its numerical resolution is not lower than the measurement resolution of the corresponding electrical performance parameter by the test equipment to which the test channel belongs.
[0018] Preferably, when performing anomaly screening on the cells to be processed, the average value is calculated based on the corrected internal resistance values of the entire batch of cells, cells in a single tray, or cells in a single storage location to which the cell to be processed belongs. and standard deviation and in accordance with Determine the internal resistance threshold , A positive dimensionless threshold coefficient; when the corrected internal resistance value of the cell to be processed is greater than the internal resistance threshold. When this happens, the battery cell to be processed is identified as an abnormal battery cell.
[0019] Preferably, the threshold coefficient n is determined based on the internal resistance data of multiple batches and the disassembly results of the corresponding battery cells.
[0020] Preferably, the threshold coefficient n is 3.
[0021] The above technical solution has the following advantages: This application uses reference data from normal, qualified battery cells as a basis, calculating the channel average for each test channel and the overall average for all test channels. The parameter compensation value for each test channel is then determined based on the difference between the overall average and the channel average. After the battery cell to be processed completes testing, the corresponding parameter compensation value is applied according to its actual test channel, reducing fixed data offsets between different testing devices and between different test channels within the same testing device. Using corrected voltage or internal resistance data for cell grading, grouping, and anomaly screening ensures that data obtained from different channels are at a more uniform comparison benchmark, reducing misclassification across grades, misjudgment of normal cells, and missed detection of abnormal cells caused by differences in test channels. This processing can be implemented after the existing electrical performance testing process without changing the original voltage and internal resistance measurement methods. Attached Figure Description
[0022] The present invention will now be described in detail with reference to specific embodiments and accompanying drawings, wherein: Figure 1 This is a flowchart of the battery test data processing method of this application. Detailed Implementation
[0023] The technical solution of this application is described below with reference to specific embodiments. These embodiments are used to illustrate the technical concept and implementation process of this application and are not intended to limit the scope of protection of this application. Data transmission, data storage, and program calling processes that can be directly understood by those skilled in the art can be implemented using conventional methods in the field.
[0024] Example 1 This embodiment provides a battery test data processing method. This method is applicable to battery testing equipment with multiple test channels, and also to a testing system composed of multiple testing devices and their test channels. The test objects are battery cells from the same production batch or adjacent production batches, and the processed electrical performance parameters include voltage and internal resistance.
[0025] This method is executed by the data processing unit of the testing equipment or a data processing system that interacts with the testing equipment. After completing the cell test, the testing equipment outputs the cell identifier, the testing equipment identifier, the test channel identifier, and the corresponding measured values of electrical performance parameters. The data processing system calls the corresponding parameter compensation value according to the test channel identifier to correct the original measured values. The corrected data is stored and used in at least one of the following processes: cell grading, grouping, and anomaly screening.
[0026] To ensure the comparability of reference data and data to be processed, the same test method, the same equipment range and sampling settings are used for the same electrical performance parameter within the same statistical period, and the cell model, state of charge, resting time and test environment temperature meet the consistent conditions specified in the production test specifications; the internal resistance is AC internal resistance or DC internal resistance, and when using AC internal resistance testing, the same test frequency is used for each test channel.
[0027] like Figure 1 As shown, Figure 1 The document outlines two types of reference data sources and the processing flow for using correction data for grading, grouping, or anomaly screening.
[0028] Step S1: Obtain reference data for calculating channel compensation values.
[0029] In this embodiment, the electrical performance parameter data of the entire batch of normal and qualified battery cells from the previous batch is selected as reference data, and the parameter compensation value obtained based on this reference data is used for data correction of the next batch of battery cells. In another embodiment, data of battery cells that have been determined to be normal and qualified according to the existing qualification judgment items and standards in the production process can be extracted from the first 1000 battery cells in this batch as reference data, and the parameter compensation value obtained based on this reference data is used for data correction of the 1001st and subsequent battery cells in this batch.
[0030] The reference data includes at least the cell identifier, the test channel identifier, and at least one electrical performance parameter value selected from voltage and internal resistance. When processing data from multiple test devices, the reference data also includes the test device identifier, and the actual test channel to which each piece of reference data belongs is determined by a combination of the test device identifier and the test channel identifier.
[0031] For each test channel, reference data from normal, qualified cells measured in that channel are compiled into a corresponding channel data set. Data from different batches are statistically analyzed separately to avoid directly mixing performance differences between different production batches into the same channel mean. The reference sample used to calculate the channel mean for each test channel should belong to normal, qualified cells from the same statistical batch, and cells with specific performance ranges should not be fixedly assigned to specific test channels to avoid misrepresenting the performance differences of the cells themselves as differences in the test channels.
[0032] When a test channel lacks valid reference data or the number of valid reference samples is insufficient to form stable statistical results, the parameter compensation value for that test channel is not generated or called temporarily. Once the number of valid reference samples for that test channel meets the conditions for forming stable statistical results, the parameter compensation value for that test channel is generated and called during subsequent data processing. When changes occur in the equipment calibration status, probe or circuit status, test range, test frequency, or test environment that are sufficient to affect the measurement deviation, the reference data is re-established and the parameter compensation value is calculated.
[0033] Step S2: Calculate the channel mean and overall mean of each test channel.
[0034] For the Test channel, assuming the test channel has Reference data for a normal, qualified battery cell, the first One reference data is Then the first Channel mean of the test channel for:
[0035] in, For testing channel indexes, For reference data index within the channel, For the first The number of reference data points for the test channel. For the first The first test channel The reference data refers to electrical performance parameters such as voltage or internal resistance.
[0036] When processing voltage and internal resistance simultaneously, calculate the average voltage value for each test channel separately. and the mean of internal resistance channels Statistical results are established separately for voltage data and internal resistance data, and different types of electrical performance parameters are not mixed in the calculation.
[0037] The overall mean can be the arithmetic mean of all reference data. In another implementation, the channel mean of each test channel can be calculated first, and then the arithmetic mean of all channel means can be calculated. When calculating the overall mean from all reference data, let the number of test channels be K, and the overall mean G be:
[0038] When calculating the overall mean from the mean of each channel, the overall mean G is:
[0039] The overall mean is used as the common center value for data leveling of each test channel. , and Use the same units as the electrical performance parameters being processed; use the same units when processing voltage. When dealing with internal resistance, use .
[0040] Step S3: Calculate the parameter compensation value for each test channel.
[0041] The parameter compensation value for each test channel is obtained by subtracting the overall mean from the channel mean of each test channel. Parameter compensation value of test channel for:
[0042] When the When the channel mean of the test channel is lower than the overall mean, the parameter compensation value is... It is a positive value. When the first... When the channel mean of the test channel is higher than the overall mean, the parameter compensation value... The value is negative. Therefore, the measurement results of each test channel can be corrected to the same overall center value.
[0043] For voltage parameters, the first Voltage compensation value of test channel for:
[0044] in, This is the overall average voltage. For the first Average voltage across the test channel.
[0045] For the internal resistance parameter, the first Internal resistance compensation value of test channel for:
[0046] in, This is the overall average internal resistance. For the first The average internal resistance of the test channel.
[0047] Parameter compensation values are stored separately for each test device and test channel. When processing multiple test devices, even if different test devices have the same channel number, their parameter compensation values are determined separately by combining the test device identifier and the test channel identifier.
[0048] The parameter compensation values are retained with a certain number of decimal places to ensure that their numerical resolution is no less than that of the corresponding testing equipment for measuring the corresponding electrical performance parameters, in order to avoid rounding errors in the parameter compensation values affecting the calibration results. The parameter compensation values and their corresponding electrical performance parameters use the same units.
[0049] Step S4: Measure the cell to be processed and obtain the corresponding raw test data.
[0050] The battery cell to be processed is placed in the testing equipment, and its voltage and internal resistance are measured through the corresponding test channels. After the test is completed, the cell identification, testing equipment identification, test channel identification, original voltage value, and original internal resistance value are obtained.
[0051] Each piece of raw measurement data is associated with the test channel on which the measurement was actually performed. The data processing system retrieves the voltage compensation value and internal resistance compensation value corresponding to that test channel from the stored parameter compensation values based on the test equipment identifier and the test channel identifier.
[0052] Step S5: Perform channel compensation on the original electrical performance parameter values.
[0053] For passing the first The test channel measured the original electrical performance parameters of the battery cell to be processed. With the Parameter compensation values corresponding to the test channel Add them together to obtain the corrected electrical performance parameter value. ,Right now:
[0054] in, For the first The raw electrical performance parameter values obtained from the test channel For the first The parameter compensation values for the test channel. To correct the electrical performance parameter values.
[0055] For voltage and internal resistance, corrections are made according to the following formulas:
[0056]
[0057] in, and The first The original voltage and original internal resistance values obtained from the test channel and These are the correction voltage value and the correction internal resistance value, respectively.
[0058] During the compensation process, the parameter compensation value must be selected based on the actual test channel used by the battery cell to be processed. The parameter compensation values of different test channels are not interchangeable. Through this correspondence, the fixed offset of each test channel relative to the overall center value can be eliminated or reduced from the measurement results.
[0059] Taking the 24-channel open-circuit voltage (OCV) test data as an example, the average voltage value for each test channel is calculated based on the voltage data of a normal, qualified battery cell. The voltage compensation value for each channel can be positive or negative. For example, the average voltage value for channel CH1 is 3243.8557mV, and its voltage compensation value is 0.2651mV. The original voltage value measured for this channel is 3244.320mV, and the compensated voltage value is 3244.5851mV. The average voltage value for channel CH12 is 3244.6360mV, and its voltage compensation value is -0.5152mV. The original voltage value measured for this channel is 3245.090mV, and the compensated voltage value is 3244.5748mV.
[0060] After processing the data from 24 test channels, the standard deviation of the original voltage test data for a single cell decreased from 0.370mV to 0.266mV, and the range decreased from 1.550mV to 1.123mV. This result indicates that the data offset caused by different test channels is reduced, and the processed voltage data exhibits better consistency.
[0061] The same data processing procedure can be used for internal resistance data. In this embodiment, the 16 test channels of the two test devices are treated as independent device channel combinations. The reference data of the 16 device channel combinations are statistically analyzed as a whole to determine a unified overall mean internal resistance. The mean internal resistance of each device channel combination is then calculated separately. Based on the difference between the overall mean internal resistance and the mean internal resistance of each device channel combination, the internal resistance compensation value for each device channel combination is determined. After compensating the original internal resistance data, the standard deviation of the internal resistance data of the 16 test channels decreases from 0.00287mΩ to 0.00124mΩ, and the range decreases from 0.00910mΩ to 0.00459mΩ.
[0062] Step S6: Process the battery cell using the corrected electrical performance parameter values.
[0063] When classifying battery cells, the calibrated voltage or calibrated internal resistance value is used as the classification basis. The data processing system compares the calibrated electrical performance parameter values with the pre-determined classification range and assigns the battery cells to the corresponding class. Since the classification basis has been compensated by the test channel, it can reduce misclassification caused by an overall high or low value in a certain test channel.
[0064] When grouping battery cells, select cells of similar grade with similar calibrated electrical performance parameters for grouping. The electrical performance parameters used for grouping can include calibrated voltage and calibrated internal resistance values. By using calibrated data, the impact of differences in testing equipment and testing channels on cell consistency evaluation can be reduced.
[0065] When performing anomaly screening, anomaly screening thresholds can be established based on the corrected internal resistance values of the entire batch of cells, cells within a single tray, or cells within a single storage location to which the cell to be processed belongs. Let the mean corrected internal resistance of the corresponding dataset be... The standard deviation is Then the internal resistance threshold for:
[0066] in, A positive dimensionless threshold coefficient. This is determined based on internal resistance data from multiple batches and the disassembly results of the corresponding battery cells. When the corrected internal resistance value of the battery cell to be processed exceeds the internal resistance threshold... If this happens, the battery cell to be processed will be identified as an abnormal battery cell.
[0067] Determine the threshold coefficient At that time, it is possible to target candidates Calculate the screening thresholds for each value, correlate the screening results of each candidate threshold with the disassembly results, statistically analyze the cases of false positives for normal cells and false negatives for abnormal cells, and select those that meet production quality requirements. Value; when changes occur in the cell model, test conditions, or equipment status that affect the internal resistance distribution, re-verification is required. value.
[0068] In a specific implementation process, the threshold coefficient Take 3. Therefore, the internal resistance threshold is determined according to... Confirmed. A certain battery cell was tested through channel CH15 of the second testing equipment, with an initial internal resistance of 0.1713 mΩ. The internal resistance compensation value of this equipment's channel combination was -0.00318 mΩ, resulting in a corrected internal resistance of 0.16812 mΩ. Under the statistical caliber used in this embodiment, the internal resistance threshold is 0.16895 mΩ. The cell's initial internal resistance value was higher than the internal resistance threshold, while the corrected internal resistance value was lower than the internal resistance threshold. Therefore, based on the corrected internal resistance value, it was determined to be a normal battery cell. Subsequent disassembly results showed that the cell had no abnormalities, indicating that the initial abnormal result originated from data deviation between the testing equipment or testing channels.
[0069] This embodiment establishes a statistical benchmark for each test channel using batch data of normal, qualified battery cells, and generates a channel-specific compensation value based on the difference between the channel mean and the overall mean. After the battery cell to be processed completes testing, the corresponding compensation value is applied according to its actual test channel. Therefore, the data used for subsequent grading, grouping, and anomaly screening can more closely approximate a unified testing benchmark, reducing the impact of differences between different testing equipment and between different test channels within the same testing equipment on the battery cell determination results.
[0070] Example 2 This embodiment provides a battery test data processing method for establishing channel compensation values based on the normal and qualified cells from the first 1000 cells in a batch. The normal and qualified cells are determined according to existing qualification criteria and standards established during the production process. The channel compensation values established based on the data of the normal and qualified cells from the first 1000 cells are used for data correction of the 1001st cell and subsequent cells in the batch. This embodiment is applicable to situations where the testing equipment is relatively stable during continuous production of the same type of cell, and where it is necessary to use the batch data to promptly correct deviations in each test channel.
[0071] The battery cells are sequentially tested using an open-circuit voltage (OCV) testing device with multiple test channels. The OCV testing device measures the voltage and internal resistance of the battery cell and outputs the cell identifier, test channel identifier, original voltage value, and original internal resistance value. The data processing system extracts data from the first 1000 cells in this batch that have been determined to be normal and qualified according to the existing qualification criteria and standards in the production process, and establishes reference data sets for each test channel.
[0072] For any given test channel, the voltage data measured in the first 1000 normal and qualified cells of that test channel is used as the voltage reference data, and the internal resistance data measured in that test channel is used as the internal resistance reference data. Reference data from different test channels are statistically analyzed separately; data with different channel identifiers are not merged into the same channel dataset. This embodiment uses the same sample comparability, test condition control, and effective reference sample quantity requirements as Embodiment 1; parameter compensation values for each test channel are generated only after the number of effective reference samples for each test channel meets the conditions for forming stable statistical results, and are then used in subsequent data processing for this batch.
[0073] The data processing system calculates the average voltage value of each test channel separately. and the mean of internal resistance channels .
[0074] This embodiment uses the channel average of each test channel to calculate the overall average. Assume the OCV testing equipment has... Each test channel has an overall average voltage. for:
[0075] Overall average internal resistance for:
[0076] Overall voltage average and the overall mean of internal resistance Serves as the data adjustment center for different test channels. The voltage compensation value for the test channel is:
[0077] No. The internal resistance compensation value of the test channel is:
[0078] After calculation, the test channel identifier, voltage compensation value, and internal resistance compensation value are stored together. The number of decimal places retained for the parameter compensation values ensures that their numerical resolution is no less than that of the OCV testing equipment for measuring the corresponding parameters.
[0079] Starting with the 1001st cell in this batch, the data processing system retrieves the corresponding voltage compensation value and internal resistance compensation value based on the test channel identifier of each cell. The original voltage value measured by the test channel is The original internal resistance value is Then the correction voltage value and the correction internal resistance value are respectively:
[0080]
[0081] The calibration voltage and calibration internal resistance values are stored in association with the cell identifier and used as input data for subsequent grading, grouping, or anomaly screening.
[0082] Using the first 1000 normal and qualified cells in this batch as a reference sample allows the channel compensation values to correspond to the current production batch and the current equipment status. This approach does not require changing the original voltage and internal resistance testing process; instead, it adds channel mean calculation and corresponding compensation processes after the test data is output.
[0083] Example 3 This embodiment provides a method for processing battery test data between multiple test devices and between different test channels within each test device. The difference from Embodiment 1 is that this embodiment uses a combination of test device identifier and test channel identifier as the index for parameter compensation values.
[0084] This embodiment uses two internal resistance testing devices. The first testing device includes channels CH1 to CH8, and the second testing device includes channels CH9 to CH16. Each device and each testing channel performs internal resistance testing on normal, qualified battery cells.
[0085] The data processing system calculates the average internal resistance channel value on a per-device and per-test-channel basis. For example, channel CH1 of test device 1 corresponds to one average internal resistance channel value, and channel CH9 of test device 2 corresponds to another average internal resistance channel value. Even if different devices use the same channel number, the corresponding statistical data and parameter compensation values are determined separately based on the device identifier.
[0086] In this embodiment, the average internal resistance of all device channel combinations is 0.16211 mΩ. The average internal resistance of the first test device channel CH1 is 0.15875 mΩ, and the corresponding internal resistance compensation value is 0.00336 mΩ. The original internal resistance of a certain cell measured by this channel is 0.1564 mΩ, therefore the corrected internal resistance value of this cell is:
[0087] The average internal resistance of channel CH16 in the second test device is 0.16420 mΩ, corresponding to an internal resistance compensation value of -0.00209 mΩ. This channel measures the original internal resistance of a certain battery cell to be 0.1621 mΩ. Therefore, the corrected internal resistance of this battery cell is:
[0088] Each equipment channel combination determines its internal resistance compensation value in the same way and is calibrated according to the actual equipment and test channel used for the battery cell to be processed.
[0089] In this set of data, the standard deviation of the original internal resistance test data for the 16 test channels was 0.00287 mΩ, and the range was 0.00910 mΩ. After compensation was performed according to the equipment channel combinations, the standard deviation of the corrected internal resistance data was 0.00124 mΩ, and the range was 0.00459 mΩ.
[0090] Therefore, the data processing not only reduces the deviation between different channels within the same testing equipment, but also reduces the overall deviation between different testing equipment. The processed internal resistance data uses a unified overall mean as the adjustment center, which facilitates subsequent processing of cells measured by different equipment according to the same grading standards or anomaly screening standards.
[0091] Example 4 This embodiment provides a method for screening abnormal battery cells based on corrected internal resistance data.
[0092] First, the corrected internal resistance value of each cell to be processed is obtained according to the method described in Example 1 or Example 3. The dataset used for anomaly screening can be the corrected internal resistance value of the entire batch of cells to which the cell to be processed belongs, or it can be the corrected internal resistance value of cells in a single tray or a single storage location to which the cell to be processed belongs.
[0093] For the selected dataset, calculate the mean of the corrected internal resistance values. and standard deviation The internal resistance threshold is determined according to the following formula:
[0094] in, The internal resistance threshold, It is a positive dimensionless threshold coefficient.
[0095] Threshold coefficient The threshold coefficient is determined based on internal resistance data from multiple production batches and the corresponding cell disassembly results. By correlating the data screening results with the cell disassembly results, the adopted threshold coefficient can be matched with the actual anomaly detection results. Candidate The screening results can be compared based on the cases of false positives for normal cells and false negatives for abnormal cells, and the cells that meet the production quality requirements can be determined accordingly. value.
[0096] When screening for anomalies in the battery cell to be processed, the corrected internal resistance value of the cell is compared with the internal resistance threshold. Comparison is performed. When the corrected internal resistance value is greater than the internal resistance threshold... When the internal resistance value is not greater than the internal resistance threshold, the cell is identified as an abnormal cell. At that time, it is not determined as an abnormal cell based on this internal resistance condition.
[0097] In one specific embodiment of this example, the threshold coefficient n is taken as 3, and the internal resistance threshold is determined according to the following formula:
[0098] After processing a set of data obtained from 16 test channels of 2 devices, the mean value of the corrected internal resistance data was 0.16121mΩ, the standard deviation was 0.00258mΩ, and the corresponding internal resistance threshold was 0.16895mΩ.
[0099] In particular, channel CH15 of the second testing device measured the original internal resistance of a certain battery cell to be 0.1713mΩ. The internal resistance compensation value of this device channel combination is -0.00318mΩ, therefore the corrected internal resistance value of this battery cell is:
[0100] The original internal resistance of the battery cell was greater than the internal resistance threshold of 0.16895 mΩ. If the original internal resistance value were used directly for judgment, the cell would be considered abnormal. The calibrated internal resistance of the cell was 0.16812 mΩ, which is less than the internal resistance threshold of 0.16895 mΩ; therefore, based on the calibrated internal resistance value, it was not judged as abnormal. No abnormalities were found after disassembling the cell.
[0101] The results indicate that when the measurement results of a certain testing device or testing channel are generally too high, directly using the raw measurement data for screening may lead to the misclassification of normal cells. Establishing anomaly screening thresholds after channel compensation can reduce the impact of equipment and channel deviations on anomaly judgment results.
[0102] Example 5 This embodiment illustrates different implementation forms of electrical performance parameter types and overall mean calculation methods.
[0103] In one implementation, channel compensation is performed only on voltage data. The data processing system acquires voltage reference data for normal, qualified battery cells, calculates the average voltage value of each test channel and the overall average voltage value, and determines the voltage compensation value for each test channel. After the battery cell to be processed completes voltage testing, a corrected voltage value is calculated based on the actual voltage compensation value of the test channel. The corrected voltage value is used for battery cell voltage grading, grouping, or voltage anomaly screening.
[0104] In another implementation, channel compensation is performed only on the internal resistance data. The data processing system acquires reference internal resistance data for normal, qualified cells, calculates the channel average internal resistance and the overall average internal resistance for each test channel, and determines the internal resistance compensation value for each test channel. After the cells to be processed complete the internal resistance test, the corrected internal resistance value is calculated based on the actual internal resistance compensation value of the test channels. The corrected internal resistance value is used for cell internal resistance grading, grouping, or anomaly screening.
[0105] In another implementation, both voltage and internal resistance are compensated simultaneously. Voltage and internal resistance parameters are statistically analyzed and calculated separately to generate voltage compensation values and internal resistance compensation values, respectively. The data processing system does not use the voltage compensation values to process internal resistance data, nor does it use the internal resistance compensation values to process voltage data.
[0106] The overall mean can be directly calculated from reference data of all normal and qualified battery cells. When using this method, the sample size of each test channel has a corresponding weight in relation to the overall mean.
[0107] The overall mean can also be calculated from the channel mean of each test channel. When using this method, the channel mean of each test channel has the same weight in the calculation of the overall mean.
[0108] Regardless of which overall mean calculation method is used, the parameter compensation value for each test channel is determined by the difference between the overall mean and the corresponding channel mean. The original measurement values of the cells to be processed are corrected based on the parameter compensation values corresponding to their actual test channels.
[0109] Example 6 This embodiment illustrates the application of calibration data in the process of cell grading and grouping.
[0110] After the data processing system completes the voltage or internal resistance correction of the battery cell to be processed, it compares the corrected electrical performance parameter values with the corresponding grading standards. The grading standards can be voltage grading ranges or internal resistance grading ranges determined during the production process. The data processing system determines the battery cell grade based on the range in which the corrected electrical performance parameter values fall.
[0111] When the initial measurement results of the same test channel are generally high, uncompensated data may cause the cells tested in that channel to be concentrated in higher parameter grades. Conversely, when the initial measurement results of the same test channel are generally low, uncompensated data may cause the cells tested in that channel to be concentrated in lower parameter grades. Using calibrated electrical performance parameter values for grading can reduce the correlation between grade distribution and test channels.
[0112] When pairing battery cells, select cells from the same grade whose caliber correction voltage and internal resistance values meet the pairing requirements. Calibration data is used to reflect the parameter differences between different cells under a unified test standard, enabling comparisons of cells measured by different testing equipment or through different test channels based on the same data standard.
[0113] In this embodiment, the data processing steps are set after electrical performance parameter testing and before grading, grouping, or anomaly screening. The existing cell testing equipment continues to perform voltage and internal resistance measurements, and the data processing system adds channel compensation processing to the test results. Therefore, this method can be used for production data processing without adding extra electrical performance testing items.
[0114] The above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit the scope of protection of this application. Those skilled in the art, without departing from the technical concept of this application, can make equivalent substitutions or adaptive adjustments to the selection method of reference data, the calculation method of the overall average, and the subsequent data application method according to the actual testing equipment, the number of test channels, the type of electrical performance parameters, and production processing requirements. All equivalent changes made to the parameter compensation relationship, data correspondence relationship, or processing steps of each test channel based on the technical content disclosed in this application should fall within the scope of the technical concept of this application.
Claims
1. A battery test data processing method, characterized in that, include: Obtain reference data selected from at least one electrical performance parameter between voltage and internal resistance, obtained from testing multiple normal and qualified cells through multiple test channels respectively, with each piece of reference data corresponding to the test channel identifier actually performed; For the i-th test channel, calculate the channel mean of the reference data corresponding to that test channel. The overall mean G is determined based on all the reference data or the channel mean of each test channel. The overall mean G and the mean of each channel are used as the basis for comparison. The difference determines the parameter compensation value for each test channel. ,in, i is the index of the test channel; The original electrical performance parameters of the battery cell to be processed are obtained by measuring the cell through any test channel. And the corresponding test channel identifier, each original measurement data corresponds to the test channel identifier of the actual measurement; The original electrical performance parameter values Compensation value determined by the test channel identifier Add them together to obtain the corrected electrical performance parameter value. ,in, ; Based on the corrected electrical performance parameter values, the cells to be processed are subjected to at least one of the following processes: grading, grouping, and anomaly screening.
2. The battery test data processing method according to claim 1, characterized in that, The reference data refers to the electrical performance parameters of the entire batch of normal and qualified battery cells from the previous batch, or the electrical performance parameters of the first 1,000 normal and qualified battery cells in this batch.
3. The battery test data processing method according to claim 2, characterized in that, The channel average is calculated using reference data of normal, qualified cells tested by the same test channel within a single batch as the statistical unit.
4. The battery test data processing method according to claim 1, characterized in that, The overall mean is the arithmetic mean of all the reference data, or the arithmetic mean of the channel means of each test channel.
5. The battery test data processing method according to claim 1, characterized in that, The electrical performance parameters include voltage and internal resistance; the average voltage value of each test channel is calculated separately. and the mean of internal resistance channels The voltage compensation value for each test channel was determined separately. and internal resistance compensation value .
6. The battery test data processing method according to claim 5, characterized in that, The voltage compensation value and internal resistance compensation value of the i-th test channel are respectively and ,in, and These are the overall average voltage and the overall average internal resistance, respectively; the original voltage value measured by the i-th test channel. and the original internal resistance value According to and Perform correction. and These are the correction voltage value and the correction internal resistance value, respectively.
7. The battery test data processing method according to claim 1, characterized in that, The multiple test channels are set on a single test device or belong to multiple test devices, and are indexed by a test channel identifier or a combination of a test device identifier and a test channel identifier. The channel mean and parameter compensation value of each test channel are calculated respectively.
8. The battery test data processing method according to claim 7, characterized in that, The single or multiple test devices are open-circuit voltage (OCV) test devices used to measure cell voltage and internal resistance.
9. The battery test data processing method according to claim 1, characterized in that, The parameter compensation value retains a certain number of decimal places so that its numerical resolution is not lower than that of the test equipment to which the test channel belongs for measuring the corresponding electrical performance parameters.
10. The battery test data processing method according to claim 1, characterized in that, When performing anomaly screening on the cells to be processed, the average value is calculated based on the corrected internal resistance values of the entire batch of cells, cells in a single tray, or cells in a single storage location to which the cell to be processed belongs. and standard deviation and in accordance with Determine the internal resistance threshold , A positive dimensionless threshold coefficient; when the corrected internal resistance value of the cell to be processed is greater than the internal resistance threshold. When this happens, the battery cell to be processed is identified as an abnormal battery cell.
11. The battery test data processing method according to claim 10, characterized in that, The threshold coefficient n is determined based on the internal resistance data of multiple batches and the disassembly results of the corresponding battery cells.
12. The battery test data processing method according to claim 10, characterized in that, The threshold coefficient n is 3.
Citation Information
Patent Citations
Matching method of lithium ion batteries
CN111786035A
Grouping method of lithium ion power batteries
CN112993377A
