Ternary lithium battery pole piece performance test method and system fused with machine learning
By combining multi-point thickness testing and bidirectional prediction using machine learning models with confidence analysis, the problem of low reliability in traditional ternary lithium battery electrode performance test data has been solved, achieving comprehensive reliability and quantification of electrode performance evaluation.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- LONGNAN JINTAIGE COBALT IND CO LTD
- Filing Date
- 2025-09-12
- Publication Date
- 2026-06-23
AI Technical Summary
Traditional ternary lithium battery electrode performance testing relies on random single-point testing, resulting in low data reliability and limited reference value for the actual performance of the electrodes.
Multiple electrode thicknesses and discharge slopes are obtained through multi-point thickness testing. Bidirectional prediction is performed using a machine learning model, and confidence analysis is combined to calculate the thickness and discharge confidence, ultimately outputting the performance test results.
This improved the representativeness and credibility of the data, quantified the reliability of electrode performance evaluation, and provided a reliable quantitative basis for quality grading and process optimization.
Smart Images
Figure CN120802035B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of performance testing, and in particular to a method and system for testing the performance of ternary lithium battery electrodes that incorporates machine learning. Background Technology
[0002] The uniformity of electrode thickness and discharge performance of ternary lithium batteries are core key indicators that directly determine the battery's energy density, cycle life, and safety.
[0003] However, traditional ternary lithium battery electrode performance testing mostly relies on random single-point detection mode. Due to the strong randomness of the detection points and the limited amount of data, the reliability of the test data is very low, which in turn makes the test results not very valuable for the reference of the true performance of the electrode.
[0004] Therefore, there is an urgent need for a precise testing method for the performance of ternary lithium battery electrodes that can effectively avoid the limitations of random single-point testing. Summary of the Invention
[0005] This invention addresses the technical problem of low reliability in the performance test data of ternary lithium battery electrodes in the prior art by providing a ternary lithium battery electrode performance test method and system that integrates machine learning.
[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:
[0007] In a first aspect, the present invention provides a method for testing the performance of ternary lithium battery electrodes by incorporating machine learning, including:
[0008] Multiple thickness tests were conducted on the ternary lithium battery electrode to obtain multiple electrode thicknesses, and multiple discharge slopes of the ternary lithium battery electrode were also obtained.
[0009] Discharge slope is predicted based on multiple electrode thicknesses to obtain multiple predicted discharge slopes. Electrode thickness is then predicted based on these multiple discharge slopes to obtain multiple predicted electrode thicknesses.
[0010] The thickness confidence level is calculated by randomly combining the multiple electrode thicknesses and multiple predicted electrode thicknesses, and the discharge confidence level is calculated by randomly combining the multiple discharge slopes and multiple predicted discharge slopes, wherein the confidence level is calculated based on the number of tests and the prediction accuracy.
[0011] Based on the thickness confidence and discharge confidence, the performance test results of the ternary lithium battery electrode are calculated by combining multiple electrode thicknesses and multiple discharge slopes.
[0012] Secondly, this invention provides a ternary lithium battery electrode performance testing system integrating machine learning, comprising:
[0013] The thickness testing module is used to perform multi-point thickness testing on ternary lithium battery electrodes to obtain multiple electrode thicknesses and test and obtain multiple discharge slopes of the ternary lithium battery electrodes.
[0014] The parameter prediction module is used to predict the discharge slope based on multiple electrode thicknesses to obtain multiple predicted discharge slopes, and to predict the electrode thickness based on the multiple discharge slopes to obtain multiple predicted electrode thicknesses.
[0015] The confidence analysis module is used to randomly combine the multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate the thickness confidence, and to randomly combine the multiple discharge slopes and multiple predicted discharge slopes to calculate the discharge confidence, wherein the confidence is calculated based on the number of tests and the prediction accuracy.
[0016] The fusion output module is used to calculate the performance test results of the ternary lithium battery electrode based on the thickness confidence and discharge confidence, combined with multiple electrode thicknesses and multiple discharge slopes.
[0017] The beneficial effects of this invention are:
[0018] Compared to existing technologies, this application first conducts multi-point thickness tests on ternary lithium battery electrodes to obtain multiple electrode thicknesses and multiple discharge slopes of the ternary lithium battery electrodes. By collecting representative and stable data, a reliable foundation is provided for subsequent machine learning predictions and confidence analysis. Secondly, discharge slopes are predicted based on the multiple electrode thicknesses to obtain multiple predicted discharge slopes. Electrode thickness is then predicted based on the multiple discharge slopes to obtain multiple predicted electrode thicknesses. A machine learning model is used to achieve bidirectional prediction of electrode thickness → predicted discharge slope and discharge slope → predicted electrode thickness. The bidirectional mapping verifies data consistency and provides a prediction benchmark for subsequent confidence calculations. Furthermore, multiple electrode thicknesses and predicted electrode thicknesses are randomly combined to calculate the thickness confidence score, and multiple discharge slopes and predicted discharge slopes are randomly combined to calculate the discharge confidence score. This transforms the consistency between measured and predicted data into quantifiable reliability indicators, fully considering the impact of test scale and model reliability on the confidence score. The final output thickness and discharge confidence scores provide clear credibility labels for electrode performance test results, addressing the pain point of traditional tests lacking quantifiable reference value. Finally, based on the thickness and discharge confidence scores, combined with multiple electrode thicknesses and discharge slopes, the performance test results of ternary lithium battery electrodes are calculated. This quantifies the impact of thickness, discharge performance, and data reliability on ternary lithium battery electrode performance, making the performance evaluation both comprehensive and reliable. It provides a directly applicable quantitative basis for electrode quality grading and process optimization.
[0019] By conducting multi-point tests on ternary lithium battery electrodes to obtain multiple electrode thicknesses and corresponding discharge slopes, a machine learning model is used to achieve bidirectional prediction of electrode thickness → predicted discharge slope and vice versa, forming a cross-validation system between measured and predicted data. A basic confidence level is then calculated by randomly combining measured and predicted data, and double corrections are made based on the number of tests (data sufficiency) and prediction accuracy (model reliability) to accurately quantify thickness and discharge confidence levels. Finally, multi-dimensional indicators are integrated to calculate the comprehensive electrode performance test results. In this way, multi-point data acquisition and bidirectional prediction effectively avoid the random errors of traditional single-point testing, improve data representativeness and correlation, and enable traceable assessment of test result reliability through confidence quantification. This enhances the representativeness and credibility of ternary lithium battery electrode performance test data, providing high-precision, quantifiable, and highly reliable technical support for ternary lithium battery electrode quality control, production process optimization, and battery performance improvement. Attached Figure Description
[0020] Figure 1 A flowchart illustrating the performance testing method for ternary lithium battery electrodes incorporating machine learning provided by this invention.
[0021] Figure 2 A schematic diagram of the structure of the ternary lithium battery electrode performance testing system that integrates machine learning, provided by the present invention.
[0022] In the attached diagram, the components represented by each number are as follows:
[0023] Thickness testing module 11, parameter prediction module 12, confidence analysis module 13, and fusion output module 14. Detailed Implementation
[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0025] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.
[0026] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.
[0027] Example 1, as Figure 1 As shown, this embodiment of the invention provides a ternary lithium battery electrode performance testing method incorporating machine learning, including:
[0028] S10: Perform multi-point thickness testing on the ternary lithium battery electrode to obtain multiple electrode thicknesses, and test and obtain multiple discharge slopes of the ternary lithium battery electrode.
[0029] Traditional ternary lithium battery electrode performance testing often relies on random single-point thickness measurement. Due to the strong randomness of the measurement points and insufficient data sample size, it is prone to problems such as insufficient data representativeness and large fluctuations, ultimately resulting in low reliability and reference value of the test results.
[0030] To address the aforementioned issues, this application conducts multi-point thickness tests on ternary lithium battery electrodes to obtain multiple electrode thicknesses and tests to acquire multiple discharge slopes of the ternary lithium battery electrodes.
[0031] Specifically, step S10 in the method includes:
[0032] Get the number of tests;
[0033] Randomly select the number of test locations on the ternary lithium battery electrode to perform multi-point thickness testing and obtain multiple electrode thicknesses.
[0034] A constant current discharge test was performed on the ternary lithium battery to obtain the discharge curve at the mid-state of charge (SOC).
[0035] According to the number of tests, the discharge curve is divided, and the slopes of the divided discharge curve segments are obtained as multiple discharge slopes.
[0036] In this embodiment, the number of tests is first determined. Specifically, the number of tests refers to the number of locations randomly selected on the ternary lithium battery electrode for multi-point thickness testing, i.e., the number of sampling points for electrode thickness testing, such as 5 or 10. The number of tests directly affects the representativeness of the data. Too few tests are easily affected by single-point random errors (such as thickness abnormalities caused by local defects in the electrode), while too many tests increase testing costs and time. Therefore, a reasonable number of tests can be preset according to factors such as the actual size of the ternary lithium battery electrode and the stability of the manufacturing process to ensure that the number of tests can reflect the overall characteristics while also taking into account testing efficiency.
[0037] Secondly, a number of test locations are randomly selected on the ternary lithium battery electrode to perform multi-point thickness tests, obtaining multiple electrode thicknesses. The thickness tests can be conducted using a laser thickness gauge, micrometer, etc. For example, based on the number of tests (e.g., 5), 5 different locations are randomly selected on the ternary lithium battery electrode, and the thickness at each of the 5 locations is measured using a laser thickness gauge, obtaining 5 electrode thicknesses, such as 122μm, 119μm, 124μm, 118μm, and 123μm. By randomly selecting test locations, subjective bias from human selection can be avoided, ensuring that the sampling covers different areas of the electrode and accurately reflects the overall distribution characteristics of the ternary lithium battery electrode thickness. This approach effectively reduces the random errors of single-point testing.
[0038] Next, a constant current discharge test was performed on the ternary lithium battery to obtain the discharge curve at the state of charge (SOC). Specifically, constant current discharge testing refers to fixing the discharge current and testing the voltage change with capacity. This is to eliminate the interference of current fluctuations on voltage changes and ensure the stability and comparability of the discharge curve. SOC (State of Charge) is a core parameter for measuring the remaining capacity of a battery, usually expressed as a percentage, with 0% representing a complete discharge and 100% representing a complete charge. Its physical essence is the ratio of the battery's current remaining capacity to its rated capacity, directly reflecting the battery's usable energy state. The discharge curve of medium SOC refers to the discharge curve of 20% to 80% of the state of charge. The discharge curve of medium SOC is obtained because the voltage of ternary lithium batteries often exhibits nonlinear characteristics with capacity changes at high SOC (e.g., >80%) or low SOC (e.g., <20%). For example, the voltage plateau is gentle at high SOC, and the voltage drops rapidly at low SOC. The slope is unstable and greatly affected by the environment (e.g., temperature). In the medium SOC range, the relationship between battery voltage and remaining capacity shows a higher linearity, with good slope repeatability, which can more stably reflect the electrochemical performance of the battery and reduce the influence of accidental interference. For example, when performing a constant current discharge test on a ternary lithium battery, a discharge curve is plotted with discharge capacity as the horizontal axis and voltage as the vertical axis. The curve segment corresponding to the SOC range (20% to 80%) in the discharge curve is extracted as the discharge curve for the medium SOC range.
[0039] Finally, the discharge curves were divided according to the number of tests, and the slopes of the divided discharge curve segments were obtained as multiple discharge slopes. Specifically, the discharge curve at mid-SOC was evenly divided into n continuous segments according to the number of tests (n), and the rate of change of voltage with respect to capacity (i.e., slope) of each segment was calculated to obtain n discharge slopes. The discharge curves were divided because although the discharge curves in the mid-SOC range are generally stable, the slopes of different sub-segments may still have slight differences. Multiple discharge slopes can comprehensively reflect the changes in the battery's discharge performance in the mid-SOC range, avoiding the inability of a single slope to capture local characteristics and further reducing the randomness of the data.
[0040] In summary, compared to existing technologies, this application performs multi-point thickness testing on ternary lithium battery electrodes to obtain multiple electrode thicknesses and tests to acquire multiple discharge slopes of the ternary lithium battery electrodes. Thus, by collecting representative and stable data, a reliable foundation is provided for subsequent machine learning predictions and confidence analysis.
[0041] S20: Based on the thickness of multiple electrodes, the discharge slope is predicted to obtain multiple predicted discharge slopes. Based on the multiple discharge slopes, the electrode thickness is predicted to obtain multiple predicted electrode thicknesses.
[0042] In the performance testing of ternary lithium battery electrodes, there is a certain correlation between electrode thickness and discharge slope. Specifically, the electrode is the core reaction region for lithium-ion insertion / extraction, and its thickness directly determines the diffusion distance of lithium ions in the electrode material. A thicker electrode prolongs the ion migration distance, leading to a faster rate of voltage decrease with capacity during discharge, thus increasing the discharge slope. Conversely, a thinner electrode, due to its shorter diffusion path and lower resistance, results in a smoother discharge slope. Furthermore, excessively thick electrodes are prone to insufficient wetting or uneven conduction, reducing the effective reaction area and further exacerbating fluctuations in the discharge slope. A moderate thickness, on the other hand, improves the utilization rate of active materials and ensures the stability of the discharge curve. Therefore, based on this clear physical correlation, the discharge slope can be predicted by electrode thickness, and the electrode thickness can be inferred from the discharge slope. This bidirectional prediction allows for accurate identification and analysis of abnormal data.
[0043] To address the aforementioned issues, this application predicts the discharge slope based on multiple electrode thicknesses to obtain multiple predicted discharge slopes, and then predicts the electrode thickness based on these multiple discharge slopes to obtain multiple predicted electrode thicknesses.
[0044] Specifically, step S20 in the method includes:
[0045] Call the discharge slope predictor;
[0046] The thicknesses of the multiple electrodes are respectively input into the discharge slope predictor, and multiple predicted discharge slopes are obtained from the prediction output.
[0047] Call the electrode thickness predictor;
[0048] The multiple discharge slopes are input into the electrode thickness predictor, and multiple predicted electrode thicknesses are obtained from the prediction output.
[0049] In this embodiment, a discharge slope predictor based on machine learning is first invoked. The discharge slope predictor can predict the theoretical discharge slope based on the measured data of electrode thickness.
[0050] Secondly, the multiple electrode thicknesses obtained from actual testing are input into the discharge slope predictor, and multiple predicted discharge slopes are obtained from the prediction output. This is because the electrode thickness directly affects the ion diffusion path length and the utilization rate of active materials. By learning the correlation between electrode thickness and discharge slope in historical data, the discharge slope predictor can output a predicted discharge slope that conforms to the theoretical law.
[0051] Next, the electrode thickness predictor is invoked. This predictor is built based on machine learning, and its training process is the same as that of the discharge slope predictor. The electrode thickness predictor uses measured data of the discharge slope to infer the theoretical electrode thickness, forming a two-way verification loop with the discharge slope predictor. The cross-prediction of the two models can reduce the impact of errors from a single model.
[0052] Finally, the multiple discharge slopes obtained from the actual test are input into the electrode thickness predictor, and multiple predicted electrode thicknesses are obtained from the prediction output. By learning the correlation between discharge slope and electrode thickness in historical data, the electrode thickness predictor can infer the theoretical thickness that leads to this performance.
[0053] Thus, the measured electrode thickness is input into the discharge slope predictor, which predicts the theoretical discharge slope. The measured discharge slope is input into the electrode thickness predictor, which predicts the theoretical electrode thickness. By comparing the consistency between the measured electrode thickness and the predicted electrode thickness, and between the measured discharge slope and the predicted discharge slope, if the deviation is too large, it indicates that there may be random errors in the measured data, providing an anomaly identification basis for subsequent confidence calculation.
[0054] Furthermore, the "calling of the discharge slope predictor" includes:
[0055] Based on historical test data of ternary lithium battery electrodes, a set of sample electrode thicknesses was collected, along with the slope of the constant current discharge curve of ternary lithium batteries under different sample electrode thicknesses, and the set of sample discharge slopes was obtained by labeling.
[0056] Construct a machine learning-based discharge slope predictor;
[0057] The discharge slope predictor is trained under supervision using the sample electrode thickness set and the sample discharge slope set, and the training is completed after the test converges.
[0058] Configure the convergent discharge slope predictor on a cloud server.
[0059] In this embodiment, firstly, based on historical test data of ternary lithium battery electrodes, a set of sample electrode thicknesses is collected, along with the slope of the constant current discharge curve of the ternary lithium battery under different sample electrode thicknesses. A set of sample discharge slopes is then labeled to form a one-to-one corresponding sample pair of electrode thickness and constant current discharge curve slope. For example, if a sample electrode thickness is 120 μm and its slope on the SOC discharge curve is -0.005 V / mAh, it is labeled as (120 μm, -0.005 V / mAh).
[0060] Secondly, a machine learning-based discharge slope predictor is constructed. For example, since electrode thickness and discharge slope have a strong linear relationship, a neural network can be used to construct the discharge slope predictor, which mainly consists of an input layer, hidden layers, and an output layer architecture: the input layer receives electrode thickness data; the first hidden layer has 32 neurons that perform preliminary nonlinear transformation on the input features; the second hidden layer has 64 neurons that, with 20% Dropout, enhance the nonlinear fitting and suppress overfitting; the third hidden layer has 32 neurons that compress and integrate higher-order features, providing a precise mapping basis for the output layer; and finally, a single neuron in the output layer linearly activates and outputs the predicted discharge slope value.
[0061] Next, the discharge slope predictor is trained under supervision using a set of sample electrode thicknesses and a set of sample discharge slopes, and training is completed after test convergence. For example, the training process of the discharge slope predictor can be implemented through the following technical path: 1. Data preparation: The set of sample electrode thicknesses and the set of sample discharge slopes are divided into training set, validation set, and test set according to a ratio of 7:1.5:1.5. The training set is used for model parameter learning, the validation set is used for generalization ability evaluation, and the test set is used to test the model accuracy. Z-score standardization is performed on the sample electrode thickness data to eliminate dimensional differences. 2. Model Training: Using the sample electrode thickness in the training set as the input feature and the corresponding sample discharge slope as the supervision label, mean squared error (MSE) is used as the loss function to quantify the prediction bias. The model weight parameters are iteratively optimized using the Adam optimizer. At the same time, an early stopping strategy (training is terminated when the loss on the validation set does not decrease for 10 consecutive rounds) and L2 regularization are introduced to prevent overfitting. The performance on the validation set is monitored in real time during training. When the number of iterations reaches the preset maximum threshold (e.g., 500 rounds) or the prediction accuracy on the validation set is ≥95%, the model is considered to have converged, and the trained discharge slope predictor is obtained.
[0062] Finally, the converged discharge slope predictor is configured on a cloud server, allowing it to be remotely accessed via a network interface. This eliminates the need for repeated training on local devices, reducing hardware resource requirements. The cloud server can centrally manage the discharge slope predictor, and when new historical data is accumulated or prediction deviations are discovered, the model can be retrained and updated in the cloud, ensuring that the model is always optimized based on the latest data and maintains high accuracy.
[0063] In summary, compared to existing technologies, this application predicts discharge slopes based on multiple electrode thicknesses to obtain multiple predicted discharge slopes, and then predicts electrode thicknesses based on these multiple discharge slopes to obtain multiple predicted electrode thicknesses. Thus, a machine learning model is used to achieve cross-prediction of electrode thickness → discharge slope and discharge slope → electrode thickness. The bidirectional mapping verifies data consistency and provides a prediction benchmark for subsequent confidence level calculations.
[0064] S30: Randomly combine the multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate the thickness confidence level; randomly combine the multiple discharge slopes and multiple predicted discharge slopes to calculate the discharge confidence level, wherein the confidence level is calculated based on the number of tests and the prediction accuracy.
[0065] Traditional methods cannot quantify the reliability of test data, resulting in a lack of objective criteria for evaluating the credibility of test results. This can easily lead to misjudgments where the data appears consistent on the surface but lacks actual representativeness, and may even result in biased conclusions based on random error data. Therefore, it is necessary to construct a confidence quantification system to accurately quantify the reliability of test data and address the shortcomings of traditional reliability assessment methods.
[0066] To address the aforementioned issues, this application randomly combines the multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate a thickness confidence level, and randomly combines the multiple discharge slopes and multiple predicted discharge slopes to calculate a discharge confidence level, wherein the confidence level is calculated based on the number of tests and the prediction accuracy.
[0067] Specifically, step S30 in the method includes:
[0068] The multiple electrode thicknesses and multiple predicted electrode thicknesses are randomly combined to obtain multiple electrode thickness groups. The similarity and mean of each group are calculated to obtain the basic thickness confidence level.
[0069] The multiple discharge slopes and multiple predicted discharge slopes are randomly combined to obtain multiple discharge slope groups. The similarity and mean of each group are calculated to obtain the confidence level of the base slope.
[0070] Based on the number of tests and the prediction accuracy, the thickness confidence and discharge confidence are calculated.
[0071] In this embodiment, multiple electrode thicknesses obtained from actual measurements and multiple predicted electrode thicknesses output by the electrode thickness predictor are first randomly combined to obtain multiple electrode thickness groups. The similarity and mean of each group are then calculated to obtain the base thickness confidence level. This random combination dilutes the impact of single-point errors, more objectively reflecting the consistency of the overall data. For example, the similarity calculation can use a normalized value of cosine similarity or Euclidean distance, such as:
[0072] ;
[0073] The similarity value ranges from 0 to 1, where 1 indicates that the electrode thickness and the predicted electrode thickness are completely consistent, and 0 indicates that the electrode thickness and the predicted electrode thickness are completely unrelated. For example, if the measured electrode thickness of a ternary lithium battery electrode is 120 μm and the predicted electrode thickness by the electrode thickness predictor is 118 μm, then the similarity of this electrode thickness group is 1 − (|120-118|) / 120 = 0.983. In the same way, the similarity of multiple electrode thickness groups is calculated, and then the arithmetic mean is calculated. For example, 0.98 is used to obtain the basic thickness confidence. The basic thickness confidence reflects the overall consistency between the measured electrode thickness and the electrode thickness predicted by the electrode thickness predictor. The higher the basic thickness confidence, the stronger the overall consistency between the measured electrode thickness and the predicted electrode thickness, that is, the more reliable the data.
[0074] Secondly, multiple discharge slopes obtained from actual measurements and multiple predicted discharge slopes output by the discharge slope predictor are randomly combined to obtain multiple discharge slope groups. The similarity of each group is calculated, and the mean is also calculated to obtain the base slope confidence level. For example, the similarity is calculated using the same logic and method as the base thickness confidence level, for example:
[0075] ;
[0076] The similarity value ranges from 0 to 1, where 1 indicates that the discharge slope and the predicted discharge slope are completely consistent, and 0 indicates that the discharge slope and the predicted discharge slope are completely unrelated. The similarity of multiple discharge slope groups is calculated, and then the arithmetic mean is calculated. For example, 0.96 is used to obtain the basic slope confidence. The basic slope confidence reflects the overall consistency between the measured discharge slope and the discharge slope predicted by the discharge slope predictor. The higher the basic slope confidence, the stronger the overall consistency between the measured discharge slope and the predicted discharge slope, that is, the more reliable the data.
[0077] Finally, based on the number of tests and the prediction accuracy, thickness confidence and discharge confidence scores that better reflect the actual scenario are calculated. This is because the basic thickness confidence and basic discharge confidence scores only reflect the overall consistency between the measured electrode thickness, measured discharge slope and the predicted electrode thickness and predicted discharge slope output by the model, without considering the impact of test scale (number of tests) and model reliability (prediction accuracy) on the confidence scores: the more tests, the more comprehensive the coverage of measured electrode thickness and discharge slope data, the more thoroughly random errors are diluted, and the stronger the data representativeness; the higher the model prediction accuracy, the closer the output predicted electrode thickness and predicted discharge slope are to the true values, and the higher the reference value of the basic confidence scores. Therefore, it is necessary to quantify the impact of the number of tests and prediction accuracy to make the final output thickness confidence and discharge confidence scores better reflect the actual test scenario and improve the objectivity and reliability of the confidence score assessment.
[0078] Specifically, the phrase "calculating the thickness confidence level and discharge confidence level based on the number of tests and the prediction accuracy" includes:
[0079] Get the number of tests performed;
[0080] Calculate the ratio of the number of tests to the average number of tests to obtain the first confidence correction coefficient;
[0081] The accuracy rates of discharge slope prediction and electrode thickness prediction are obtained, and the average values are calculated to obtain the second confidence correction coefficient.
[0082] Based on the first confidence correction coefficient and the second confidence correction coefficient, the basic thickness confidence and the basic discharge confidence are corrected and calculated to obtain the thickness confidence and the discharge confidence.
[0083] In this embodiment of the application, the number of tests to be conducted is first obtained. For example, if the number of tests is 5, it means that there are 5 measured electrode thicknesses and 5 measured discharge slopes. The number of tests can reflect the representativeness of the data. Generally, the more tests conducted, the wider the coverage of the electrode area and the finer the slope segmentation. The higher the degree of dilution of random errors, the stronger the data reliability.
[0084] Secondly, the ratio of the number of tests to the average number of tests is calculated to obtain the first confidence level correction coefficient. The first confidence level correction coefficient = number of tests / average number of tests. The average number of tests refers to the historically accumulated average number of tests, which can be used as a benchmark to measure the sufficiency of the data in this test. If the first confidence level correction coefficient is greater than 1, it indicates that the number of tests in this test is greater than the historical average, the data coverage is more comprehensive, and random errors are diluted more effectively, thus positively improving the baseline confidence level. If the first confidence level correction coefficient is less than 1, it indicates that the number of tests in this test is less than the historical average, the data representativeness is insufficient, and the baseline confidence level needs to be lowered to reflect the lack of data sufficiency, thereby making the confidence level result more consistent with the actual sample size differences in the tests. For example, if the average number of tests is 6 and the number of tests in this test is 5, then the first confidence level correction coefficient = 5 / 6 = 0.83.
[0085] Next, the discharge slope prediction accuracy of the discharge slope predictor and the electrode thickness prediction accuracy of the electrode thickness predictor are obtained, and their averages are calculated to obtain the second confidence correction coefficient. For example, the prediction accuracy of the discharge slope predictor and the electrode thickness predictor are tested separately using an independent test set. For instance, if the prediction accuracy of the discharge slope predictor is 96% and the prediction accuracy of the electrode thickness predictor is 94%, then the second confidence correction coefficient = (96% + 94%) / 2 = 95%. The larger the second confidence correction coefficient, the higher the confidence level of the multiple predicted electrode thicknesses and predicted discharge slopes output by the discharge slope predictor and the electrode thickness predictor, and the higher the calculated base confidence level.
[0086] Finally, based on the first and second confidence level correction coefficients, the baseline thickness confidence and baseline discharge confidence are corrected and calculated to obtain the thickness confidence and discharge confidence. The thickness confidence = baseline thickness confidence * first confidence level correction coefficient * second confidence level correction coefficient, and the discharge confidence = baseline discharge confidence = first confidence level correction coefficient * second confidence level correction coefficient. The thickness and discharge confidence integrate three dimensions: data consistency, data sufficiency, and model reliability. Higher thickness and discharge confidence levels indicate more reliable measured data, which can be used as a basis for electrode performance evaluation. Low thickness and discharge confidence levels indicate unreliable measured data, requiring retesting or model optimization. For example, if the confidence level of the base thickness is 0.98, the confidence level of the base discharge is 0.96, the first confidence level correction factor is 0.83, and the second confidence level correction factor is 95%, then the confidence level of the thickness is 0.98 * 0.83 * 95% = 0.77, and the confidence level of the discharge is 0.96 * 0.83 * 95% = 0.76.
[0087] In summary, compared to existing technologies, this application randomly combines multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate a thickness confidence score, and randomly combines multiple discharge slopes and multiple predicted discharge slopes to calculate a discharge confidence score. The confidence score is calculated based on the number of tests and the prediction accuracy. This transforms the consistency between measured and predicted data into quantifiable reliability indicators, and fully considers the impact of test scale and model reliability on the confidence score. The final output thickness and discharge confidence scores provide clear credibility labels for electrode performance test results, solving the pain point of traditional tests lacking quantifiable quantifiable reference value.
[0088] S40: Based on the thickness confidence and discharge confidence, the performance test results of the ternary lithium battery electrode are calculated by combining multiple electrode thicknesses and multiple discharge slopes.
[0089] The physical properties (thickness), electrochemical performance (discharge performance), and data reliability of the electrode all affect the performance test results of ternary lithium battery electrodes. Therefore, it is necessary to integrate these three factors into the evaluation system through a systematic approach in order to obtain performance test results that comprehensively reflect the true quality level of the electrode.
[0090] To address the aforementioned issues, this application calculates the performance test results of ternary lithium battery electrodes based on the aforementioned thickness confidence and discharge confidence, combined with multiple electrode thicknesses and multiple discharge slopes.
[0091] Specifically, step S40 in the method includes:
[0092] The average similarity between the thickness of multiple electrodes and the thickness of a standard electrode is calculated, and the thickness performance parameters are obtained by combining the thickness confidence score.
[0093] The average similarity between multiple discharge slopes and the standard discharge slope is calculated, and the discharge performance parameters are obtained by combining the discharge confidence level.
[0094] The performance test results are calculated based on the thickness performance parameters and discharge performance parameters.
[0095] In this embodiment, the average similarity between multiple electrode thicknesses and the standard electrode thickness is first calculated. Combined with the thickness confidence level, a thickness performance parameter is calculated. The standard electrode thickness refers to the ideal electrode thickness that meets design requirements, such as 120μm±5μm for the positive electrode and 200μm±8μm for the negative electrode. This thickness is typically predetermined by process standards or performance targets and serves as a benchmark for evaluating whether the physical properties of the electrode meet the standards. The thickness performance parameter is calculated as: average similarity * thickness confidence level. For example, the similarity between multiple electrode thicknesses and the standard electrode thickness is calculated, and then the arithmetic mean of the multiple similarities is calculated. The similarity can be calculated using the following formula:
[0096] ;
[0097] For example, if the standard electrode thickness is 120 μm with an allowable deviation of ±5 μm, and five measured values are 122 μm, 119 μm, 124 μm, 118 μm, and 123 μm, then the similarities are 1−2 / 5=0.6, 1−1 / 5=0.8, 1−4 / 5=0.2, 1−2 / 5=0.6, and 1−3 / 5=0.4, respectively. The mean similarity is (0.6+0.8+0.2+0.6+0.4) / 5= If the thickness confidence level is 0.77, then the thickness performance parameter = 0.52 * 0.77 = 0.40. The thickness performance parameter reflects both the degree of agreement between the electrode thickness and the standard electrode thickness, and the reliability of the measured data is reflected through the thickness confidence level. Even if the average similarity between the electrode thickness and the standard electrode thickness is high, if the thickness confidence level is low (such as insufficient test quantity), the final thickness performance parameter will also be reduced, thus avoiding overestimating the value of unreliable data.
[0098] Secondly, the average similarity between multiple discharge slopes and the standard discharge slope is calculated. Combined with the discharge confidence score, the discharge performance parameters are obtained. The standard discharge slope refers to the ideal discharge slope that meets performance requirements, determined by battery design goals (such as energy density and cycle life), and serves as the benchmark for evaluating electrochemical performance. The discharge performance parameter = average similarity * discharge confidence score. For example, the average similarity between multiple discharge slopes and the standard discharge slope is calculated using the same logic and method as described above. For instance, if the standard discharge slope is -0.005V / mAh, with an allowable deviation of ±0.0005V / mAh, and the five measured discharge slopes are -0.0052V / mAh, -0.0049V / mAh, -0.0052V / mAh, -0.0048V / mAh, and -0.0051V / mAh, the calculated similarity scores are 1−0.0002 / 0.0005=0.6 and 1−0.0001 / 0. =0.8, 1−0.0002 / 0.0005=0.6, 1−0.0002 / 0.0005=0.6, 1−0.0001 / 0.0005=0.8, the mean similarity is (0.6+0.8+0.6+0.6+0.8) / 5=0.68. If the discharge confidence level is 0.76, then the discharge performance parameter is 0.68*0.76=0.52. The discharge performance parameter comprehensively reflects the degree of agreement between the electrode electrochemical performance and the standard value, as well as the reliability of the data, avoiding misjudgments caused by high similarity but low confidence (such as inaccurate model prediction).
[0099] Finally, the performance test results are calculated based on the thickness performance parameters and discharge performance parameters. The performance test result = w1 * thickness performance parameters + w2 * discharge performance parameters, where w1 and w2 are the weights of the thickness performance parameters and discharge performance parameters on the overall performance of the battery. w1 + w2 = 1. Those skilled in the art can dynamically set these parameters according to the actual situation, for example, w1 = 0.6 and w2 = 0.4. For example, if the thickness performance parameter is 0.40, the discharge performance parameter is 0.52, w1=0.6, and w2=0.4, then the performance test result at this time = 0.6*0.40+0.4*0.52=0.448. The performance test result is a quantitative value of 0% to 100%. The higher the value, the better the comprehensive performance of the electrode, which meets the physical standard (thickness) and has good electrochemical performance (discharge performance), and the test data is reliable. Based on this, a fixed threshold can be set to determine whether the performance of the ternary lithium battery electrode meets the standard. For example, the performance test result threshold can be set to 70%. Ternary lithium battery electrodes with performance results below the performance test result threshold do not meet the standard, and it is necessary to investigate the problem of uneven thickness or abnormal reaction.
[0100] In summary, compared to existing technologies, this application calculates the performance test results of ternary lithium battery electrodes based on the aforementioned thickness confidence and discharge confidence, combined with multiple electrode thicknesses and multiple discharge slopes. This quantifies the impact of thickness, discharge performance, and data reliability on the performance of ternary lithium battery electrodes, making the performance evaluation both comprehensive and reliable, and providing a directly applicable quantitative basis for electrode quality grading and process optimization.
[0101] In summary, the embodiments of this application have at least the following technical effects:
[0102] Compared to existing technologies, this application first conducts multi-point thickness tests on the ternary lithium battery electrode sheets to obtain multiple electrode thicknesses, and then tests and obtains multiple discharge slopes of the ternary lithium battery electrode sheets. In this way, by collecting representative and stable data, a reliable foundation is provided for subsequent machine learning predictions and confidence analysis.
[0103] Secondly, this application predicts the discharge slope based on multiple electrode thicknesses to obtain multiple predicted discharge slopes, and then predicts the electrode thickness based on these multiple discharge slopes to obtain multiple predicted electrode thicknesses. Thus, a machine learning model is used to achieve cross-prediction of electrode thickness → discharge slope and discharge slope → electrode thickness. This bidirectional mapping verifies data consistency and provides a prediction benchmark for subsequent confidence level calculations.
[0104] Furthermore, this application randomly combines the multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate the thickness confidence score, and randomly combines the multiple discharge slopes and multiple predicted discharge slopes to calculate the discharge confidence score. The confidence score is calculated based on the number of tests and the prediction accuracy. In this way, the consistency between measured and predicted data is transformed into a quantifiable reliability indicator, and the impact of test scale and model reliability on the confidence score is fully considered. The final output thickness and discharge confidence scores provide clear credibility labels for electrode performance test results, solving the pain point of traditional tests lacking quantifiable quantifiable reference value.
[0105] Finally, based on the aforementioned thickness confidence level and discharge confidence level, this application calculates the performance test results of the ternary lithium battery electrode by combining multiple electrode thicknesses and multiple discharge slopes. In this way, the impact of thickness, discharge performance, and data reliability on the performance of the ternary lithium battery electrode is quantified, making the performance evaluation both comprehensive and reliable, and providing a directly applicable quantitative basis for electrode quality grading and process optimization.
[0106] Through the above technical solution, this application obtains multiple electrode thicknesses and corresponding discharge slopes by conducting multi-point tests on ternary lithium battery electrodes. Based on a machine learning model, it achieves bidirectional prediction of electrode thickness → predicted discharge slope and discharge slope → predicted electrode thickness, forming a cross-validation system between measured and predicted data. Then, by randomly combining measured and predicted data, a basic confidence level is calculated, and dual corrections are made based on the number of tests (data sufficiency) and prediction accuracy (model reliability) to accurately quantify thickness and discharge confidence levels. Finally, multi-dimensional indicators are integrated to calculate the comprehensive performance test results of the electrode. Thus, multi-point data acquisition and bidirectional prediction effectively avoid the random errors of traditional single-point testing, improve data representativeness and correlation, and achieve traceable assessment of test result reliability through confidence quantification. This improves the representativeness and credibility of ternary lithium battery electrode performance test data, providing high-precision, quantifiable, and highly reliable technical support for ternary lithium battery electrode quality control, production process optimization, and battery performance improvement.
[0107] Example 2, as Figure 2 As shown, based on the same inventive concept as the ternary lithium battery electrode performance testing method integrating machine learning provided in Embodiment 1, this embodiment of the invention also provides a ternary lithium battery electrode performance testing system integrating machine learning, including:
[0108] Thickness testing module 11 is used to perform multi-point thickness testing on ternary lithium battery electrodes to obtain multiple electrode thicknesses and test and obtain multiple discharge slopes of ternary lithium battery electrodes.
[0109] The parameter prediction module 12 is used to predict the discharge slope based on multiple electrode thicknesses to obtain multiple predicted discharge slopes, and to predict the electrode thickness based on the multiple discharge slopes to obtain multiple predicted electrode thicknesses.
[0110] The confidence analysis module 13 is used to randomly combine the multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate the thickness confidence, and to randomly combine the multiple discharge slopes and multiple predicted discharge slopes to calculate the discharge confidence, wherein the confidence is calculated based on the number of tests and the prediction accuracy.
[0111] The fusion output module 14 is used to calculate the performance test results of the ternary lithium battery electrode based on the thickness confidence and discharge confidence, combined with multiple electrode thicknesses and multiple discharge slopes.
[0112] The thickness testing module 11 is specifically used for:
[0113] Get the number of tests;
[0114] Randomly select the number of test locations on the ternary lithium battery electrode to perform multi-point thickness testing and obtain multiple electrode thicknesses.
[0115] A constant current discharge test was performed on the ternary lithium battery to obtain the discharge curve at the mid-state of charge (SOC).
[0116] According to the number of tests, the discharge curve is divided, and the slopes of the divided discharge curve segments are obtained as multiple discharge slopes.
[0117] Specifically, the parameter prediction module 12 is used for:
[0118] Call the discharge slope predictor;
[0119] The thicknesses of the multiple electrodes are respectively input into the discharge slope predictor, and multiple predicted discharge slopes are obtained from the prediction output.
[0120] Call the electrode thickness predictor;
[0121] The multiple discharge slopes are input into the electrode thickness predictor, and multiple predicted electrode thicknesses are obtained from the prediction output.
[0122] Specifically, the "calling of the discharge slope predictor" includes:
[0123] Based on historical test data of ternary lithium battery electrodes, a set of sample electrode thicknesses was collected, along with the slope of the constant current discharge curve of ternary lithium batteries under different sample electrode thicknesses, and the set of sample discharge slopes was obtained by labeling.
[0124] Construct a machine learning-based discharge slope predictor;
[0125] The discharge slope predictor is trained under supervision using the sample electrode thickness set and the sample discharge slope set, and the training is completed after the test converges.
[0126] Configure the convergent discharge slope predictor on a cloud server.
[0127] The confidence analysis module 13 is specifically used for:
[0128] The multiple electrode thicknesses and multiple predicted electrode thicknesses are randomly combined to obtain multiple electrode thickness groups. The similarity and mean of each group are calculated to obtain the basic thickness confidence level.
[0129] The multiple discharge slopes and multiple predicted discharge slopes are randomly combined to obtain multiple discharge slope groups. The similarity and mean of each group are calculated to obtain the confidence level of the base slope.
[0130] Based on the number of tests and the prediction accuracy, the thickness confidence and discharge confidence are calculated.
[0131] Specifically, the phrase "calculating the thickness confidence level and discharge confidence level based on the number of tests and the prediction accuracy" includes:
[0132] Get the number of tests performed;
[0133] Calculate the ratio of the number of tests to the average number of tests to obtain the first confidence correction coefficient;
[0134] The accuracy rates of discharge slope prediction and electrode thickness prediction are obtained, and the average values are calculated to obtain the second confidence correction coefficient.
[0135] Based on the first confidence correction coefficient and the second confidence correction coefficient, the basic thickness confidence and the basic discharge confidence are corrected and calculated to obtain the thickness confidence and the discharge confidence.
[0136] The fusion output module 14 is specifically used for:
[0137] The average similarity between the thickness of multiple electrodes and the thickness of a standard electrode is calculated, and the thickness performance parameters are obtained by combining the thickness confidence score.
[0138] The average similarity between multiple discharge slopes and the standard discharge slope is calculated, and the discharge performance parameters are obtained by combining the discharge confidence level.
[0139] The performance test results are calculated based on the thickness performance parameters and discharge performance parameters.
[0140] In summary, the embodiments of this application have at least the following technical effects:
[0141] Compared to existing technologies, this application first uses a thickness testing module to perform multi-point thickness tests on ternary lithium battery electrodes, obtaining multiple electrode thicknesses and multiple discharge slopes. By collecting representative and stable data, a reliable foundation is provided for subsequent machine learning predictions and confidence analysis. Secondly, using a parameter prediction module, discharge slopes are predicted based on the multiple electrode thicknesses, obtaining multiple predicted discharge slopes. Electrode thickness is then predicted based on these multiple discharge slopes, obtaining multiple predicted electrode thicknesses. A machine learning model is used to achieve bidirectional prediction between electrode thickness and predicted discharge slope, and vice versa. The bidirectional mapping verifies data consistency, providing a prediction benchmark for subsequent confidence calculations. Furthermore, the confidence analysis module randomly combines multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate the thickness confidence score, and randomly combines multiple discharge slopes and multiple predicted discharge slopes to calculate the discharge confidence score. This transforms the consistency between measured and predicted data into quantifiable reliability indicators, fully considering the impact of test scale and model reliability on the confidence score. The final output thickness and discharge confidence scores provide clear credibility labels for electrode performance test results, addressing the pain point of traditional tests lacking quantifiable reference value. Finally, the fusion output module calculates the performance test results of ternary lithium battery electrodes based on the thickness and discharge confidence scores, combined with multiple electrode thicknesses and multiple discharge slopes. This quantifies the impact of thickness, discharge performance, and data reliability on the performance of ternary lithium battery electrodes, making the performance evaluation both comprehensive and reliable, and providing directly applicable quantitative evidence for electrode quality grading and process optimization. In this way, the random errors of traditional single-point testing are effectively avoided by collecting data from multiple measurement points and making bidirectional predictions, which improves the representativeness and correlation of the data. Furthermore, the reliability of the test results can be traced and evaluated through confidence quantification, which improves the representativeness and credibility of the performance test data of ternary lithium battery electrodes. This provides high-precision, quantifiable and highly reliable technical support for the quality control of ternary lithium battery electrodes, the optimization of production processes and the improvement of battery performance.
[0142] It should be noted that the descriptions of each embodiment in the above embodiments have different focuses. For parts that are not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.
[0143] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0144] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.
[0145] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.
[0146] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.
[0147] Although preferred embodiments of the invention have been described, those skilled in the art, once they have learned the basic inventive concept, can make other changes and modifications to these embodiments.
[0148] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.
Claims
1. A performance testing method for ternary lithium battery electrodes incorporating machine learning, characterized in that, include: Multiple thickness tests were conducted on the ternary lithium battery electrode to obtain multiple electrode thicknesses, and multiple discharge slopes of the ternary lithium battery electrode were also obtained. Discharge slope is predicted based on multiple electrode thicknesses to obtain multiple predicted discharge slopes. Electrode thickness is then predicted based on these multiple discharge slopes to obtain multiple predicted electrode thicknesses. The thickness confidence level is calculated by randomly combining the multiple electrode thicknesses and multiple predicted electrode thicknesses, and the discharge confidence level is calculated by randomly combining the multiple discharge slopes and multiple predicted discharge slopes, wherein the confidence level is calculated based on the number of tests and the prediction accuracy. Based on the thickness confidence and discharge confidence, the performance test results of the ternary lithium battery electrode are calculated by combining multiple electrode thicknesses and multiple discharge slopes. Specifically, the thickness confidence level is calculated by randomly combining the multiple electrode thicknesses and multiple predicted electrode thicknesses, and the discharge confidence level is calculated by randomly combining the multiple discharge slopes and multiple predicted discharge slopes, including: The multiple electrode thicknesses and multiple predicted electrode thicknesses are randomly combined to obtain multiple electrode thickness groups. The similarity and mean of each group are calculated to obtain the basic thickness confidence level. The multiple discharge slopes and multiple predicted discharge slopes are randomly combined to obtain multiple discharge slope groups. The similarity and mean of each group are calculated to obtain the confidence level of the base slope. Get the number of tests performed; Calculate the ratio of the number of tests to the average number of tests to obtain the first confidence correction coefficient; The accuracy rates of discharge slope prediction and electrode thickness prediction are obtained, and the average values are calculated to obtain the second confidence correction coefficient. Based on the first confidence correction coefficient and the second confidence correction coefficient, the basic thickness confidence and the basic discharge confidence are corrected and calculated to obtain the thickness confidence and the discharge confidence.
2. The ternary lithium battery electrode performance testing method integrating machine learning according to claim 1, characterized in that, Multi-point thickness tests were performed on the ternary lithium battery electrodes to obtain multiple electrode thicknesses, and multiple discharge slopes of the ternary lithium battery electrodes were also measured, including: Get the number of tests; The number of test locations are randomly selected on the ternary lithium battery electrode to perform multi-point thickness testing and obtain multiple electrode thicknesses. Constant current discharge test was performed on ternary lithium batteries to obtain the discharge curve at mid-SOC. According to the number of tests, the discharge curve is divided, and the slopes of the divided discharge curve segments are obtained as multiple discharge slopes.
3. The ternary lithium battery electrode performance testing method integrating machine learning according to claim 1, characterized in that, Discharge slopes are predicted based on multiple electrode thicknesses to obtain multiple predicted discharge slopes. Electrode thicknesses are then predicted based on these multiple discharge slopes to obtain multiple predicted electrode thicknesses, including: Call the discharge slope predictor; The thicknesses of the multiple electrodes are respectively input into the discharge slope predictor, and multiple predicted discharge slopes are obtained from the prediction output. Call the electrode thickness predictor; The multiple discharge slopes are input into the electrode thickness predictor, and multiple predicted electrode thicknesses are obtained from the prediction output.
4. The ternary lithium battery electrode performance testing method integrating machine learning according to claim 3, characterized in that, Calling the discharge slope predictor includes: Based on historical test data of ternary lithium battery electrodes, a set of sample electrode thicknesses was collected, along with the slope of the constant current discharge curve of ternary lithium batteries under different sample electrode thicknesses, and the set of sample discharge slopes was obtained by labeling. Construct a machine learning-based discharge slope predictor; The discharge slope predictor is trained under supervision using the sample electrode thickness set and the sample discharge slope set, and the training is completed after the test converges. Configure the convergent discharge slope predictor on a cloud server.
5. The ternary lithium battery electrode performance testing method integrating machine learning according to claim 1, characterized in that, Based on the aforementioned thickness confidence level and discharge confidence level, and combining multiple electrode thicknesses and multiple discharge slopes, the performance test results of the ternary lithium battery electrodes are calculated, including: The average similarity between the thickness of multiple electrodes and the thickness of a standard electrode is calculated, and the thickness performance parameters are obtained by combining the thickness confidence score. The average similarity between multiple discharge slopes and the standard discharge slope is calculated, and the discharge performance parameters are obtained by combining the discharge confidence level. The performance test results are calculated based on the thickness performance parameters and discharge performance parameters.
6. A ternary lithium battery electrode performance testing system integrating machine learning, characterized in that, For performing the method according to any one of claims 1-5, comprising: The thickness testing module is used to perform multi-point thickness testing on ternary lithium battery electrodes to obtain multiple electrode thicknesses and test and obtain multiple discharge slopes of the ternary lithium battery electrodes. The parameter prediction module is used to predict the discharge slope based on multiple electrode thicknesses to obtain multiple predicted discharge slopes, and to predict the electrode thickness based on the multiple discharge slopes to obtain multiple predicted electrode thicknesses. The confidence analysis module is used to randomly combine the multiple electrode thicknesses and multiple predicted electrode thicknesses to calculate the thickness confidence, and to randomly combine the multiple discharge slopes and multiple predicted discharge slopes to calculate the discharge confidence, wherein the confidence is calculated based on the number of tests and the prediction accuracy. The fusion output module is used to calculate the performance test results of the ternary lithium battery electrode based on the thickness confidence and discharge confidence, combined with multiple electrode thicknesses and multiple discharge slopes.