Confidence evaluation method and system for dynamic skipping of capacity grading process

Abstract

The invention relates to the technical field of battery manufacturing, in particular to a confidence evaluation method and system for dynamic skipping of a capacity grading process. According to the method, the problem that the parameter interaction cannot be evaluated by a traditional method is solved by analyzing the association of the process processing parameters, the material characteristic parameters and the historical battery capacity and constructing the parameter-capacity mapping matrix, then the confidence reference interval and the design threshold are generated in combination with the process capability index, the adaptive adjustment of the evaluation standard is realized, and the evaluation accuracy is improved. Then combining static historical mapping with dynamic electrical performance through electrical characteristic parameters to obtain a confidence evaluation interval, and meanwhile, calculating an input disturbance coefficient and a quality risk index to realize decoupling judgment of data reliability and real risk, break through the limitation that a traditional method cannot distinguish a fluctuation root, and finally, integrating various indexes so as to realize the prediction of the data reliability and the real risk. And a nonlinear algorithm is adopted to calculate the dynamic skipping confidence coefficient, and flexible processing is performed on the critical-state battery, so that the precise control of the capacity grading process of the lithium ion battery is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery manufacturing, and in particular to a confidence evaluation method and system for dynamically skipping a capacity division process. Background Art

[0002] In the lithium-ion battery production process, the "capacity grading process" is a critical inspection step to ensure product quality. Its core function is to accurately measure the actual capacity of the battery and group the batteries according to their capacity values. By screening batteries with similar parameters such as capacity, internal resistance, and voltage, the consistency of the assembled battery packs is effectively improved. The "dynamic skip" technology is a revolutionary breakthrough in the field of intelligent lithium-ion battery manufacturing. It focuses on data-driven real-time decision-making. While ensuring product quality, it can selectively omit the traditional capacity grading test, thereby significantly improving production efficiency and reducing manufacturing costs. The core technology that enables this intelligent decision-making is "confidence assessment." By establishing a precise mapping between the battery's pre-process parameters and capacity characteristics, this technology can calculate the confidence level of each battery cell in real time to skip the capacity grading test. When the confidence level reaches a preset threshold, the system automatically marks the battery as exempt from the capacity grading test and directly enters the subsequent process. If the threshold is not reached, the battery must still enter the traditional capacity grading process. This shift from relying on "experience judgment" to relying on "data-driven" not only reflects the upgrading and iteration of the industry, but also highlights the key leap in the lithium-ion battery manufacturing industry in the process of intelligent and digital transformation, and provides an important practical paradigm for the industry to build an efficient and precise intelligent manufacturing system.

[0003] Existing confidence assessment methods for "dynamic skipping of the capacity sizing process" are primarily based on multi-source data fusion and statistical modeling techniques. By collecting key process parameters from the battery's pre-process and combining them with historical production data to construct a capacity prediction model, the probability of the battery capacity meeting the standard is calculated as the confidence level for the capacity sizing skip. However, this assessment method suffers from a significant problem: it lacks reasonable differentiation and flexible handling of critical states. When the confidence level for capacity sizing skip falls within the critical range, it is unable to distinguish whether fluctuations are due to model input errors or actual battery quality risks. Predicted fluctuations under boundary conditions are mistakenly interpreted as performance differences. Consequently, for batteries with actual capacity differences of less than 0.3%, the extreme measures of "forced capacity sizing" and "direct skipping" are implemented, respectively. Even small parameter fluctuations trigger a step change in decision-making, resulting in inaccurate battery confidence assessments and ultimately a coexistence of over-detection and missed detection. Summary of the Invention

[0004] The main purpose of the present invention is to provide a confidence assessment method and system for dynamic skipping of a volume fractionation process, aiming to solve the technical problems in the prior art.

[0005] The present invention proposes a confidence assessment method for dynamic skipping of a capacity division process, comprising: Acquire multi-dimensional manufacturing characteristic data and historical reference data of the battery, wherein the multi-dimensional manufacturing characteristic data includes process processing parameters, material characteristic parameters, electrical characteristic parameters, and environmental equipment data; Obtaining a parameter-capacity mapping matrix of the battery according to the process processing parameters, material characteristic parameters and historical reference data; Obtaining a design threshold and a confidence reference interval of a battery according to the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix; Obtaining a confidence assessment interval of the battery according to the parameter-capacity mapping matrix, the electrical characteristic parameters, and the historical reference data; Obtaining an input disturbance coefficient and a quality risk index of the battery according to the process parameters, environmental equipment data, and the electrical characteristic parameters; Obtaining a dynamic skip confidence of the battery according to the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient, and the quality risk index; Determining whether the dynamic skip confidence is greater than the designed threshold; If the dynamic skip confidence is greater than the design threshold, it is determined that the battery is allowed to skip the capacity division process; If the dynamic skip confidence is not greater than the design threshold, it is determined that the battery is not allowed to skip the capacity division process.

[0006] Preferably, the step of obtaining a parameter-capacity mapping matrix of the battery according to the process parameters, material characteristic parameters and historical reference data includes: Obtaining core processing parameters according to the process processing parameters, wherein the core processing parameters include coating area density, thickness of the electrode sheet after rolling, and baking temperature, and obtaining core material parameters according to the material characteristic parameters, wherein the core material parameters include positive electrode slurry viscosity, active material median diameter, and diaphragm porosity; Obtaining historical battery capacity and historical confidence intervals based on the historical reference data; Obtaining a capacity impact weight of each parameter based on the historical confidence interval, the core processing parameters, and the material characteristic parameters; Arrange and combine a plurality of the core processing parameters and a plurality of the material characteristic parameters to obtain an interaction parameter group, and obtain an interaction effect heat map according to the interaction parameter group; The plurality of core processing parameters and the plurality of material characteristic parameters are screened according to the capacity influence weight and the interaction effect heat map to obtain key dominant parameters, and a parameter-capacity mapping matrix is ​​obtained according to the key dominant parameters and the historical battery capacity.

[0007] Preferably, the step of obtaining a design threshold and a confidence reference interval of a battery according to the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix comprises: Acquire reference sample data according to historical reference data, and acquire an initial confidence interval of the battery according to the reference sample data and the parameter-capacity mapping matrix; Acquire material fluctuation data and process fluctuation data according to the multi-dimensional manufacturing feature data, and acquire an overall consistency coefficient of the battery according to the material characteristic parameters and the material fluctuation data; Obtaining a historical minimum consistency coefficient and a median consistency coefficient according to the reference sample data, and obtaining a design threshold value according to the historical minimum consistency coefficient, the median consistency coefficient, the overall consistency coefficient, and the initial confidence interval; An overall process capability index of the battery is obtained according to the process processing parameters and the process fluctuation data, and a confidence reference interval is obtained according to the confidence initial interval, the process fluctuation data and the overall process capability index.

[0008] Preferably, the step of obtaining the confidence assessment interval of the battery according to the parameter-capacity mapping matrix, the electrical characteristic parameters and the historical reference data includes: Obtaining an initial confidence evaluation interval according to the parameter-capacity mapping matrix, the process parameters, and the material characteristic parameters; Obtaining real-time charge integral, real-time static open voltage decay rate, and real-time first charge and discharge efficiency according to the electrical characteristic parameters; Acquire multiple complete data sets based on the historical reference data, and acquire a performance impact weight group and a sensitivity coefficient group based on the complete data sets, the real-time charge integral, the real-time static open voltage decay rate, and the real-time first charge and discharge efficiency; Acquire electrical performance fluctuation data according to the multi-dimensional manufacturing characteristic data, and acquire a charge correction coefficient, an opening voltage attenuation correction coefficient, and an efficiency correction coefficient according to the sensitivity coefficient group and the electrical performance fluctuation data; A comprehensive correction coefficient is obtained according to the charge correction coefficient, the opening voltage attenuation correction coefficient, the efficiency correction coefficient and the performance impact weight group, and a confidence assessment interval is obtained according to the comprehensive correction coefficient and the initial confidence assessment interval.

[0009] Preferably, the step of obtaining the input disturbance coefficient and quality risk index of the battery according to the process parameters, environmental equipment data and the electrical characteristic parameters includes: Acquiring process setting values ​​according to the process processing parameters, wherein the process setting values ​​include coating speed, roller gap, and slitting tension, and acquiring real-time monitoring values ​​according to the environmental equipment data, wherein the real-time monitoring values ​​include oven temperature fluctuation variance, hydraulic pressure deviation, and vibration spectrum energy; Obtaining a process stability characteristic vector according to the process setting value and the real-time monitoring value, and obtaining a quality risk characteristic vector according to the electrical characteristic parameter; Acquire multiple environmental equipment noise components according to the process stability characteristic vector and the quality risk characteristic vector, and acquire lithium ion embedding entropy change and static open-voltage relaxation constant according to the quality risk characteristic vector; An input disturbance coefficient is obtained according to the plurality of environmental equipment noise components and the static open-voltage relaxation constant, and a quality risk index is obtained according to the lithium ion intercalation entropy change and the static open-voltage relaxation constant.

[0010] Preferably, the step of obtaining the dynamic skip confidence of the battery according to the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient and the quality risk index includes: Acquire a boundary proximity determination threshold according to the historical reference data, and acquire an interval relationship identifier of the battery according to the confidence evaluation interval, the confidence reference interval, and the boundary proximity determination threshold; Acquire an evaluation interval analysis strategy according to the interval relationship identifier, and acquire a correction coefficient according to the evaluation interval analysis strategy, the confidence evaluation interval, and the confidence reference interval, wherein the correction coefficient includes an overlap ratio correction coefficient and a boundary distance correction coefficient; Acquiring influencing factor data, and acquiring noise suppression parameters and risk sensitivity parameters based on the influencing factor data; A reference confidence is obtained according to the confidence evaluation interval, and a dynamic skip confidence is obtained according to the reference confidence, a correction coefficient, a noise suppression parameter, and a risk-sensitive parameter.

[0011] The present application also provides a confidence assessment system for dynamically skipping a capacity division process, including: A data acquisition module, configured to acquire multi-dimensional manufacturing characteristic data and historical reference data of the battery, wherein the multi-dimensional manufacturing characteristic data includes process processing parameters, material characteristic parameters, electrical characteristic parameters, and environmental equipment data; A matrix construction module, configured to obtain a parameter-capacity mapping matrix of a battery based on the process parameters, material characteristic parameters, and historical reference data; A benchmark analysis module, configured to obtain a design threshold and a confidence benchmark interval of a battery based on the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix; A dynamic evaluation module, configured to obtain a confidence evaluation interval of the battery based on the parameter-capacity mapping matrix, the electrical characteristic parameters, and the historical reference data; a risk diagnosis module, configured to obtain an input disturbance coefficient and a quality risk index of the battery based on the process parameters, environmental equipment data, and the electrical characteristic parameters; A comprehensive evaluation module, configured to obtain a dynamic skip confidence level of a battery based on the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient, and the quality risk index; A capacity decision module, configured to determine whether the dynamic skip confidence is greater than the designed threshold; If the dynamic skip confidence is greater than the design threshold, it is determined that the battery is allowed to skip the capacity division process; If the dynamic skip confidence is not greater than the design threshold, it is determined that the battery is not allowed to skip the capacity division process.

[0012] Preferably, the risk diagnosis module includes: a data acquisition unit, configured to acquire process setting values ​​based on the process parameters, wherein the process setting values ​​include coating speed, roller gap, and slitting tension, and acquire real-time monitoring values ​​based on the environmental equipment data, wherein the real-time monitoring values ​​include oven temperature fluctuation variance, hydraulic pressure deviation, and vibration spectrum energy; a feature extraction unit, configured to obtain a process stability feature vector based on the process setting value and the real-time monitoring value, and to obtain a quality risk feature vector based on the electrical characteristic parameter; a noise analysis unit, configured to obtain a plurality of environmental equipment noise components according to the process stability characteristic vector and the quality risk characteristic vector, and to obtain a lithium ion insertion entropy change and a static open-voltage relaxation constant according to the quality risk characteristic vector; A risk quantification unit is used to obtain an input disturbance coefficient based on the multiple environmental equipment noise components and the static open-voltage relaxation constant, and to obtain a quality risk index based on the lithium ion embedding entropy change and the static open-voltage relaxation constant.

[0013] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the confidence assessment method for dynamically skipping the capacity division process are implemented.

[0014] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the confidence assessment method for dynamically skipping the capacity division process.

[0015] The beneficial effects of the present invention are as follows: the present invention collects multi-dimensional manufacturing feature data of the battery in real time through a multi-source sensor network deployed on the production line, and obtains historical reference data at the same time, thereby providing historical experience reference for the current battery evaluation, and then constructs a parameter-capacity mapping matrix by analyzing the relationship between process processing parameters, material characteristic parameters and historical battery capacity. By using this method of quantifying the comprehensive impact of parameter interaction on capacity, the problem in the prior art that the synergistic effect of parameters cannot be evaluated, resulting in distortion of the evaluation caused by parameter fluctuations under critical states, is solved. Then, based on the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix, combined with the dynamic adjustment of the process capability index, the design threshold and confidence reference interval are dynamically generated. This method of adaptively adjusting the evaluation standard with material fluctuations and process stability can solve the limitations of the fixed threshold in the prior art and the problem of "one-size-fits-all standards leading to over-detection or missed detection". Historical reference data is used to correct the interval of electrical characteristic parameters in real time. The "static historical mapping" is combined with the "dynamic electrical performance" to obtain the confidence assessment interval. Then, the input disturbance coefficient is calculated using the environmental equipment data. Combined with the electrical characteristic analysis, the quality risk index is obtained to achieve the decoupling judgment of "data reliability" and "real risk", breaking through the limitation that traditional methods cannot distinguish the root causes of fluctuations. This method provides a dual verification basis for accurate decision-making under critical conditions, "assessment not only looks at the probability, but also the credibility of the probability". Finally, the confidence interval, benchmark interval, input disturbance coefficient and quality risk index are combined to calculate the dynamic skip confidence. By adopting a nonlinear correction mechanism, batteries with critical confidence levels are reasonably distinguished and flexibly processed to solve the problem of taking extreme decisions on batteries with actual capacity differences less than 0.3% in the existing technology, as well as the problem of "coexistence of over-detection and missed detection", which is conducive to improving the precise control of the lithium-ion battery capacity separation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of a method flow according to an embodiment of the present invention.

[0017] Figure 2 FIG. 1 is a schematic diagram of a system structure according to an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of the internal structure of a computer device according to an embodiment of the present application.

[0019] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0020] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0021] like Figure 1 As shown, the present application provides a confidence assessment method for dynamic skipping of a capacity division process, including: S1. Acquire multi-dimensional manufacturing characteristic data and historical reference data of the battery, wherein the multi-dimensional manufacturing characteristic data includes process processing parameters, material characteristic parameters, electrical characteristic parameters, and environmental equipment data; S2. Obtaining a battery parameter-capacity mapping matrix based on the process parameters, material characteristic parameters, and historical reference data; S3. Obtaining a design threshold and a confidence reference interval of the battery according to the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix; S4. Obtaining a confidence assessment interval of the battery according to the parameter-capacity mapping matrix, the electrical characteristic parameters, and the historical reference data; S5. Obtaining an input disturbance coefficient and a quality risk index of the battery according to the process parameters, environmental equipment data, and the electrical characteristic parameters; S6. Obtaining a dynamic skip confidence of the battery according to the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient, and the quality risk index; S7, determining whether the dynamic skip confidence is greater than the designed threshold; If the dynamic skip confidence is greater than the design threshold, it is determined that the battery is allowed to skip the capacity division process; If the dynamic skip confidence is not greater than the design threshold, it is determined that the battery is not allowed to skip the capacity division process.

[0022] As described in the above steps S1-S7, the present invention collects multi-dimensional manufacturing feature data of batteries in real time through a multi-source sensor network deployed on the production line, wherein the multi-dimensional manufacturing feature data refers to a key data set covering the entire battery production process, including process processing parameters, material characteristic parameters, electrical characteristic parameters and environmental equipment data, wherein the process processing parameters refer to the process parameters of the battery manufacturing process, the material characteristic parameters refer to the parameters reflecting the physical and chemical properties of the raw materials, the electrical characteristic parameters refer to the parameters characterizing the electrochemical performance of the battery, and the environmental equipment data refer to the parameters reflecting the production environment and equipment status. Through these data, the battery production status and performance characteristics can be fully characterized, providing a data basis for confidence assessment, and obtaining historical Reference data, where historical reference data refers to a collection of process parameters, material characteristic parameters, electrical characteristic parameters, environmental equipment data, and corresponding capacity separation test results (such as actual capacity, whether the capacity separation process was skipped, etc.) of similar batteries produced in the past under the same process conditions. It can provide historical experience references for current battery evaluation, and be used to build mapping relationships and dynamically adjust evaluation standards. Then, by analyzing the relationship between process parameters, material characteristic parameters, and historical battery capacity, a parameter-capacity mapping matrix is ​​constructed. Wherein, the parameter-capacity mapping matrix refers to a matrix that quantifies the relationship between parameter combinations and capacity. This method of converting the relationship between scattered parameters and capacity into a structured matrix and quantifying the interaction effects of parameters solves the current problem. There is a problem in the technology that it is impossible to quantify the comprehensive impact of parameter combinations on capacity, which leads to distortion in the evaluation of the impact of parameter fluctuations under critical conditions (for example, it is impossible to determine whether a parameter fluctuation acts alone or in conjunction with other parameters). Then, based on multi-dimensional manufacturing feature data and parameter-capacity mapping matrix, the design threshold and confidence reference interval are dynamically generated. Among them, the design threshold refers to the critical value for judging whether the battery is allowed to skip capacity division, and the confidence reference interval refers to the normal probability range of the battery capacity meeting the standard in historical production. Combined with the dynamic adjustment of the process capability index (such as shrinking the interval when the process stability is high and expanding it when the stability is low), it is used as the reference interval for evaluating the current battery. This method allows the evaluation standard to be adaptively adjusted with material fluctuations and process stability. The method can solve the limitations of the fixed threshold in the existing technology and solve the problem of "one-size-fits-all standards leading to over-detection or missed detection" (for example, fixed thresholds are still used for batches with large material fluctuations, causing over-detection or missed detection). Then, based on the parameter-capacity mapping matrix and historical reference data, the real-time correction interval of the electrical characteristic parameters is used to combine the "static historical mapping" with the "dynamic electrical performance" to obtain the confidence assessment interval, where the confidence assessment interval refers to the capacity compliance probability range calculated based on the current battery parameter combination and real-time electrical characteristic parameters, and the width of the interval changes dynamically with electrical performance fluctuations. Then, the input disturbance coefficient and quality risk index of the battery are obtained based on the process processing parameters, environmental equipment data and electrical characteristic parameters, whereThe input disturbance coefficient refers to an indicator that quantifies the degree of interference of environmental equipment noise (such as vibration and electromagnetic interference) on the evaluation input data. The quality risk index refers to an indicator for evaluating the quality risk of the battery itself. Through these two data, data reliability and battery body risk are quantified respectively, "noise interference" and "real quality risk" are decoupled, breaking through the limitation of ignoring the source of risk in existing technologies and solving the core problem of "inability to distinguish the cause of fluctuations". This method of quantifying data reliability and body risk separately enables the system to clearly determine whether it is a "false fluctuation" caused by environmental noise or a "real risk" caused by the battery's own quality problems when a certain parameter fluctuates, realizing "assessment not only looks at the probability, but also the credibility of the probability", thereby providing a double verification basis for accurate decision-making under critical conditions. Finally, through the comprehensive confidence interval, reference interval, input disturbance coefficient and quality The dynamic skip confidence level is calculated based on the risk index. The dynamic skip confidence level is a quantitative indicator that comprehensively evaluates the feasibility of skipping the capacity division process. By adopting a nonlinear correction mechanism, the confidence level near the critical value changes smoothly with parameter fluctuations, thereby rationally distinguishing and flexibly handling batteries with critical confidence levels. This solves the problem of taking extreme decisions for batteries with actual capacity differences less than 0.3% in existing technologies, and avoids the drastic switching between "forced capacity division" and "direct skipping" caused by small fluctuations. Finally, by comparing the dynamic skip confidence level with the design threshold, the capacity division decision is obtained, ensuring that high-risk batteries 100% enter capacity division and low-risk batteries can reliably skip capacity division. This breaks through the "black-or-white" decision-making model in existing technologies, solves the problem of "over-detection and missed detection coexisting", and is conducive to improving the precision control of the lithium-ion battery capacity division process.

[0023] In one embodiment, the step S2 of acquiring a battery parameter-capacity mapping matrix based on the process parameters, material characteristic parameters, and historical reference data includes: S21. Obtaining core processing parameters based on the process processing parameters, wherein the core processing parameters include coating area density, electrode thickness after rolling, and baking temperature, and obtaining core material parameters based on the material characteristic parameters, wherein the core material parameters include positive electrode slurry viscosity, active material median diameter, and separator porosity; S22. Obtaining historical battery capacity and historical confidence intervals according to the historical reference data; S23. Obtaining a capacity impact weight of each parameter based on the historical confidence interval, the core processing parameters, and the material characteristic parameters; S24, arranging and combining the plurality of core processing parameters and the plurality of material characteristic parameters to obtain an interaction parameter group, and obtaining an interaction effect heat map according to the interaction parameter group; S25. Screen the plurality of core processing parameters and the plurality of material characteristic parameters according to the capacity influence weight and the interaction effect heat map to obtain key dominant parameters, and obtain a parameter-capacity mapping matrix according to the key dominant parameters and the historical battery capacity.

[0024] As described in the above steps S21-S25, the present invention extracts core processing parameters from the process parameters, wherein the core processing parameters refer to the key process parameters that play a decisive role in the battery performance and quality during the battery production process, including coating surface density, electrode thickness after rolling, and baking temperature. At the same time, core material parameters are extracted from the material characteristic parameters, wherein the core material parameters refer to the key characteristic parameters of the battery material that have a core influence on the performance of the battery, including positive electrode slurry viscosity, active material median diameter, and diaphragm porosity. The existing technology often incorporates all parameters equally into the model, resulting in large noise interference and dilution of the influence of key parameters. This solution focuses on the core parameters to improve the sensitivity of the model to key variables, thereby solving the problem of insufficient prediction accuracy caused by parameter redundancy in the existing technology. Then, the historical battery capacity and historical confidence interval corresponding to the above parameters in time series are extracted from the historical reference data, wherein the historical battery capacity refers to In the past production, testing or use process, the actual capacity data of the battery was recorded. The historical confidence interval refers to the statistical confidence range of a batch of batteries calculated based on historical data, which provides a reference for the subsequent analysis of the impact of parameter fluctuations on the prediction results. Subsequently, these core processing parameters and core material parameters are used as input features, and the historical battery capacity is used as the output label. The model is trained using the XGBoost machine learning feature selection algorithm, and the model outputs the capacity influence weight of each parameter on the capacity prediction. Among them, the capacity influence weight refers to an indicator used to measure the degree of influence of each parameter (such as core processing parameters, core material parameters, etc.) on the battery capacity, so as to clarify the difference in the degree of influence of different parameters on the battery capacity. The existing technology lacks quantitative distinction between parameter influences, resulting in minor parameter fluctuations interfering with decision-making. This solution solves the problem of critical state misjudgment caused by fuzzy parameter influence in the existing technology by quantifying the importance of parameters.

[0025] Then, from all the obtained core processing parameters and core material parameters, generate interactive parameter groups that may have mutual influence (such as "coating surface density and positive electrode slurry viscosity", "baking temperature and active material middle diameter", "thickness of electrode sheet after rolling and porosity of diaphragm", etc.), where the interactive parameter group refers to a set of two parameters that interact with each other. For each group of interactive parameter groups, based on the trained XGBoost model, calculate the SHAP value (i.e. contribution value) of each group of parameters to the prediction result. At the same time, perform SHAP interaction value calculation on all combinations to obtain the joint interaction contribution value of any two parameters (A and B) in each combination, and then draw a graph for each group of interactive parameter groups. Draw a SHAP dependency graph: take parameter A as the x-axis and the SHAP value of parameter A as the y-axis. The graph marks the continuous values ​​of parameter B by color (red indicates high B value, blue indicates low B value), and judge the interaction relationship between the two parameters based on the graphical characteristics: if the SHAP value of A increases rapidly as A increases when the B value is high, it means that the positive impact of A and B on the output increases when both increase, and there is a synergistic effect between the two; if the SHAP value of A decreases as A increases (even from positive to negative), it means that B will weaken the positive impact of A, and there is a counteracting effect between the two. Then, for each pair of parameter combinations, calculate the interaction effect strength value: this value is the sum of the "joint interaction contribution value of the two parameters and the two" in all samples. The average value of the absolute difference of the sum of the SHAP values ​​of the individual parameters is used to quantify the strength of nonlinear interactions between parameters beyond linear superposition: if it is a positive value, it indicates that there is a synergistic effect (the combined effect is stronger than the sum of the individual effects); if it is a negative value, it indicates that there is a counteracting effect (the combined effect is weaker than the sum of the individual effects). Then, the interaction effect strength values ​​of all parameter pairs are integrated into a symmetric matrix. The off-diagonal elements in the matrix correspond to the interaction effect strength values ​​of the row and column parameters (positive values ​​are synergistic effects, negative values ​​are counteracting effects, and the absolute value reflects the strength of the interaction). Finally, all core feature parameters are used as row and column labels to form a matrix, and the off-diagonal cells represent the interaction effect strength of the corresponding row and column parameters. This scheme By quantifying the interaction effect through SHAP values, the problem of inaccurate confidence assessment caused by ignoring parameter interactions is solved. This enables the model to identify the impact of "parameter combination fluctuations" on capacity rather than the impact of single parameter fluctuations. This solves the problem of existing technologies that only consider the independent effects of individual parameters and ignore the interactions between parameters (for example, an increase in a single parameter is beneficial to capacity, but an increase together with another parameter may be detrimental), which leads to prediction bias in critical states. A two-color gradient mapping rule is used for visual encoding: red (dark red to light pink) indicates a positive synergistic effect (the higher the intensity, the darker the color), blue (dark blue to light blue) indicates a negative offsetting effect (the lower the intensity, the darker the color), and white indicates no significant interaction (the absolute value of the intensity is less than 0.05), thereby constructing an interaction effect heat map, which can clearly quantify whether the relationship between different parameters is synergistic or offsetting, providing a basis for subsequent optimization of process parameters.

[0026] Then, the core processing parameters and core material parameters with great influence on capacity, as well as the parameters contained in the interactive parameter group with strong interactive influence intensity are selected as key dominant parameters, and the fluctuation data corresponding to the core processing parameters and core material parameters are obtained according to the multi-dimensional manufacturing feature data. Then, multiple key dominant parameter groups are obtained according to the key dominant parameters and material data through the method of permutation and combination. These key dominant parameter groups (such as "coating surface density = 46mg / cm²+rolling thickness = 80μm+active material middle diameter = 16μm", etc.) are used as row dimension content, and the column dimension is divided into continuous battery capacity intervals (such as 3.5Ah~3.6Ah, 3.6Ah~3.7Ah, etc.). Similarly, these key dominant parameters are used as input features, and the corresponding historical battery capacity is used as the output label. The XGBoost machine learning feature selection algorithm is used to train the model so that the model learns the nonlinear relationship between parameters (including the relationship between single parameters). Dynamic influence and interactive effect of interactive parameter groups); after the model training is completed, a large number of artificially constructed key dominant parameters are input (covering the actual value range of each parameter), and the capacity prediction value and confidence interval corresponding to each key dominant parameter group are output. They are converted into parameter-capacity interval probability values ​​and used as matrix values. Among them, the parameter-capacity interval probability value refers to the probability that the battery capacity falls into the corresponding interval under a specific parameter or interactive parameter group. The matrix values ​​are filled into the matrix to finally form a parameter-capacity mapping matrix. The existing technology relies on a single threshold judgment and cannot handle the continuity and uncertainty of parameter fluctuations. However, this scheme converts the parameter combination into a probability distribution of capacity intervals through a mapping matrix, realizing the transition from "black-or-white" threshold decision-making to "probabilistic interval" decision-making, and solving the critical state step decision-making problem caused by the lack of parameter-capacity quantitative mapping in the existing technology (for example, the actual capacity difference <0.3% is classified as a different treatment method).

[0027] In one embodiment, the step S3 of obtaining a design threshold and a confidence reference interval of a battery according to the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix includes: S31, obtaining reference sample data according to historical reference data, and obtaining an initial confidence interval of the battery according to the reference sample data and the parameter-capacity mapping matrix; S32. Acquire material fluctuation data and process fluctuation data according to the multi-dimensional manufacturing feature data, and acquire an overall consistency coefficient of the battery according to the material characteristic parameters and the material fluctuation data; S33. Obtaining a historical minimum consistency coefficient and a median consistency coefficient according to the reference sample data, and obtaining a design threshold according to the historical minimum consistency coefficient, the median consistency coefficient, the overall consistency coefficient, and the initial confidence interval; S34. Obtain an overall process capability index of the battery according to the process processing parameters and the process fluctuation data, and obtain a confidence reference interval according to the initial confidence interval, the process fluctuation data and the overall process capability index.

[0028] As described in steps S31-S34 above, the present invention obtains reference sample data through historical reference data, wherein the reference sample data refers to battery samples that have historically "skipped capacity division and subsequently passed verification", as well as the corresponding processing parameters, material parameters, and actual capacity values ​​that have been confirmed to meet the standards. For each battery sample's historical processing parameters and historical material parameters, the corresponding key leading parameter groups are matched from the parameter-capacity mapping matrix, and the capacity compliance probabilities corresponding to these key leading parameter groups in the matrix are extracted. The extracted capacity compliance probabilities (e.g., 0.78 , 0.81, 0.83…0.95, 0.97) to form a "capacity compliance probability distribution", and then calculate the 95% confidence interval of the distribution through statistical methods to obtain the confidence reference interval, specifically: sort the summarized capacity compliance probability values ​​from small to large, remove 2.5% of the extreme values ​​at both ends, and the minimum and maximum values ​​of the remaining data form the confidence reference interval (for example, the sorted data is 0.78, 0.81, 0.83…0.95, 0.97, after removing 2.5% at both ends, the interval is [0.81, 0 .94]), and then calibrate the confidence reference interval according to the degree of influence of the interaction effect of the key dominant parameters on the actual capacity of the battery and the interaction effect strength value: when there is a strong synergistic effect between the parameters (interaction strength value 0.6~0.8), the upper and lower limits of the interval are reduced by 0.05; when the interaction strength value is greater than 0.8, the upper and lower limits of the interval are reduced by 0.08; if there is a significant offsetting effect between the parameters (interaction strength value -0.3~-0.5), the upper and lower limits of the interval are expanded by 0.05; when the interaction strength value is less than -0.5, the upper and lower limits of the interval are increased by 0.05. The lower limit difference is expanded by 0.08 to obtain the initial confidence interval. This method of calibrating the interval according to the interaction effect intensity value is an in-depth understanding and application of the complex relationship between parameters in the battery production process. The existing technology usually does not fully consider this interaction between parameters. The present invention is conducive to improving the reliability of the confidence interval in this way. Among them, the initial confidence interval refers to the capacity compliance probability range preliminarily determined by matching the parameter-capacity mapping matrix, which provides an initial probability range basis for the subsequent determination of the design threshold and confidence reference interval.

[0029] Then, material fluctuation data and process fluctuation data are obtained based on the multi-dimensional manufacturing feature data, wherein the material fluctuation data refers to the data reflecting the degree of fluctuation of the characteristic parameters of the materials used in the current batch of battery production deviating from the standard values ​​in actual production, and the process fluctuation data refers to the data reflecting the degree of fluctuation of the process processing parameters in the current batch of battery manufacturing process deviating from the set standard values ​​in actual production, and the standard deviation and average value corresponding to each parameter in the material characteristic parameters of the current batch of materials are calculated based on this, and then the ratio of 3 times the standard deviation of the parameter to the average value is calculated, and the difference between "1" and the ratio is calculated to obtain the consistency coefficient corresponding to the parameter, and then the average value of the consistency coefficients corresponding to multiple parameters is taken as the overall consistency coefficient of the material characteristics of the batch of batteries, wherein the overall The consistency coefficient refers to an indicator that measures the overall consistency of a batch of battery materials. At the same time, the historical consistency coefficients of the characteristics of multiple batches of battery materials are obtained based on historical reference data, and the median consistency coefficient and the historical minimum consistency coefficient are obtained through multiple historical consistency coefficients. Among them, the median consistency coefficient refers to the value in the middle position after the consistency coefficients of multiple batches of battery materials produced in history are sorted by size. The historical minimum consistency coefficient refers to the minimum value screened out from the consistency coefficients of multiple batches of battery materials produced in history. When the overall consistency coefficient is greater than the median consistency coefficient, the median of the initial confidence interval is used as the design threshold. When the overall consistency coefficient is not greater than the median consistency coefficient, the design threshold is obtained according to the degree of deviation. The formula " ", calculate the design threshold, where T represents the design threshold, M represents the median of the initial confidence interval, K represents the overall consistency coefficient, and K min Represents the historical minimum consistency coefficient, K median It represents the median of the consistency coefficient. The essence of this calculation formula is to dynamically assign the strictness of the threshold value through the degree of deviation of the material consistency. The higher the overall consistency coefficient of the material (the closer it is to 1), the smaller the fluctuation of the material parameters and the higher the quality stability of the product produced. At this time, the threshold requirement can be relatively loose (close to 90% of the median of the initial confidence interval); on the contrary, the lower the overall consistency coefficient (the farther it deviates from the median of the consistency coefficient), the greater the material fluctuation and the higher the quality risk. It is necessary to strictly screen with a higher threshold (close to 70% of the median of the initial confidence interval) to avoid the outflow of defective products. When the overall consistency coefficient is between the historical minimum consistency coefficient and the median of the consistency coefficient, the design threshold needs to be "smoothly adjusted" with the change of the overall consistency coefficient, rather than a step-by-step jump. In the formula, " The essence of the "part" is to normalize the overall consistency coefficient to the range of 0-1 (that is, the "degree of deviation": when the overall consistency coefficient is equal to the minimum consistency coefficient, it is 0, and when the overall consistency coefficient is equal to the median of the consistency coefficient, it is 1), and then multiply it by a floating ratio of 0.2, and finally achieve a linear increase in the design threshold from 70% of the median of the initial confidence interval to 90% of the median of the initial confidence interval. The existing technology usually does not fully consider the impact of fluctuations in material characteristic parameters on battery quality. Therefore, it is difficult to ensure the stability of product quality when the material fluctuates. The method of this scheme dynamically adjusts the design threshold according to the material consistency coefficient and flexibly changes the strictness of the threshold according to the material fluctuation situation reflects the refined management of battery production quality control. Instead of using a fixed threshold, it can be adaptively adjusted according to the actual material conditions, so that the production standard is more in line with the actual production situation, which helps to improve the stability of product quality.

[0030] Then, according to the process fluctuation data, the upper limit, lower limit, mean and standard deviation of the processing parameters corresponding to each processing parameter of the current batch of materials are obtained, and then the formula " ” Calculate the process capability index corresponding to each processing parameter, where CPK represents the process capability index, USL represents the upper limit of the processing parameter, μ represents the mean value of the processing parameter, LSL represents the lower limit of the processing parameter, and σ represents the standard deviation of the processing parameter. This formula evaluates the process stability and the ability to meet quality requirements by quantifying the degree of matching between the fluctuation of the processing parameters in the actual production process and the preset qualified range. The distance between the mean value of the processing parameter and the upper and lower limits of the processing parameter is calculated through the "USL-μ" and "μ-LSL" items respectively, and then divided by 3 times the standard deviation of the processing parameter, and the smaller value of the two is taken as the process capability index, so as to take into account whether the processing parameter deviates from the center and whether the fluctuation is too large. The larger the process capability index value, the stronger the process capability and the more stable the output of products that meet the standards. Then take the average of the process capability indices corresponding to multiple processing parameters. The mean is used as the overall process capability index of the batch of battery materials, where the overall process capability index refers to an indicator for measuring the overall capability of a batch of battery manufacturing processes. When the overall process capability index is greater than 1.33, the upper and lower limits of the initial confidence interval are simultaneously shrunk inward by 5%-10%; when the overall process capability index is less than 1.0, the upper and lower limits of the initial confidence interval are simultaneously expanded outward by 5%-15%, thereby forming an adaptive confidence reference interval. This method evaluates the stability of the battery production process from a process perspective, quantifies the process level through the process capability index, and then adjusts the initial confidence interval according to the process level to form the final confidence reference interval. This method enables the interval to more accurately reflect the confidence of the battery capacity under the current process conditions, thereby solving the problem of insufficient quantitative analysis of the relationship between the process and battery quality in the existing technology.

[0031] In one embodiment, the step S4 of obtaining the confidence assessment interval of the battery according to the parameter-capacity mapping matrix, the electrical characteristic parameters and the historical reference data includes: S41, obtaining an initial confidence evaluation interval according to the parameter-capacity mapping matrix, the process parameters, and the material characteristic parameters; S42, obtaining a real-time charge integral, a real-time static open voltage decay rate, and a real-time first charge and discharge efficiency according to the electrical characteristic parameters; S43. Acquire multiple complete data sets based on the historical reference data, and acquire a performance impact weight group and a sensitivity coefficient group based on the complete data sets, the real-time charge integral, the real-time static open voltage decay rate, and the real-time first charge and discharge efficiency; S44, acquiring electrical performance fluctuation data according to the multi-dimensional manufacturing characteristic data, and acquiring a charge correction coefficient, an opening voltage attenuation correction coefficient, and an efficiency correction coefficient according to the sensitivity coefficient group and the electrical performance fluctuation data; S45. Obtain a comprehensive correction coefficient according to the charge correction coefficient, the opening voltage attenuation correction coefficient, the efficiency correction coefficient, and the performance impact weight group, and obtain a confidence assessment interval according to the comprehensive correction coefficient and the initial confidence assessment interval.

[0032] As described in the above steps S41-S45, the present invention matches the process processing parameters and material characteristic parameters of the current battery with a pre-constructed parameter-capacity mapping matrix, thereby obtaining an initial confidence assessment interval based on the probability distribution corresponding to the current parameters in the matrix, wherein the initial confidence assessment interval refers to the capacity compliance probability corresponding to the key dominant parameter group matched by the parameter-capacity mapping matrix, and the capacity compliance probability range preliminarily determined by statistical analysis, and then obtains the real-time charge integral, real-time static open voltage decay rate and real-time first charge and discharge efficiency from the battery's electrical characteristic parameters, and then collects a large number of complete data sets from historical reference data, wherein the complete data set refers to various data sets covering the entire battery manufacturing process, including but not limited to processing parameters (such as temperature, pressure), material parameters (such as purity, particle size), historical capacity data, process fluctuation records, material consistency coefficient, etc., and obtains the corresponding historical charge integral, historical static open voltage decay rate and historical first charge and discharge efficiency from the historical electrical characteristic parameters, and then performs Z-score standardization on these indicators (based on historical The mean and standard deviation of historical data are used to eliminate dimensional effects, and outliers outside the range of plus or minus three times the standard deviation are removed (to ensure data reliability) to obtain a preprocessed historical electrical characteristic indicator dataset. For the real-time charge integral, real-time static opening voltage decay rate, and real-time first charge and discharge efficiency, Z-score normalization conversion is performed according to the standardization rules of historical data (i.e., using the mean and standard deviation of historical data) to obtain the standardized charge integral, standardized static opening voltage decay rate, and standardized first charge and discharge efficiency. These are used as model inputs. Z-score normalization eliminates dimensional differences between different electrical characteristic indicators (such as charge integral and opening voltage decay rate) and removes extreme outliers to ensure comparability between historical data and real-time data. This unifies electrical parameters with different physical meanings into the same analysis dimension, avoiding weight allocation bias caused by differences in units or numerical ranges. This solves the problem in existing technologies where high-value range parameters (such as charge integral) excessively dominate the evaluation results when directly mixing parameters of different dimensions. This lays a data foundation for subsequent weight calculation and correction coefficient generation.

[0033] Since the charge integral directly reflects the degree of lithiation of the active material, it has the highest weight on the capacity; the static open-voltage decay rate characterizes the stability of the SEI film and has the second highest influence on the first-week capacity; the first efficiency reflects the irreversible lithium loss, mainly affects the initial capacity retention rate, and has the lowest weight. This method of determining the priority of the impact of each parameter on the capacity by combining physical mechanism analysis can ensure that the weight distribution conforms to the internal electrochemical laws of the battery (such as the charge integral is directly related to the utilization rate of the active material and is the core determinant of the capacity), and avoids the "counter-common-sense" weighting caused by sample bias in the data-driven model (such as giving too high weight to secondary parameters and then using the predicted value to predict the capacity). The processed historical electrical characteristic indicator data set is used as the independent variable, and the corresponding historical measured capacity is used as the dependent variable. A multivariate linear regression equation is constructed, and the regression coefficient is calculated by the least squares method. It is then standardized (divided by its respective standard deviation) to obtain the standardized regression coefficient reflecting the influence intensity of each parameter. The standardized regression coefficient is then normalized (so that the sum of the weights is 1) to obtain the initial weight group, and the business constraint data is obtained. Business constraints are imposed on the initial weight group based on the business constraint data: if the weight of the charge is lower than 0.4 (such as 0.35), it is forced to be increased to 0.4, and the weights of other parameters are proportionally reduced (such as the static opening voltage attenuation). The rate is reduced from 0.3 to 0.25, and the first efficiency is kept at 0.2), ensuring that the total weight is 1. When the actual impact of the static opening voltage decay rate of a batch exceeds the model prediction (for example, the historical average decay rate of 1% corresponds to a 0.3% capacity loss, but the actual corresponding rate for this batch is 0.5%), the deviation compensation value is calculated (such as +0.2%), and the weight of the static opening voltage decay rate is increased from 0.3 to 0.35 (i.e. 0.3+0.05). Finally, the performance impact weight group is obtained, where the performance impact weight group refers to a combination of multiple values ​​reflecting the degree of dominance of the parameters on the performance, including the charge impact weight, the static opening voltage decay impact weight, and the first Efficiency affects weights. This correction method combines data-driven regression results with industrial practice experience, ensuring that the weight distribution conforms to physical mechanisms and can dynamically adapt to parameter fluctuations in the production process. Before each batch of production, the latest 500 sets of data are used to recalculate the regression coefficients and update the weights, and a threshold of no more than 10% is set for a single weight adjustment (to avoid drastic fluctuations). This method dynamically adjusts weights through business constraints and compensates weights with actual batch deviations, so that the weight distribution conforms to theoretical laws and adapts to production fluctuations. This method can break through the limitations of pure algorithm regression, integrate industrial practice experience, and achieve "dynamic self-calibration" of weights.

[0034] Subsequently, the standardized charge, standardized static voltage decay rate and standardized first efficiency in the historical data were used as input, and the capacity correction coefficient was used as output. HuberLoss was used as the loss function (balancing the influence of outliers and normal data) for model training. The training set and test set were divided according to the production time sequence. Time cross-validation was used to simulate the scenario of "using past data to predict the future" in real production to ensure that the correction coefficient output by the model can make the corrected capacity error ≤1.5% (that is, the absolute value of the difference between the model-predicted capacity and the measured capacity ≤1.5%), and the misjudgment rate of critical batteries (batteries with a capacity close to the qualified threshold) <0.5% (avoiding unqualified batteries from being judged as qualified or vice versa). By strongly binding model training to production timing, evaluation distortion caused by "future data leakage" can be avoided.

[0035] Then, the trained model is deployed to the real-time system of the production line (such as the manufacturing execution system), and the database is connected through the interface to realize automatic operation: during real-time calculation, the charge integral, static open circuit voltage decay rate, and first charge and discharge efficiency of the current batch of batteries are loaded from the database, and the standardized charge, standardized open circuit voltage decay rate, and standardized first efficiency are calculated according to the standardization rules of historical data (that is, using the mean and standard deviation of historical data). Based on the training correction model, the standardized charge, standardized open circuit voltage decay rate, and standardized first efficiency are input into the model to output a sensitivity coefficient group, where the sensitivity coefficient group refers to a set of coefficients that describes the degree of influence of fluctuations of various parameters (or parameter groups) on battery performance (such as capacity), including charge sensitivity coefficient, open circuit voltage sensitivity coefficient, and charge and discharge efficiency sensitivity coefficient. Then, the electrical performance fluctuation data is obtained based on the multi-dimensional manufacturing feature data, where the electrical performance fluctuation data refers to the actual fluctuation record of the electrical performance indicators of the battery during manufacturing or use, including the average charge integral, charge integral standard deviation, maximum allowable static open circuit voltage decay rate, maximum efficiency, and minimum efficiency, etc., and according to the formula " ” Calculate the charge correction coefficient, where α represents the charge correction coefficient, θ1 represents the charge sensitivity coefficient, and Q int Indicates the current integral of the battery charge, Q avg Indicates the average charge integral of batteries in the same batch, Q std Indicates the integrated standard deviation of the charge of batteries in the same batch, through " "Normalize the current battery charge integral to the standard deviation relative to the batch mean, multiply the normalized result by the charge sensitivity coefficient, and convert it into a correction ratio for the capacity; according to the formula " ” Calculate the static opening pressure attenuation correction coefficient, where β represents the static opening pressure attenuation correction coefficient, θ2 represents the static opening pressure sensitivity coefficient, V decay Indicates the current battery's 24-hour static voltage decay rate, Vmax Indicates the preset maximum allowable static opening pressure decay rate, through the " "The attenuation rate is mapped to the interval [0,1]. The larger the attenuation rate, the smaller the static opening voltage attenuation correction coefficient (the minimum is 0.9), indicating a greater capacity loss. According to the formula " ” Calculate the efficiency correction coefficient, where γ represents the efficiency correction coefficient, θ3 represents the efficiency sensitivity coefficient, ε represents the first charge and discharge efficiency of the current battery, and ε max Indicates the maximum efficiency, ε min Indicates the minimum efficiency, the formula is expressed by " Efficiency is mapped to the interval [0,1]. The higher the efficiency, the larger the efficiency correction coefficient (up to 1.15), indicating a higher reversible capacity. The normalized efficiency is converted into a correction ratio for capacity. This method of converting the absolute value of the electrical characteristic parameter into a correction ratio relative to the batch mean achieves "interpretable calculation" of the correction coefficient, solving the problem of existing technologies that rely on empirical judgment and lack mathematical basis for correction of electrical characteristic fluctuations, thereby making corrections under critical conditions more objective.

[0036] Finally, the weighted summation method is used to calculate the comprehensive correction coefficient based on the charge correction coefficient, static opening voltage attenuation correction coefficient and efficiency correction coefficient, as well as the corresponding charge influence weight, static opening voltage attenuation influence weight and first efficiency influence weight, and the formula " ” Calculate the lower limit of the confidence interval, where E L Indicates the lower limit of the confidence assessment interval, E init-L Indicates the lower limit of the initial confidence assessment interval, C represents the comprehensive correction coefficient, according to the formula " "Calculate the upper limit of the confidence assessment interval, where E U Indicates the upper limit of the confidence assessment interval, E init-URepresents the upper limit of the initial confidence assessment interval, C represents the comprehensive correction coefficient. The calculation formula takes the initial confidence assessment interval as the benchmark framework, and dynamically shrinks or expands the upper and lower limits of the interval through the comprehensive correction coefficient. The comprehensive correction coefficient is calculated by the sensitivity coefficient group of the electrical characteristic parameters, reflecting the degree of deviation of the current battery electrical characteristics from the historical benchmark. When the electrical characteristics perform better than the historical benchmark (such as higher charge and lower decay rate), the comprehensive correction coefficient moves the lower limit of the interval upward and the upper limit downward (interval contraction) to improve the evaluation accuracy; otherwise, it moves the lower limit of the interval downward and the upper limit upward (interval expansion) to reduce the risk of misjudgment. At the same time, the interval adjusted by the max and min function constraints never exceeds the initial interval range, ensuring that the evaluation result reflects both the electrical characteristics and the historical benchmark. The correction effect of the characteristic parameters is achieved while avoiding physical irrationality caused by excessive correction (such as the lower limit of the interval is higher than the upper limit or exceeds the theoretical limit of the material), realizing adaptive optimization of the evaluation interval, and finally the lower limit and upper limit of the interval obtained by calculation together constitute the final evaluation interval, thereby forming a confidence evaluation interval that is both corrected and in line with the basic theoretical constraints. This method not only utilizes real-time electrical characteristic parameters to optimize the evaluation interval, but also avoids excessive correction through theoretical constraints, achieving a balance of "local correction + global constraint", which is conducive to solving the problem in the existing technology that the interval changes drastically due to fluctuations in a single parameter under critical conditions, causing decision-making steps, so that the confidence evaluation interval presents a "smooth transition" within the critical range, reducing extreme decisions caused by small fluctuations.

[0037] In one embodiment, the step S5 of obtaining the input disturbance coefficient and quality risk index of the battery according to the process parameters, environmental equipment data, and the electrical characteristic parameters includes: S51. Acquire process setting values ​​based on the process parameters, wherein the process setting values ​​include coating speed, roller gap, and slitting tension, and acquire real-time monitoring values ​​based on the environmental equipment data, wherein the real-time monitoring values ​​include oven temperature fluctuation variance, hydraulic pressure deviation, and vibration spectrum energy; S52, obtaining a process stability characteristic vector according to the process setting value and the real-time monitoring value, and obtaining a quality risk characteristic vector according to the electrical characteristic parameter; S53, obtaining multiple environmental equipment noise components according to the process stability characteristic vector and the quality risk characteristic vector, and obtaining lithium ion insertion entropy change and static open-voltage relaxation constant according to the quality risk characteristic vector; S54. Obtain an input disturbance coefficient according to the multiple environmental equipment noise components and the static open-voltage relaxation constant, and obtain a quality risk index according to the lithium ion embedding entropy change and the static open-voltage relaxation constant.

[0038] As described in the above steps S51-S54, the present invention extracts coating speed, roller gap and slitting tension from the process processing parameters. These process setting values ​​refer to the standard values ​​of key process parameters set manually in the battery production process, which directly reflect the controllable variables in the production process. Their stability plays a decisive role in the physical properties of the electrode (such as coating uniformity and compaction density). At the same time, the oven temperature fluctuation variance, hydraulic pressure deviation and vibration spectrum energy are obtained from the environmental equipment data. These real-time monitoring values ​​refer to indicators collected in real time from the environmental equipment data that reflect the equipment operating status and workshop environment fluctuations. They are a direct reflection of the equipment operating status and workshop environment fluctuations and can be used as real-time "health indicators" of process stability. Through these data, the "input-state-output" three-dimensional data system of the production process can be integrated, breaking through the limitation of the existing technology that only focuses on data of a single link. Then, based on the battery production time sequence, the process setting values ​​and the real-time monitoring values ​​are synchronized at the millisecond level, for example, the coating speed is synchronized with the corresponding time. The oven temperature of the time stamp is accurately matched to eliminate the error caused by data collection delay. Then, the weight of each parameter is dynamically assigned to the synchronized multi-source data through the multi-head attention mechanism. When a parameter is abnormal (such as roller gap abnormality), its weight will be increased accordingly (such as from 0.3 to 0.7), and finally a 128-dimensional process stability feature vector is generated. The 128 dimensions of this vector correspond to the process stability performance of different aspects in the production process. For example, the 1st to 20th dimensions reflect the stability of the coating process (including the fluctuation amplitude of the coating speed, the uniformity of the coating thickness, etc.), the 21st to 40th dimensions reflect the stability of the roller process (covering the stability of the roller gap, the uniformity of the pressure, etc.), and the 41st to 60th dimensions correspond to the stability of the slitting process (involving the fluctuation of the slitting tension, the flatness of the cut, etc.). The remaining dimensions comprehensively reflect the impact of environmental factors on the process. The value of each dimension represents the level of stability in that aspect. Its mathematical expression is: P=Attention(W q X, W k X, W v X), where P represents the process stability feature vector, X represents the aligned multi-source data matrix, and W q X, W k X, W v X represents a learnable parameter. By combining "time synchronization" with "dynamic weighting," this method allows the feature vector to focus on the parameters that truly affect stability. This solves the problem of "abnormal signals being drowned out" by fixed parameter weights in existing technologies. In particular, it can highlight the impact of fluctuations in key parameters under critical conditions.

[0039] The real-time charge integral, real-time static open voltage decay rate and real-time first charge and discharge efficiency extracted from the electrical characteristic parameters directly reflect the quality of the electrochemical behavior inside the battery, such as the quality of SEI film formation, interface stability, etc. These parameters are then input into the generative adversarial network (GAN), and the encoder extracts potential features (that is, abstract information extracted from the electrical characteristic parameters that cannot be directly observed but can reflect the essential characteristics of the battery, such as features that reflect the chemical reaction laws inside the battery, the performance of the material, etc.). The discriminator is used to force the feature to be in the same distribution as the process stability feature vector (that is, the 128-dimensional process stability feature vector mentioned above) (even if the two are consistent in the distribution law of the data, so as to perform effective comparison and correlation analysis). Finally, a 128-dimensional quality risk feature vector is output, and feature decoupling is achieved in this process. By adding specific constraints during model training, quality risk The first 64 dimensions of the feature vector are strongly correlated with the material characteristic parameters to reflect the intrinsic characteristics of the material, and the last 64 dimensions are strongly correlated with the process parameters to reflect process defects. At the same time, when the static voltage decay rate is greater than 0.5mV / h, an L1 regularization constraint (a restriction mechanism) is imposed on the abstract data space where the encoder extracts the latent features, making the features related to anomalies in the latent space more prominent and enhancing sensitivity to abnormal data. This method, which uses the GAN network to force the distribution of process characteristics and electrical characteristics to be consistent, achieves effective correlation of cross-dimensional data, distinguishes material and process risks through feature decoupling, and combines L1 regularization to enhance anomaly sensitivity, breaking through the bottleneck of "separation of process and electrical characteristic parameters" in existing technologies, achieving precise positioning of risk sources, and solving the problem of critical state misjudgment caused by the inability to distinguish between "material intrinsic risk" and "process defect risk" in existing technologies.

[0040] Then, a process-electrical characteristic correlation matrix is ​​constructed by transposing the 128-dimensional process stability feature vector (rows to columns) and performing matrix multiplication with the 128-dimensional quality risk feature vector to obtain a 128×128 matrix. Each element in the matrix represents the correlation strength between the process stability feature and the quality risk feature. Then, the non-correlated components are extracted through singular value decomposition (a part after matrix decomposition, representing information that is not related to process characteristics and quality risk characteristics, mainly mechanical vibration during equipment operation, electromagnetic interference in the environment and other noises), and the environmental equipment noise components of each process are calculated. Among them, the environmental equipment noise component refers to the extracted component that is not related to process stability and quality risk, mainly from noise such as mechanical vibration of equipment and environmental electromagnetic interference. This component quantifies the degree of interference of noise on process signals. For example, the environmental equipment noise component of the coating process is calculated as follows: the L2 norm of the noise component corresponding to the coating process is divided by the L2 norm of the process stability feature vector corresponding to the coating process and 10 -6The ratio is the noise energy ratio of the coating process. The essence is to calculate the proportion of noise in the process signal. The higher the ratio, the greater the influence of noise on the process. The denominator adds the minimum value 10 -6 This is to prevent the error of dividing by 0 when the process signal strength is 0. This method can convert abstract noise into a calculable "energy proportion", making the noise impact from "unknowable" to "quantifiable", thereby solving the problem in existing technologies of ignoring noise interference and misjudging "parameter fluctuations caused by noise" as "real quality differences".

[0041] Then, the differential entropy change of the charge-discharge curve is analyzed from the risk feature vector. Specifically, the features related to the charge-discharge curve in the risk feature vector are used (these features are learned and extracted by the model from the charge-discharge curve contained in the electrical characteristic parameters). Combined with the principles of thermodynamics, the change law of the charge-discharge curve is first extracted, and then the differential entropy change of the curve is calculated to obtain the lithium ion embedding entropy change. Among them, the lithium ion embedding entropy change refers to the change in the degree of system disorder when lithium ions are embedded in the electrode material. The normal range is 0.2~0.5J / mol・K. Then, the static voltage decay curve is fitted through the risk feature vector, that is, the static voltage decay curve is used to fit the static voltage decay curve. The characteristics related to voltage decay (also learned and extracted by the model from the static voltage decay curve included in the electrical characteristic parameters) use mathematical curves (such as exponential curves) to match the actual data of voltage decay over time when static. By adjusting the curve parameters to make the curve as close to the actual data as possible, the static opening voltage relaxation constant is calculated based on the fitted curve equation. Among them, the static opening voltage relaxation constant refers to the rate at which the battery voltage decays when static. This method of converting the electrical characteristic curve into thermodynamic parameters (entropy change) and kinetic parameters (relaxation constant) realizes the leap from "curve morphology observation" to "mechanism parameter quantification", and then through the formula " ” Calculate the input disturbance coefficient, where K noise represents the input disturbance coefficient, max(φ noise ) represents the maximum noise energy ratio in each process, t represents the process memory cycle, the default is 8h, τ represents the static pressure relaxation constant, as time goes by, the influence of noise will gradually weaken, the maximum noise influence multiplied by the time attenuation result is the actual disturbance degree of noise on the model input at the current moment, and it is calculated by the formula " Calculate the quality risk index, where R riskIt represents the quality risk index, ΔS represents the entropy change of lithium ion embedding, t represents the process memory cycle, the default is 8h, and τ represents the static open-voltage relaxation constant. This formula comprehensively considers the SEI film disorder and voltage relaxation rate to evaluate the battery quality risk. The greater the entropy change of lithium ion embedding (the more disordered the SEI film), the smaller the static open-voltage relaxation constant (the faster the voltage decay), and the closer the entire score is to 1, the higher the quality risk of the body. The noise impact (input disturbance coefficient) and the body risk (quality risk index) are quantified through the formula to achieve a two-dimensional evaluation of "data reliability-quality risk" to solve the problem of over-detection or missed detection caused by fixed evaluation standards in existing technologies.

[0042] In one embodiment, the step S6 of obtaining the dynamic skip confidence of the battery according to the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient and the quality risk index includes: S61, obtaining a boundary proximity determination threshold according to the historical reference data, and obtaining an interval relationship identifier of the battery according to the confidence evaluation interval, the confidence reference interval, and the boundary proximity determination threshold; S62: obtaining an evaluation interval analysis strategy according to the interval relationship identifier, and obtaining correction coefficients according to the evaluation interval analysis strategy, the confidence evaluation interval, and the confidence reference interval, wherein the correction coefficients include an overlap ratio correction coefficient and a boundary distance correction coefficient; S63. Acquire influencing factor data, and acquire noise suppression parameters and risk sensitivity parameters based on the influencing factor data; S64. Obtain a reference confidence according to the confidence evaluation interval, and obtain a dynamic skip confidence according to the reference confidence, a correction coefficient, a noise suppression parameter, and a risk-sensitive parameter.

[0043] As described in the above steps S61-S64, the present invention obtains a boundary proximity determination threshold based on historical reference data, wherein the boundary proximity determination threshold refers to a critical value used to define the "boundary proximity" range, and compares the confidence assessment interval and the confidence reference interval. At the same time, the boundary proximity determination threshold is used as a reference to obtain the interval relationship identifier of the battery, wherein the interval relationship identifier refers to a classification label used to characterize the positional relationship between the confidence assessment interval and the confidence reference interval. The interval relationship identifier is divided into the following four categories: Complete inclusion (Category A): refers to the confidence assessment interval being completely covered by the confidence reference interval, that is, the lower limit of the confidence assessment interval is greater than or equal to the lower limit of the confidence reference interval. The confidence assessment interval has a lower limit, and the upper limit of the confidence assessment interval is less than or equal to the upper limit of the confidence reference interval. At this time, the two intervals show a relationship of inclusion and being included, and the confidence assessment interval as a whole is within the range of the confidence reference interval; Partial overlap (Class B): The confidence assessment interval and the confidence reference interval partially overlap but do not completely contain each other, that is, part of the confidence assessment interval falls within the confidence reference interval, and the other part falls outside the confidence reference interval. The two intervals have overlapping widths, but each also has non-overlapping parts, showing a cross-overlapping state; Complete separation (Class C): The confidence assessment interval and the confidence reference interval do not have any overlapping parts, and are completely independent of each other, that is, the confidence The upper limit of the evaluation interval is less than the lower limit of the confidence reference interval (the confidence evaluation interval is on the left side of the confidence reference interval), or the lower limit of the confidence evaluation interval is greater than the upper limit of the confidence reference interval (the confidence evaluation interval is on the right side of the confidence reference interval), the two intervals are separated from each other, there is no overlapping area, and the distance between the two intervals is greater than the boundary proximity judgment threshold; Boundary proximity (Class D): The confidence evaluation interval and the confidence reference interval have no overlapping parts and are completely independent of each other, that is, the upper limit of the confidence evaluation interval is less than the lower limit of the confidence reference interval (the confidence evaluation interval is on the left side of the confidence reference interval), or the lower limit of the confidence evaluation interval is greater than the upper limit of the confidence reference interval. limit (the confidence assessment interval is to the right of the confidence reference interval), the two intervals are separated from each other and there is no overlapping area, but the distance between the two intervals is less than the boundary proximity judgment threshold, that is, the edges of the two intervals are in a close state, although they do not cross but the distance is close. By clarifying the four types of interval relationships, the positional relationship between the confidence assessment interval and the confidence reference interval is systematically classified, breaking through the limitation of the existing technology of "black or white" judgment of the confidence assessment results, and using clear mathematical definitions to distinguish between states such as "complete inclusion" and "partial overlap", especially accurately defining the difference between "boundary proximity" and "complete separation", so as to provide a clear basis for the subsequent flexible processing of critical states.

[0044] The evaluation interval analysis strategy is obtained based on the interval relationship identifier. The specific evaluation interval analysis strategy is as follows: for Class C batteries, its baseline confidence is directly set to 0, so its dynamic skip confidence is also 0, and this type of battery will be forced to enter the capacity separation process; for Class A batteries, the median of its confidence evaluation interval is taken as the dynamic skip confidence, and the processing decisions for Class A batteries and Class C batteries achieve "efficient processing in a clear state" to avoid excessive calculation.

[0045] For Class B batteries, the median of the confidence evaluation interval is taken as the benchmark confidence, and the preset weight parameters are obtained based on the matching data of the historical Class B battery capacity classification results and the skip decision. Then, the formula " ” Calculate the overlap ratio correction coefficient, where ω1 represents the overlap ratio correction coefficient, O (ECI，RCI) W represents the width of the overlap between the confidence evaluation interval (ECI) and the confidence reference interval (RCI), ECI Indicates the total width of the confidence evaluation interval (ECI), μ represents the preset weight parameter, which is usually in the range of 0≤μ≤1. " is used to calculate the overlap ratio, which measures the degree of overlap between the confidence assessment interval and the confidence reference interval. The value is 1 when there is complete overlap and 0 when there is no overlap at all. The overlap ratio is nonlinearly adjusted by the weight parameter μ: when μ=1, the correction coefficient is directly equal to the overlap ratio, and the system relies entirely on the interval overlap judgment; when μ=0, the correction coefficient is always 1, and the system completely ignores the interval overlap. This design enables the system to dynamically adjust its confidence in battery quality according to the degree of interval overlap, and achieve accurate decision-making in the capacity separation process: the more overlap, the larger the correction coefficient, and the more likely the battery is to skip the process, thereby optimizing production efficiency while ensuring quality. This processing decision for Class B batteries takes into account both interval matching and historical experience, and uses nonlinear adjustment to balance "quality assurance" and "production efficiency", avoiding extreme decisions on partially overlapping batteries, in order to solve the problem of over-detection or missed detection caused by the one-size-fits-all strategy of existing technologies, especially achieving a smooth transition in the critical overlap state.

[0046] For Class D batteries, the median of the confidence assessment interval is taken as the baseline confidence. The baseline confidence refers to the starting point for calculating the dynamic skip confidence, which reflects the initial confidence level that the battery can skip the capacity separation process without considering factors such as interval relationship correction, noise and risk, and is expressed through the formula " ” Calculate the boundary distance correction coefficient, where ω2 represents the boundary distance correction coefficient, L eval Indicates the lower limit of the confidence evaluation interval (ECI), L base Indicates the lower limit of the confidence interval (RCI), Represents the boundary proximity judgment threshold, which is used to define the range of "boundary proximity". The formula quantifies the impact of boundary proximity on battery quality assessment by the distance between the lower limit of the confidence assessment interval and the lower limit of the confidence reference interval. When the absolute difference between the lower limits of the two intervals is " The smaller the value is, the closer the confidence assessment interval is to the confidence reference interval on the left side. At this time, the closer the boundary distance correction coefficient is to 1, the more reliable the battery quality is and the higher the confidence in skipping the capacity separation process is. On the contrary, if the difference exceeds the preset boundary proximity judgment threshold, the boundary distance correction coefficient begins to decrease significantly. When the difference is equal to the preset boundary proximity judgment threshold, the boundary distance correction coefficient drops to 0, which means that the battery quality risk is high and the capacity separation process needs to be forced to start. The preset boundary proximity judgment threshold plays the role of a "sensitivity regulator": a larger boundary proximity judgment threshold makes the system more tolerant to the boundary distance and allows more batteries to skip capacity separation; a smaller boundary proximity judgment threshold makes the system more tolerant to the boundary distance and allows more batteries to skip capacity separation; A small boundary proximity judgment threshold makes the system more sensitive, ensuring that batteries near the quality boundary are strictly inspected. This design mechanism mathematically quantifies the fuzzy concept of "boundary proximity", converts the interval position relationship into a computable confidence correction factor, and realizes refined decision-making control of the battery capacity separation process. This processing decision for Class D batteries converts the fuzzy concept of "boundary proximity" into a computable correction coefficient, and dynamically adjusts the confidence by comparing the distance with the threshold, so that batteries "close to the reference interval but not overlapping" can be reasonably evaluated, so as to solve the rough decision-making problem of either discharging or intercepting all batteries near the boundary in the existing technology.

[0047] For Class B and Class D batteries, noise suppression parameters and risk-sensitive parameters are first obtained based on the influencing factor data. The influencing factor data refers to various types of relevant data that affect battery quality assessment and capacity separation process decision-making, including the noise level of model input, the quality risk characteristics of the battery body, and production environment fluctuations. The noise suppression parameter refers to the parameter used to reduce the impact of the input disturbance coefficient on the dynamic skip confidence level, and the risk-sensitive parameter refers to the parameter used to amplify the impact of the quality risk index on the dynamic skip confidence level. Finally, the formula " ” Calculate the dynamic skip confidence, where C skip represents the dynamic skip confidence, C base represents the reference confidence level, ω represents the correction coefficient, which is calculated from the interval relationship (B / D type), K noise represents the input disturbance coefficient, R risk represents the quality risk index, m represents the noise suppression parameter, represents the mth power of the input disturbance coefficient, k represents the risk sensitive parameter, It represents the kth power of the quality risk index. The formula integrates four factors, namely, benchmark confidence, interval relationship correction, model noise and ontological risk, through a product mechanism to realize the quantification of conservative decision-making under multi-dimensional constraints: taking the benchmark confidence as the basic value, the interval relationship correction coefficient is used to adjust the initial credibility. In view of the uncertainty of the model input, the noise coefficient is attenuated by power operation through the noise suppression parameter to ensure that the skipping confidence is reduced when the data is unreliable. At the same time, the quality risk index is amplified by power operation to significantly reduce the skipping confidence of high-risk batteries. Finally, through the product of the four factors, any unfavorable factors in any dimension (high noise or high risk) will lead to a reduction in the overall skipping confidence, thereby ensuring the battery Under the premise of quality, the efficiency of the capacity division process is optimized to achieve a scientific balance between data reliability, interval matching and quality risk. Then, a high-risk forced interception step is executed. Through this method, the scattered influencing factors are systematically associated so that small parameter fluctuations only cause a gentle change in confidence, avoiding decision-making steps. In addition, if the quality risk index is greater than the preset risk threshold, the dynamic skip confidence of the battery is forced to be set to 0, forming a "last line of defense". This method of adding an absolute risk interception mechanism in addition to the comprehensive assessment gives priority to quality requirements to solve the problem that high risks may be concealed by comprehensive calculations and lead to missed inspections, especially for batteries that are close to qualified but have extremely high intrinsic risks. Effective interception is achieved.

[0048] The present application also provides a confidence assessment system for dynamically skipping a capacity division process, including: A data acquisition module, configured to acquire multi-dimensional manufacturing characteristic data and historical reference data of the battery, wherein the multi-dimensional manufacturing characteristic data includes process processing parameters, material characteristic parameters, electrical characteristic parameters, and environmental equipment data; A matrix construction module, configured to obtain a parameter-capacity mapping matrix of a battery based on the process parameters, material characteristic parameters, and historical reference data; A benchmark analysis module, configured to obtain a design threshold and a confidence benchmark interval of a battery based on the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix; A dynamic evaluation module, configured to obtain a confidence evaluation interval of the battery based on the parameter-capacity mapping matrix, the electrical characteristic parameters, and the historical reference data; a risk diagnosis module, configured to obtain an input disturbance coefficient and a quality risk index of the battery based on the process parameters, environmental equipment data, and the electrical characteristic parameters; A comprehensive evaluation module, configured to obtain a dynamic skip confidence level of a battery based on the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient, and the quality risk index; A capacity decision module, configured to determine whether the dynamic skip confidence is greater than the designed threshold; If the dynamic skip confidence is greater than the design threshold, it is determined that the battery is allowed to skip the capacity division process; If the dynamic skip confidence is not greater than the design threshold, it is determined that the battery is not allowed to skip the capacity division process.

[0049] In one embodiment, the risk diagnosis module includes: a data acquisition unit, configured to acquire process setting values ​​based on the process parameters, wherein the process setting values ​​include coating speed, roller gap, and slitting tension, and acquire real-time monitoring values ​​based on the environmental equipment data, wherein the real-time monitoring values ​​include oven temperature fluctuation variance, hydraulic pressure deviation, and vibration spectrum energy; a feature extraction unit, configured to obtain a process stability feature vector based on the process setting value and the real-time monitoring value, and to obtain a quality risk feature vector based on the electrical characteristic parameter; a noise analysis unit, configured to obtain a plurality of environmental equipment noise components according to the process stability characteristic vector and the quality risk characteristic vector, and to obtain a lithium ion insertion entropy change and a static open-voltage relaxation constant according to the quality risk characteristic vector; A risk quantification unit is used to obtain an input disturbance coefficient based on the multiple environmental equipment noise components and the static open-voltage relaxation constant, and to obtain a quality risk index based on the lithium ion embedding entropy change and the static open-voltage relaxation constant.

[0050] The present invention also provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the confidence assessment method for dynamically skipping the capacity division process are implemented.

[0051] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the confidence assessment method for dynamically skipping the capacity division process.

[0052] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media provided in this application and used in the embodiments may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct RAM bus dynamic RAM (DRDRAM), and RAM bus dynamic RAM (RDRAM).

[0053] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0054] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present invention, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A confidence assessment method for dynamic skipping of a capacity division process, characterized in that: include: Acquire multi-dimensional manufacturing characteristic data and historical reference data of the battery, wherein the multi-dimensional manufacturing characteristic data includes process processing parameters, material characteristic parameters, electrical characteristic parameters, and environmental equipment data; Obtaining a parameter-capacity mapping matrix of the battery according to the process processing parameters, material characteristic parameters and historical reference data; Obtaining a design threshold and a confidence reference interval of a battery according to the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix; Obtaining a confidence assessment interval of the battery according to the parameter-capacity mapping matrix, the electrical characteristic parameters, and the historical reference data; Obtaining an input disturbance coefficient and a quality risk index of the battery according to the process parameters, environmental equipment data, and the electrical characteristic parameters; Obtaining a dynamic skip confidence of the battery according to the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient, and the quality risk index; Determining whether the dynamic skip confidence is greater than the designed threshold; If the dynamic skip confidence is greater than the design threshold, it is determined that the battery is allowed to skip the capacity division process; If the dynamic skip confidence is not greater than the design threshold, it is determined that the battery is not allowed to skip the capacity division process.

2. The confidence evaluation method for dynamic skipping of a capacity division process according to claim 1, characterized in that: The step of obtaining a parameter-capacity mapping matrix of the battery according to the process parameters, material characteristic parameters and historical reference data includes: Obtaining core processing parameters according to the process processing parameters, wherein the core processing parameters include coating area density, thickness of the electrode sheet after rolling, and baking temperature, and obtaining core material parameters according to the material characteristic parameters, wherein the core material parameters include positive electrode slurry viscosity, active material median diameter, and diaphragm porosity; Obtaining historical battery capacity and historical confidence intervals based on the historical reference data; Obtaining a capacity impact weight of each parameter based on the historical confidence interval, the core processing parameters, and the material characteristic parameters; Arrange and combine a plurality of the core processing parameters and a plurality of the material characteristic parameters to obtain an interaction parameter group, and obtain an interaction effect heat map according to the interaction parameter group; The plurality of core processing parameters and the plurality of material characteristic parameters are screened according to the capacity influence weight and the interaction effect heat map to obtain key dominant parameters, and a parameter-capacity mapping matrix is ​​obtained according to the key dominant parameters and the historical battery capacity.

3. The confidence evaluation method for dynamic skipping of a capacity division process according to claim 1, characterized in that: The step of obtaining a design threshold and a confidence reference interval of a battery according to the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix includes: Acquire reference sample data according to historical reference data, and acquire an initial confidence interval of the battery according to the reference sample data and the parameter-capacity mapping matrix; Acquire material fluctuation data and process fluctuation data according to the multi-dimensional manufacturing feature data, and acquire an overall consistency coefficient of the battery according to the material characteristic parameters and the material fluctuation data; Obtaining a historical minimum consistency coefficient and a median consistency coefficient according to the reference sample data, and obtaining a design threshold value according to the historical minimum consistency coefficient, the median consistency coefficient, the overall consistency coefficient, and the initial confidence interval; An overall process capability index of the battery is obtained according to the process processing parameters and the process fluctuation data, and a confidence reference interval is obtained according to the confidence initial interval, the process fluctuation data and the overall process capability index.

4. The confidence evaluation method for dynamic skipping of a capacity division process according to claim 1, characterized in that: The step of obtaining a confidence assessment interval of the battery according to the parameter-capacity mapping matrix, the electrical characteristic parameters and the historical reference data includes: Obtaining an initial confidence evaluation interval according to the parameter-capacity mapping matrix, the process parameters, and the material characteristic parameters; Obtaining real-time charge integral, real-time static open voltage decay rate, and real-time first charge and discharge efficiency according to the electrical characteristic parameters; Acquire multiple complete data sets based on the historical reference data, and acquire a performance impact weight group and a sensitivity coefficient group based on the complete data sets, the real-time charge integral, the real-time static open voltage decay rate, and the real-time first charge and discharge efficiency; Acquire electrical performance fluctuation data according to the multi-dimensional manufacturing characteristic data, and acquire a charge correction coefficient, an opening voltage attenuation correction coefficient, and an efficiency correction coefficient according to the sensitivity coefficient group and the electrical performance fluctuation data; A comprehensive correction coefficient is obtained according to the charge correction coefficient, the opening voltage attenuation correction coefficient, the efficiency correction coefficient and the performance impact weight group, and a confidence assessment interval is obtained according to the comprehensive correction coefficient and the initial confidence assessment interval.

5. The confidence evaluation method for dynamic skipping of a capacity division process according to claim 1, characterized in that: The step of obtaining the input disturbance coefficient and quality risk index of the battery according to the process parameters, environmental equipment data and the electrical characteristic parameters includes: Acquiring process setting values ​​according to the process processing parameters, wherein the process setting values ​​include coating speed, roller gap, and slitting tension, and acquiring real-time monitoring values ​​according to the environmental equipment data, wherein the real-time monitoring values ​​include oven temperature fluctuation variance, hydraulic pressure deviation, and vibration spectrum energy; Obtaining a process stability characteristic vector according to the process setting value and the real-time monitoring value, and obtaining a quality risk characteristic vector according to the electrical characteristic parameter; Acquire multiple environmental equipment noise components according to the process stability characteristic vector and the quality risk characteristic vector, and acquire lithium ion embedding entropy change and static open-voltage relaxation constant according to the quality risk characteristic vector; An input disturbance coefficient is obtained according to the plurality of environmental equipment noise components and the static open-voltage relaxation constant, and a quality risk index is obtained according to the lithium ion intercalation entropy change and the static open-voltage relaxation constant.

6. The confidence evaluation method for dynamic skipping of a capacity division process according to claim 1, characterized in that: The step of obtaining the dynamic skip confidence of the battery according to the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient, and the quality risk index includes: Acquire a boundary proximity determination threshold according to the historical reference data, and acquire an interval relationship identifier of the battery according to the confidence evaluation interval, the confidence reference interval, and the boundary proximity determination threshold; Acquire an evaluation interval analysis strategy according to the interval relationship identifier, and acquire a correction coefficient according to the evaluation interval analysis strategy, the confidence evaluation interval, and the confidence reference interval, wherein the correction coefficient includes an overlap ratio correction coefficient and a boundary distance correction coefficient; Acquiring influencing factor data, and acquiring noise suppression parameters and risk sensitivity parameters based on the influencing factor data; A reference confidence is obtained according to the confidence evaluation interval, and a dynamic skip confidence is obtained according to the reference confidence, a correction coefficient, a noise suppression parameter, and a risk-sensitive parameter.

7. A confidence assessment system for dynamic skipping of a capacity division process, characterized in that: include: A data acquisition module, configured to acquire multi-dimensional manufacturing characteristic data and historical reference data of the battery, wherein the multi-dimensional manufacturing characteristic data includes process processing parameters, material characteristic parameters, electrical characteristic parameters, and environmental equipment data; A matrix construction module, configured to obtain a parameter-capacity mapping matrix of a battery based on the process parameters, material characteristic parameters, and historical reference data; A benchmark analysis module, configured to obtain a design threshold and a confidence benchmark interval of a battery based on the multi-dimensional manufacturing feature data and the parameter-capacity mapping matrix; A dynamic evaluation module, configured to obtain a confidence evaluation interval of the battery based on the parameter-capacity mapping matrix, the electrical characteristic parameters, and the historical reference data; a risk diagnosis module, configured to obtain an input disturbance coefficient and a quality risk index of the battery based on the process parameters, environmental equipment data, and the electrical characteristic parameters; A comprehensive evaluation module, configured to obtain a dynamic skip confidence level of a battery based on the confidence evaluation interval, the confidence reference interval, the input disturbance coefficient, and the quality risk index; A capacity decision module, configured to determine whether the dynamic skip confidence is greater than the designed threshold; If the dynamic skip confidence is greater than the design threshold, it is determined that the battery is allowed to skip the capacity division process; If the dynamic skip confidence is not greater than the design threshold, it is determined that the battery is not allowed to skip the capacity division process.

8. The confidence evaluation system for dynamic skipping of a capacity division process according to claim 7, characterized in that: The risk diagnosis module includes: a data acquisition unit, configured to acquire process setting values ​​based on the process parameters, wherein the process setting values ​​include coating speed, roller gap, and slitting tension, and acquire real-time monitoring values ​​based on the environmental equipment data, wherein the real-time monitoring values ​​include oven temperature fluctuation variance, hydraulic pressure deviation, and vibration spectrum energy; a feature extraction unit, configured to obtain a process stability feature vector based on the process setting value and the real-time monitoring value, and to obtain a quality risk feature vector based on the electrical characteristic parameter; a noise analysis unit, configured to obtain a plurality of environmental equipment noise components according to the process stability characteristic vector and the quality risk characteristic vector, and to obtain a lithium ion insertion entropy change and a static open-voltage relaxation constant according to the quality risk characteristic vector; A risk quantification unit is used to obtain an input disturbance coefficient based on the multiple environmental equipment noise components and the static open-voltage relaxation constant, and to obtain a quality risk index based on the lithium ion embedding entropy change and the static open-voltage relaxation constant.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images