Battery life prediction method compatible with different design capacities

Abstract

The invention discloses a battery life prediction method compatible with different design capacities, and the method comprises the steps: optimizing the frequency factors and activation energy parameters of a positive electrode and a negative electrode through a parameter normalization technology and a positive and negative electrode separation modeling frame, dynamically associating the temperature, the design capacities and the aging rate, and employing a Levenberg-Marquardt algorithm, and integration of data input, dynamic prediction and sensitivity analysis is realized in combination with a visual interaction platform, the technical bottlenecks of low precision, poor environmental adaptability and complex operation of cross-capacity battery life prediction are overcome, and a high-precision and high-compatibility intelligent prediction tool is provided for electric vehicles, energy storage systems and other scenes.

Description

Technical Field

[0001] The present invention relates to the technical field of battery life modeling and state of health assessment, and specifically to a battery life prediction method that can be compatible with different design capacities. Background Art

[0002] In the context of the accelerating transformation of the energy structure, lithium-ion batteries have become the core support technology for electric mobility, intelligent electronic devices, and renewable energy storage systems. With high energy density, excellent rate performance, and long cycle life, lithium batteries have been widely used in multiple key fields. However, with the continuous expansion of the application scale, the problems of battery aging and performance degradation during long-term operation have become increasingly prominent, becoming the key bottleneck affecting the safety, stability, and economy of the system.

[0003] Battery aging usually manifests as phenomena such as capacity decline, internal resistance increase, and power output ability reduction. These degradation processes not only shorten the battery life but also may cause safety risks such as unstable charge and discharge and thermal runaway. The aging mechanism itself involves complex multi-physical field coupling processes, including multiple dimensions such as electrochemical reaction decline, material structure change, and side reaction accumulation. Among them, the microstructural evolution and active material loss of the positive and negative electrode materials are considered the main sources of capacity attenuation. During long-cycle cycling, the two electrode materials frequently experience volume expansion, interface rupture, and the cumulative effect of ion migration obstacles, ultimately resulting in the difficulty of recovering the battery capacity.

[0004] Although there have been a large number of attempts to establish prediction models for battery aging, the current mainstream methods still have certain limitations. Most models are based on empirical fitting or simplified assumptions of ideal conditions, unable to fully consider the differences at the material level, nor accurately reflect the actual impact of external operating conditions (such as temperature, current density, etc.) on the aging rate. This modeling method often has problems such as large prediction deviations and poor adaptability when facing batteries with multiple capacity specifications and different material systems, restricting its popularization and application in practical engineering. Therefore, there is an urgent need to construct a new type of aging modeling method with a more physical basis and better scalability, which can take into account the coupling relationship between material behavior, environmental variables, and the life evolution process, so as to achieve scientific prediction and management of the entire life cycle of the battery.

[0005] Therefore, it is desired to have a battery life prediction method that can be compatible with different design capacities, and combine the core mechanism of capacity decline to construct a high-precision life prediction framework for batteries with different design capacities. Summary of the Invention

[0006] To solve the above technical problems, this application is proposed. Embodiments of this application provide a battery life prediction method that can be compatible with different design capacities. By separately modeling the degradation behaviors of the positive and negative electrode materials and introducing the Arrhenius formula to describe the temperature-dependent aging rate, the adaptability and prediction accuracy of the model are significantly improved. Combining actual test data for parameter fitting and sensitivity analysis can provide strong theoretical support and technical means for subsequent battery design optimization, thermal management strategy formulation, and state-of-health assessment.

[0007] According to one aspect of this application, there is provided a battery life prediction method that can be compatible with different design capacities, which includes: S101, performing long-term cycling tests on lithium-ion batteries with multiple design capacities to obtain discharge capacities, voltage curves, and temperature data at different temperatures, rates, and numbers of cycles; S102, constructing a first-order kinetic reaction model based on the separation of the positive and negative electrodes, separately constructing positive and negative electrode capacity decay equations, and incorporating the influence of temperature on the aging rate into the model through the Arrhenius equation, with the total capacity expressed as the sum of the capacity decay contributions of the positive and negative electrodes; S103, using the nonlinear least squares method combined with the Levenberg-Marquardt algorithm to optimize and fit the frequency factors and activation energies of the positive and negative electrodes, minimizing the sum of the squared residuals between the model prediction values and the measured values; S104, based on the trained model, inputting the positive and negative electrode design capacities and operating condition parameters of the target battery, dynamically predicting the evolution trend of the remaining capacity with the number of cycles, and generating a state-of-health assessment result; S105, providing a visualization operation platform for data input, model calculation, and result display.

[0008] Preferably, in step S101, the design capacities of the lithium-ion batteries include 280 Ah, 50 Ah, and 40 Ah, and the test ambient temperature is at least one of 25°C, 30°C, and 35°C.

[0009] Preferably, in step S101, a high-temperature warning is triggered when the battery temperature exceeds 55°C, and the test is forced to abort when it reaches 70°C.

[0010] Preferably, in step S102, the positive and negative electrode capacity decay equations satisfy:

[0011] The positive electrode capacity decay model is:

[0012]

[0013] The negative electrode capacity decay model is:

[0014]

[0015] Where and respectively represent the available capacities contributed by the positive electrode and the negative electrode in the t-th cycle; and are the initial design capacities of the positive electrode and the negative electrode respectively; A (+) and A (-) are the frequency factors of the aging reactions of the positive electrode and the negative electrode; and respectively represent the aging activation energies of the positive electrode and the negative electrode.

[0016] Preferably, in the step S103, the objective function of parameter estimation is:

[0017]

[0018] where N is the total number of test cycles, C measured,i is the measured discharge capacity value in the i-th cycle, C model,i is the predicted value of the model, the pre-exponential factor A and the activation energy E a are model parameters, and the optimal parameter combination is obtained through iterative optimization.

[0019] Preferably, in the step S104, the model is dynamically updated by a machine learning algorithm, and the prediction curve is adjusted in combination with the real-time input temperature data.

[0020] Preferably, the step S105 includes: supporting the input of battery design capacity, temperature and number of cycles; automatically generating a capacity-time curve and an aging rate trend chart; providing a temperature sensitivity heat map and an analysis of the influence degree of key parameters.

[0021] Preferably, the method further includes a model verification step: dividing the test data into a training set and a verification set, and evaluating the generalization ability of the model through residual analysis and cross-validation.

[0022] According to one aspect of the present application, there is also provided an electronic device, characterized in that it includes a memory and a processor, the memory stores an executable program, and when the processor runs the executable program, it implements the battery life prediction method compatible with different design capacities as described above.

[0023] According to one aspect of the present application, there is also provided a computer-readable storage medium, characterized in that the computer-readable storage medium stores computer execution instructions, and when the computer execution instructions are executed by a processor, they are used to implement the battery life prediction method compatible with different design capacities as described above.

[0024] The present invention designs a battery life prediction method that can be compatible with different design capacities. By constructing a first-order kinetics - Arrhenius composite model with independent positive and negative electrodes, high-precision life assessment of cross-capacity batteries is achieved. This method innovatively models the capacity attenuation processes of the positive and negative electrodes separately, combines the nonlinear least squares method to optimize core parameters such as activation energy and frequency factor, and dynamically corrects the aging rate through a temperature compensation mechanism. Based on the cycle test data of multi-specification batteries in the 25 - 35 °C environment, an interactive platform supporting visual input and dynamic prediction is developed, which can output the remaining capacity, health state, and sensitivity analysis results in real time, solving the technical bottlenecks of traditional models in cross-capacity adaptability and temperature response accuracy, and providing an intelligent solution for the full life cycle management of batteries.

[0025] The beneficial effects of the present invention include:

[0026] The modeling method proposed by the present invention shows significant advantages in the field of battery aging prediction and management, especially in enhancing model adaptability, strengthening physical interpretability, and expanding the scope of practical applications. By introducing a positive and negative electrode separation modeling structure, this method can meticulously depict the aging behaviors of different electrode materials, enabling the prediction of capacity attenuation to not only have high precision but also reflect the performance evolution at the material level. Secondly, this method has cross-capacity compatibility. By inputting the design capacity parameters of the positive and negative electrodes, the model can complete predictions without retraining when facing different specifications of battery cells or battery packs. This characteristic greatly improves the reuse efficiency of the model and reduces the development and verification costs for battery manufacturing and integration enterprises. For the operation and management of battery systems, the model of the present invention can not only identify potential degradation trends in the early stage but also dynamically adjust the aging rate in combination with the actual operating temperature and working conditions. This enables battery management systems in scenarios such as electric vehicles and grid energy storage to achieve higher-precision health state assessment and life prediction, thereby optimizing maintenance strategies and scheduling plans, extending the service life of the system, and enhancing the overall operational economy. Description of the Drawings

[0027] By describing the embodiments of the present application in more detail in conjunction with the drawings, the above and other objects, features, and advantages of the present application will become more apparent. The drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0028] Figure 1 It is a flowchart of a battery life prediction method that can be compatible with different design capacities according to an embodiment of the present application.

[0029] Figure 2Test examples of the battery and the capacity aging curves of all batteries for the battery life prediction method compatible with different design capacities according to the embodiments of the present application.

[0030] Figure 3 The software login interface and operation interface developed for the battery life prediction method compatible with different design capacities according to the embodiments of the present application.

[0031] Figure 4 The life prediction diagram of the 280Ah energy storage battery cell for the battery life prediction method compatible with different design capacities according to the embodiments of the present application.

[0032] Figure 5 The life prediction diagram of the 50Ah energy storage battery cell for the battery life prediction method compatible with different design capacities according to the embodiments of the present application.

[0033] Figure 6 The life prediction diagram of the 40Ah energy storage battery cell for the battery life prediction method compatible with different design capacities according to the embodiments of the present application.

[0034] Figure 7 The heat map of the correlation (sensitivity) analysis between life prediction and 6 temperature test sites for the battery life prediction method compatible with different design capacities according to the embodiments of the present application. Detailed implementation manners

[0035] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0036] The special term "exemplary" here means "serving as an example, embodiment or illustration". Any embodiment described as "exemplary" here does not have to be construed as superior to or better than other embodiments.

[0037] In addition, in order to better illustrate the present application, numerous specific details are given in the following detailed implementation manners. Those skilled in the art should understand that the present application can also be implemented without some specific details. In some instances, methods, means, elements and circuits well known to those skilled in the art are not described in detail so as to highlight the gist of the present application.

[0038] In addition, the terms "first" and "second" are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of this application, "a plurality" means two or more, unless otherwise specifically defined.

[0039] Different from traditional life prediction methods, the present invention fully considers the influence of the difference in design capacity among different battery specifications on the aging rate, and widely adapts the model through parameter normalization technology. The temperature response function embedded in the model combines empirical data with physical mechanisms, which can accurately reflect the modulation effect of the environment and working conditions on the battery attenuation rate, and significantly improves the prediction accuracy and stability under actual working conditions. This method first obtains the cycle data of the battery under different temperature, rate, and capacity conditions through a standardized test process, and uses machine learning optimization algorithms to fit the model parameters. The model output can be updated in real time, supporting the dynamic prediction of the remaining useful life (RUL) and state of health (SOH) of the battery throughout its life cycle. To improve the practicality of the model and the user experience, the present invention has developed a set of user interfaces with visual operations. This interface supports users to input battery operation data, set initial capacity and temperature parameters, and the system will automatically complete data preprocessing, model fitting, and life prediction analysis. The platform also integrates various chart generation functions, which can intuitively display prediction curves, key parameter trends, and aging risk assessment results, facilitating users to quickly grasp the battery health status and degradation path. The life prediction method proposed by the present invention has extremely high generality and expandability, and is applicable to battery management systems (Battery Management System, BMS) with different capacity specifications in various scenarios such as consumer electronics, electric vehicles, and energy storage power stations. It can not only provide a basis for battery manufacturers to optimize product design, but also provide strong technical support for safety control, pre-maintenance planning, and energy efficiency management in terminal applications. The proposal of the present invention provides a systematic and intelligent solution for the full life cycle management of batteries, and is expected to promote the development of the energy system towards a more efficient and intelligent direction, with significant engineering value and broad market application prospects.

[0040] The present invention will be further described in detail below by combining the accompanying drawings and specific implementation methods:

[0041] Figure 1 It is a flowchart of a battery life prediction method compatible with different design capacities according to an embodiment of the present application. As Figure 1As shown, the battery life prediction method compatible with different design capacities according to the embodiments of the present application includes aging data collection, adding prior knowledge (positive and negative electrode design capacities), extracting data such as capacity and temperature, building an Arrhenius model, model training and testing, user interface, and automated processing. Specifically, it includes: S101, performing long-term cycle tests on lithium-ion batteries with multiple design capacities to obtain discharge capacities, voltage curves, and temperature data at different temperatures, rates, and cycle numbers; S102, constructing a first-order kinetic reaction model based on the separation of the positive and negative electrodes, respectively constructing positive and negative electrode capacity attenuation equations, and incorporating the influence of temperature on the aging rate into the model through the Arrhenius equation, with the total capacity expressed as the sum of the capacity attenuation contributions of the positive and negative electrodes; S103, using the nonlinear least squares method combined with the Levenberg-Marquardt algorithm to optimize and fit the frequency factors and activation energies of the positive and negative electrodes, minimizing the sum of the squared residuals between the model predicted values and the measured values; S104, based on the trained model, inputting the positive and negative electrode design capacities and operating condition parameters of the target battery, dynamically predicting the evolution trend of the remaining capacity with the number of cycles, and generating a health state assessment result; S105, providing a visualization operation platform for data input, model calculation, and result display.

[0042] In the embodiments of the present application, in step S101, long-term cycle tests are performed on lithium-ion batteries with multiple design capacities to obtain discharge capacities, voltage curves, and temperature data at different temperatures, rates, and cycle numbers. It should be understood that to accurately model the life of lithium-ion batteries of various specifications, systematic data collection for batteries with different design capacities is crucial. At this stage, commercial energy storage lithium-ion battery samples with capacities of 280Ah, 50Ah, and 40Ah are selected, and long-term cycle tests are carried out on a unified test platform to ensure the consistency and comparability of the data. During the test process, using high-precision test instruments, under constant temperature and constant current / constant voltage conditions, the discharge capacity, charge-discharge voltage curve, and battery case temperature after each cycle are recorded. To cover the usage characteristics of different batteries in various application scenarios, multiple test conditions are also set, including different rates, different ambient temperatures, and various charge-discharge strategies.

[0043] To ensure that the model has good generalization ability among batteries with different capacities, special attention is paid to the influence of the initial design capacity of the battery on the aging behavior during data acquisition, such as the modulation effect of capacity on the internal resistance growth rate and the voltage platform attenuation rate. At the same time, to support subsequent parameter normalization and model training, the factory nominal parameters, measured initial capacity, and corresponding rate stability of each battery are recorded. By systematically tracking the entire aging process of batteries with different design capacities, not only can a training data set for cross-capacity models be constructed, but also a solid data foundation is provided for subsequent in-depth analysis of the coupling effect of capacity on the aging mechanism.

[0044] Specifically, in the data collection stage, the Xunpai battery module test system was used to conduct full-life cycle tests on multiple groups of lithium-ion batteries to obtain high-quality data required for model training and verification. A total of 40 lithium-ion battery monomers with different design capacities were selected in this experiment, covering multiple capacity levels. All samples passed pre-treatment screening to ensure their initial performance was consistent and there were no manufacturing defects. The tests were carried out in a laboratory environment with constant temperature and humidity, and temperature gradient tests were conducted at 25, 30, and 35 °C. Each battery performed standard charge and discharge cycles during the test, including constant current charging, constant voltage charging, rest, constant current discharging, and another rest stage. All operations were automatically completed through the preset program of the Xunpai system.

[0045] To fully capture the performance changes that the battery may experience during actual use, at least 1000 complete charge and discharge cycles were implemented for each group of batteries, and test extensions were carried out under multiple typical working conditions. The test system has high-precision voltage and current measurement and control capabilities and multi-channel synchronous test functions, ensuring the accuracy and timeliness of the data. Accordingly, the test examples of the battery and the test data of all batteries are as Figure 2 shown.

[0046] Furthermore, in an example of the present application, in step S101, the design capacities of the lithium-ion batteries include 280 Ah, 50 Ah, and 40 Ah, and the test environment temperature is at least one of 25 °C, 30 °C, and 35 °C. When the battery temperature exceeds 55 °C, a high-temperature warning is triggered, and when it reaches 70 °C, the test is forced to abort.

[0047] In the embodiment of the present application, in step S102, a first-order kinetic reaction model based on the separation of the positive and negative electrodes is constructed. The capacity decay equations of the positive and negative electrodes are respectively constructed, and the influence of temperature on the aging rate is incorporated into the model through the Arrhenius equation. The total capacity is expressed as the sum of the capacity decay contributions of the positive and negative electrodes.

[0048] Specifically, in the model establishment stage, based on the essential process of the capacity decay of lithium-ion batteries, a first-order kinetic reaction model was established to describe the aging behavior of the battery during cycling. The decay rate of the battery capacity is considered to be proportional to the current remaining capacity to reflect the characteristic of the gradual loss of materials during use.

[0049] The basic form of the model is:

[0050]

[0051] Among them, \(C(t)\) is the remaining discharge capacity of the battery at the \(t\)-th cycle, and \(k\) is the first-order reaction rate constant, which determines the speed of capacity decay. As the battery continues to be used, its effective active materials undergo slow and irreversible losses, and the capacity shows an exponential decline trend accordingly. It can be seen that the application of the Arrhenius equation to battery aging modeling conforms to physical intuition and mathematical logic.

[0052] By separating variables and integrating both sides of the above differential equation, the analytical formula of this model is obtained:

[0053] \(C(t)=C(0)\cdot e\) -kt

[0054] Considering the significant influence of temperature on the aging rate, the rate constant \(k\) is expressed as a function of temperature, and the Arrhenius relationship is introduced:

[0055]

[0056] Among them, \(A\) is the frequency factor; \(E\) a is the activation energy, representing the energy barrier that needs to be overcome for the reaction to occur; \(R\) is the ideal gas constant; \(T\) is the working temperature of the battery (unit: K). To further improve the model's ability to depict the actual battery aging process, based on the first-order kinetics modeling, the idea of separating the positive electrode from the negative electrode is introduced. That is: the decay of the total battery capacity is regarded as the result of the combined action of the independent aging processes of the positive electrode and negative electrode materials. This method can more carefully reflect the contributions of different electrode materials to capacity decline and enhance the physical interpretability of the model.

[0057] During the aging process, the positive electrode active materials will undergo failure mechanisms such as structural collapse and metal dissolution, resulting in a decrease in the ability to release reversible lithium ions; while the negative electrode materials may experience phenomena such as SEI film thickening and lithium dendrite deposition, hindering the insertion and extraction of lithium ions. Based on this, first-order decay models are established for both electrodes respectively:

[0058] Positive electrode aging reaction part:

[0059]

[0060] Negative electrode aging reaction part:

[0061]

[0062] Among them, and respectively represent the available capacities contributed by the positive electrode / negative electrode at the \(t\)-th cycle; and are the initial (designed) capacities of the positive electrode / negative electrode respectively; \(A\) (+) and \(A\) (-) are the frequency factors of the positive electrode / negative electrode aging reactions; and represent the aging activation energy of the positive / negative electrode, respectively.

[0063] Considering that the effective capacity of a lithium-ion battery is limited by the relatively "shorter board" in the positive and negative electrodes, the total capacity in the model is simplified as the sum of the aging contributions of the positive and negative electrodes. By combining the positive and negative electrode aging models, the total capacity prediction is obtained. The modeling logic is implemented through the Python programming language, and the core calculation module is written using the NumPy and SciPy mathematical libraries, supporting batch data processing and matrix fitting operations. This module is encapsulated as an independent function interface and can be flexibly integrated with the data preprocessing system and the parameter optimization module.

[0064] In the embodiment of the present application, in step S103, the frequency factors and activation energies of the positive and negative electrodes are optimized and fitted by using the non-linear least squares method combined with the Levenberg-Marquardt algorithm, minimizing the sum of the squared residuals between the model prediction value and the measured value. In the parameter estimation stage, around the constructed first-order reaction model for separating the positive and negative electrodes, the non-linear least squares method is used to fit the core parameters involved in the model, including the frequency factors and activation energies of the positive and negative electrode aging reactions. In the specific implementation process, based on the battery capacity data collected from experiments, by minimizing the error between the model prediction value and the measured value, the parameter combination is iteratively optimized to ensure that the obtained parameters can truly reflect the material aging behavior. In order to improve the credibility of the estimation results, residual analysis, goodness-of-fit evaluation and cross-validation are performed on the fitting results, and confidence interval calculation and parameter sensitivity analysis are introduced to systematically evaluate the influence of each parameter on the model stability and output results. The finally obtained parameters not only have high numerical accuracy but also have good physical interpretability, and can be widely applied to the life prediction of batteries with different working conditions and various design capacities.

[0065] Parameter estimation is a crucial step in this aging modeling system, and its goal is to determine the key kinetic parameters involved in the model, including the frequency factors, activation energies, etc. of the positive and negative electrode reaction processes, as well as the temperature-dependent reaction rate constant jointly determined by them. These parameters not only determine the prediction accuracy of the model but also directly reflect the inherent thermodynamics and kinetics characteristics during the aging process of the electrode material.

[0066] Parameter estimation is completed through a numerical optimization method based on experimental data. First, the capacity decay data under different temperature and cycle number conditions are extracted from battery samples with multiple design capacities, and compared point by point with the theoretical prediction values of the model to construct an objective error function:

[0067]

[0068] where N is the total number of test cycles, C measured,iis the measured discharge capacity value during the i-th cycle, C model,i is the predicted value of the model, the pre-exponential factor A and the activation energy E a are model parameters. To accurately identify the key parameters in the aging model, the Nonlinear Least Squares (NLS) method is used as the main parameter estimation method. This method minimizes the sum of the squared residuals between the model predicted values and the experimental observed values, thereby inversely calculating the optimal kinetic parameters. And, since it involves the superposition of exponential functions, it belongs to a typical nonlinear multi-parameter fitting problem. Therefore, the Levenberg-Marquardt algorithm (LMA) is used for solving. This algorithm combines the fast convergence characteristics of the Newton method and the robustness of the gradient descent method, and is especially suitable for nonlinear fitting of models with exponential terms. By initializing parameters, calculating the Jacobian matrix, adjusting the step size, updating parameters, and determining convergence and other steps, the optimal parameters are output.

[0069] In the embodiment of the present application, in step S104, based on the trained model, the designed capacities of the positive and negative electrodes of the target battery and the operating condition parameters are input, the evolution trend of the remaining capacity with the number of cycles is dynamically predicted, and a health state evaluation result is generated. After the aging model is trained and verified, this model can be widely applied to the life prediction and health state evaluation of batteries with different designed capacities. For lithium-ion batteries of various specifications, the designed capacities of the positive and negative electrodes (virtual capacities), a prior parameter, are introduced to enhance the adaptability and generalization ability of the model to different initial configurations of the batteries. In specific applications, first, according to engineering experience or data provided by battery manufacturers, the designed capacities of the positive and negative electrodes of the battery are entered, that is and The introduction of these designed capacities is equivalent to setting the "starting point" of the battery aging process for the model, that is: from what level of electrode material loading the battery starts to experience performance degradation. This setting is a key link to realize the compatibility prediction of the battery life under various designed capacity configurations by the model.

[0070] Based on the trained aging kinetic model, combined with the input designed capacity parameters, the model will automatically simulate and predict the evolution process of the remaining capacity of the battery under specified working temperatures and cycling conditions. Through the numerical solution process driven by machine learning algorithms, the system can quickly output the predicted capacity of the battery at any number of cycles or time nodes, and then estimate its service life and health state. Not only paying attention to the shape of the capacity degradation curve, but also further carrying out the analysis of the influence of key factors. Through the sensitivity analysis of temperature, rate, and material parameters, the influence degree of each factor on the battery degradation rate can be quantified. These analysis results have important engineering reference value for battery structure optimization, charge and discharge strategy formulation, and new material research and development.

[0071] In the embodiment of the present application, in step S105, a visualization operation platform is provided, which is used for data input, model calculation, and result display. To improve the practicability and operability of the model, a set of user-oriented visualization operation interfaces are developed. Users only need to input the designed capacity, operating temperature, and test data of the target battery, and the system can automatically complete data preprocessing, model calculation, and result display. The platform is also equipped with a graphical analysis module, which supports generating capacity-time curves, aging rate graphs, parameter sensitivity heat maps, etc., to help users intuitively understand the whole process of battery degradation.

[0072] Through the above method, the aging model proposed by the present invention not only has good accuracy and physical interpretation ability, but also realizes the compatibility prediction of batteries with various designed capacities, has broad engineering application prospects and product integration potential, and provides strong theoretical and technical support for battery health management and life optimization.

[0073] In the embodiment of the present application, a model verification step is further included: dividing the test data into a training set and a verification set, and evaluating the generalization ability of the model through residual analysis and cross-validation.

[0074] In the prediction and analysis stage, a battery aging analysis platform based on a graphical interface is developed, which integrates the first-order reaction model with separated positive and negative electrodes, enabling users without programming background to complete capacity degradation prediction. Users can import battery test data through the interface, input the designed capacity and environmental parameters, and the system will automatically perform data preprocessing, model calculation, and visualization result display. The platform is developed based on Python, uses PyQt5 to build the front-end interface, and utilizes Matplotlib and Seaborn to realize the dynamic generation of capacity-cycle curves, error analysis graphs, and sensitivity heat maps. Users can intuitively view the aging trend and key influencing factors, providing effective support for battery design optimization and life management.

[0075] The software login interface and operation interface developed by this method are as Figure 3 shown.

[0076] For the 280Ah energy storage battery cell, the prediction effect by this method is as Figure 4 shown (only part of the battery cells are shown).

[0077] For the 50Ah energy storage battery cell, the prediction effect by this method is as Figure 5 shown (only part of the battery cells are shown).

[0078] For the 40Ah energy storage battery cell, the prediction effect by this method is as Figure 6 shown (only part of the battery cells are shown).

[0079] The correlation (sensitivity) analysis of life prediction with 6 temperature test sites is asFigure 7 as shown (only part of the battery cells are shown).

[0080] The present invention aims to provide a capacity degradation modeling method for batteries with different designed capacities, and solve the deficiencies existing in the existing aging prediction models in terms of adaptability, physical interpretability, and parameter transferability. This method takes the independent aging behaviors of the positive and negative electrode materials of the battery as the starting point, and constructs a separate modeling framework based on the first-order kinetic reaction theory to accurately simulate the capacity change trajectory of the battery during long-term use. Different from traditional models, the present invention separately models the degradation processes of the positive and negative electrode materials for the first time, and combines them mathematically as a time function of the total capacity, so as to achieve a more refined expression of the battery aging behavior. By introducing the designed capacities of the positive and negative electrodes as prior inputs, the model can identify the structural starting point of battery aging, and dynamically adjust the expression of the aging rate function accordingly. This design enables the model to have the adaptability to batteries with multiple designed capacities, breaking through the limitation that traditional models need to separately train multiple capacity levels. In addition, to comprehensively capture the modulation effect of external working conditions on the aging process, the present invention introduces the Arrhenius reaction mechanism in the aging rate modeling, and takes temperature as the exponential function parameter of the rate constant. This method not only enhances the model's ability to express temperature sensitivity, but also improves its prediction stability in complex environments. In terms of parameter solution, the present invention adopts a non-linear least squares optimization algorithm to fit the key parameters such as activation energy and frequency factor involved in the model. Through repeated comparison and iteration with the measured data, it is ensured that the obtained parameters have both numerical stability and physical rationality. After training, the model can be used to simulate the future capacity evolution of the battery under any temperature and cycling conditions, and has high interpretability and engineering practical value.

[0081] To further improve the practicality of the model, the present invention has developed a visual analysis interface, enabling users to conveniently input the designed capacity and working condition data, and automatically obtain the capacity prediction results, aging trend curves, and related key indicators. The system also integrates a sensitivity analysis module, which can be used to evaluate the influence degree of each parameter on the aging results, and provide decision-making support for battery structure optimization and operation strategy formulation.

[0082] In summary, the capacity aging modeling method proposed by the present invention is superior to traditional models in terms of accuracy, applicable range, and interaction performance, and is applicable to multiple application scenarios such as battery R & D, management system integration, and product reliability assessment, and has significant theoretical significance and promotion value.

[0083] The embodiments of the present disclosure have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to technologies in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein.

Claims

1. A battery life prediction method compatible with different design capacities, characterized in that, Including: S101, conduct long-term cycling tests on lithium-ion batteries with multiple design capacities, and obtain discharge capacities, voltage curves, and temperature data at different temperatures, rates, and number of cycles; S102, construct a first-order kinetic reaction model based on the separation of the positive and negative electrodes, respectively construct the capacity attenuation equations for the positive and negative electrodes, and incorporate the influence of temperature on the aging rate into the model through the Arrhenius equation. The total capacity is expressed as the sum of the capacity attenuation contributions of the positive and negative electrodes; S103, use the non-linear least squares method combined with the Levenberg-Marquardt algorithm to optimize and fit the frequency factors and activation energies of the positive and negative electrodes, and minimize the sum of the squared residuals between the model prediction values and the measured values; S104, based on the trained model, input the design capacities of the positive and negative electrodes and the operating condition parameters of the target battery, dynamically predict the evolution trend of the remaining capacity with the number of cycles, and generate a health state assessment result; S105, provide a visual operation platform for data input, model calculation, and result display.

2. The battery life prediction method compatible with different design capacities according to claim 1, characterized in that, In step S101, the design capacities of the lithium-ion batteries include 280 Ah, 50 Ah, and 40 Ah, and the test ambient temperature is at least one of 25°C, 30°C, and 35°C.

3. The battery life prediction method compatible with different design capacities according to claim 2, characterized in that, In step S101, when the battery temperature exceeds 55°C, a high-temperature warning is triggered, and when it reaches 70°C, the test is forced to abort.

4. The battery life prediction method compatible with different design capacities according to claim 3, wherein In step S102, the capacity attenuation equations for the positive and negative electrodes satisfy: The positive electrode capacity attenuation model is: The negative electrode capacity attenuation model is: Among them, and represent the available capacities contributed by the positive and negative electrodes under the t-th cycle, respectively; and are the initial design capacities of the positive and negative electrodes, respectively; A (+) and A (-) are the frequency factors of the aging reactions of the positive and negative electrodes; and represent the aging activation energies of the positive and negative electrodes, respectively.

5. The battery life prediction method compatible with different design capacities according to claim 4, wherein In step S103, the objective function for parameter estimation is: Among them, N is the total number of test cycles, and C measured,i is the measured discharge capacity value during the i-th cycle, and C model,i is the predicted value of the model, where the pre-exponential factor A and the activation energy E a are model parameters, and the optimal parameter combination is obtained through iterative optimization.

6. The battery life prediction method compatible with different design capacities according to claim 5, characterized in that, In step S104, the model is dynamically updated through a machine learning algorithm, and the prediction curve is adjusted by combining the real-time input temperature data.

7. The battery life prediction method compatible with different design capacities according to claim 6, characterized in that Step S105 includes: Support input of battery design capacity, temperature, and number of cycles; Automatically generate capacity-time curves and aging rate trend charts; Provide a temperature sensitivity heat map and analysis of the influence degree of key parameters.

8. The battery life prediction method capable of being compatible with different design capacities according to claim 1, characterized in that, The method further includes a model verification step: dividing the test data into a training set and a verification set, and evaluating the generalization ability of the model through residual analysis and cross-validation.

9. An electronic device, characterized in that, Including a memory and a processor, the memory stores an executable program, and when the processor runs the executable program, it implements the battery life prediction method compatible with different design capacities as described in any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, and when the computer-executable instructions are executed by the processor, they are used to implement the battery life prediction method compatible with different design capacities as described in any one of claims 1 to 8.

Images