SOC-based lithium ion power battery cycle service life estimation method

Abstract

The invention discloses an SOC-based lithium ion power battery cycle service life estimation method, which comprises the following steps: acquiring impedance data of a battery under different charge states and cycle times through an electrochemical impedance spectroscopy technology, and obtaining a capacitance characteristic initial data set containing an electric double-layer capacitor and a Faraday capacitor; performing statistical analysis on the relevance between the capacitance characteristics and the state of charge in each cycle index section in the independent change rule to obtain attenuation trends of the two capacitors along with the cycle index; updating the three-dimensional relation model according to a life estimation result, carrying out iterative optimization on model parameters by adopting real-time impedance spectroscopy data, and determining a capacitance characteristic change rule reflecting the latest aging mechanism; and obtaining the optimized three-dimensional relation model, and repeating the step of separating the electric double-layer capacitance and the Faraday capacitance for newly collected impedance spectrum data to obtain dynamically adjusted attenuation trend parameters.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery life prediction, and particularly relates to a method for estimating the cycle service life of lithium-ion power batteries based on SOC. Background Art

[0002] As a core component in the new energy field, the life of lithium-ion power batteries directly determines the reliability and economy of key applications such as electric vehicles and energy storage systems. The research on battery life is not only the cornerstone for promoting energy transformation but also a key link in achieving sustainable development. However, current methods for estimating battery life mostly rely on single parameters, such as capacity decay or internal resistance increase. These methods are often difficult to accurately reflect the true aging state of the battery in complex usage scenarios, especially the dynamic change characteristics under different states of charge (SOC) are difficult to capture. Existing solutions usually ignore the diversity of the internal electrochemical processes of the battery, resulting in limited accuracy and universality of life prediction.

[0003] In this field, the core challenge lies in how to comprehensively characterize the aging mechanism of the battery during cycling, especially the deep mechanism of the impact of SOC on life has not been fully understood. Among them, the capacitance characteristics, as an important indicator reflecting the battery health state, the correlation between its change law and SOC and the number of cycles still needs to be deeply explored. Different types of capacitors, such as electric double-layer capacitors and Faraday capacitors, have significant differences in their contributions during the aging process, but existing research lacks systematic discrimination and quantitative analysis of these capacitance parameters. In addition, the monitoring of the capacitance decay trend and its coupling relationship with the SOC usage history are also difficult to achieve accurate prediction due to the lack of effective modeling means. These technical factors have not been resolved, resulting in the dual problems of insufficient accuracy and excessive model complexity in life estimation in practical applications. Summary of the Invention

[0004] The present invention proposes a method for estimating the cycle service life of lithium-ion power batteries based on SOC to solve the problems existing in the above-mentioned prior art.

[0005] To achieve the above object, the present invention provides a method for estimating the cycle service life of lithium-ion power batteries based on SOC, including the following steps:

[0006] Collect impedance data of the battery under different states of charge and the number of cycles to obtain an initial data set of capacitance characteristics including electric double-layer capacitance and Faraday capacitance;

[0007] Separate the contributions of the electric double-layer capacitance and the Faraday capacitance according to the initial data set, fit the impedance spectrum data using an equivalent circuit model, and determine the independent change laws of the two capacitance characteristics under different states of charge;

[0008] Statistical analysis is carried out on the correlation between the capacitance characteristics and the state of charge within each cycle number segment of the independent change law, and the attenuation trends of the two capacitances with the cycle number are obtained;

[0009] A three-dimensional relationship model of capacitance characteristics, cycle number, and state of charge is established through the attenuation trend data;

[0010] The capacitance characteristic attenuation values at specific states of charge and cycle numbers are extracted from the three-dimensional relationship model, and weighted processing is performed on the differences between the electric double-layer capacitance and the Faraday capacitance to obtain the comprehensive attenuation trend parameter;

[0011] If the comprehensive attenuation trend parameter exceeds the preset threshold, the remaining battery life is predicted through the support vector regression algorithm, and the life estimation result based on the capacitance characteristics at the current cycle number is obtained;

[0012] The three-dimensional relationship model is updated according to the life estimation result, and the model parameters are iteratively optimized using real-time impedance spectrum data;

[0013] The optimized three-dimensional relationship model is obtained, and the dynamically adjusted attenuation trend parameter is obtained based on the newly collected impedance spectrum data;

[0014] The support vector regression algorithm is re-run through the dynamically adjusted attenuation trend parameter to obtain the updated predicted value of the remaining battery life.

[0015] Preferably, obtaining the initial data set of capacitance characteristics includes:

[0016] The impedance data of the battery at different states of charge and cycle numbers are collected through electrochemical impedance spectroscopy technology to obtain the initial data set;

[0017] The capacitance characteristic data including the electric double-layer capacitance and the Faraday capacitance are extracted from the initial data set to obtain the capacitance characteristic data subset.

[0018] Preferably, the determination of the independent change laws of the two capacitance characteristics under different states of charge includes:

[0019] An equivalent circuit model is loaded through the impedance spectrum data to obtain the fitted parameter set;

[0020] The contributions of the electric double-layer capacitance and the Faraday capacitance are extracted from the fitted parameter set to obtain the separated capacitance data set;

[0021] The separated capacitance data set is used to determine the characteristic parameters of the electric double-layer capacitance under different states of charge;

[0022] The distribution of the Faraday capacitance is calculated through the characteristic parameters to obtain its independent change curve under different states of charge;

[0023] If the independent change curve deviates from the preset threshold, the parameters of the equivalent circuit model are adjusted to obtain optimized capacitance characteristic data;

[0024] Based on the optimized capacitance characteristic data, the dynamic responses of the two capacitors during the change of the state of charge are judged to obtain the final independent change law.

[0025] Preferably, obtaining the attenuation trends of the two capacitors with the number of cycles includes:

[0026] By loading the data set corresponding to the number of cycles, the capacitance characteristic data within each cycle number segment are obtained, the distribution characteristics of the state of charge are extracted from the obtained capacitance characteristic data, the correlation parameters are obtained, the correlation parameters are processed by statistical analysis, the attenuation trend of the electric double layer capacitor is determined, and by calculating the dynamic response of the Faraday capacitor, its change law with the number of cycles is obtained;

[0027] If the change law deviates from the preset threshold, the parameters of the statistical analysis are adjusted to obtain the optimized attenuation trend. Based on the optimized attenuation trend, the independent distribution characteristics of the two capacitors in the state of charge are judged, and a complete data set is generated through the independent distribution characteristics to obtain the final attenuation trend.

[0028] Preferably, establishing the three-dimensional relationship model of capacitance characteristics, number of cycles and state of charge through the attenuation trend data includes:

[0029] The preliminary mapping relationship between capacitance characteristics and the number of cycles is generated through the attenuation trend data, and the mapping relationship is processed by the polynomial regression algorithm to obtain the preliminary characteristics of dynamic change;

[0030] The distribution parameters of the state of charge are extracted from the preliminary characteristics, the data basis of the three-dimensional relationship is determined according to the distribution parameters, the parameters of the polynomial regression are adjusted according to the three-dimensional relationship data, the optimized dynamic change characteristics are obtained, and the intermediate expression of the aging mechanism is calculated through the optimized dynamic change characteristics to obtain the three-dimensional relationship model of mathematical expression.

[0031] Preferably, obtaining the comprehensive attenuation trend parameters includes:

[0032] The attenuation values corresponding to the state of charge and the number of cycles are extracted through the three-dimensional relationship model, and the differences between the electric double layer capacitor and the Faraday capacitor are quantified by weighted processing to obtain the comprehensive attenuation parameters;

[0033] The time series distribution of the attenuation values is obtained from the comprehensive attenuation parameters, linear interpolation processing is performed on the change trend of the number of cycles to determine the continuous expression of the attenuation values, the dynamic distribution characteristics of the state of charge are extracted according to the continuous expression, and the characteristics are classified by the random forest algorithm to judge the contribution ratio of the electric double layer capacitor and the Faraday capacitor;

[0034] If the contribution ratio exceeds a preset threshold, recalculate the comprehensive decay parameter by adjusting the parameters of the weighted processing to obtain a corrected comprehensive decay trend.

[0035] Preferably, the obtaining of the life estimation result based on the capacitance characteristics at the current cycle number includes:

[0036] Process the comprehensive decay trend parameters through a support vector regression algorithm to obtain a preliminary predicted value of the remaining battery life, and perform time series correction on the preliminary predicted value using the cycle number to determine the corrected life estimation distribution;

[0037] Extract the change characteristics of the capacitance characteristics from the corrected life estimation distribution to obtain the dynamic distribution data of the characteristics. If the dynamic distribution data exceeds a preset threshold, perform quadratic fitting on the characteristics through a regression algorithm to obtain a corrected prediction result;

[0038] Analyze the correlation between the cycle number and the capacitance characteristics according to the corrected prediction result, judge the significance of the correlation, and perform weighted adjustment on the comprehensive decay trend parameters through the significance to obtain an adjusted trend parameter distribution;

[0039] Extract the life estimation value in the current state from the adjusted trend parameter distribution to determine the final predicted output.

[0040] Preferably, the determining of the change rule of the capacitance characteristics reflecting the latest aging mechanism includes:

[0041] Update the three-dimensional relationship model through real-time impedance spectrum data, adjust the model parameters using an iterative optimization method to obtain the characteristic distribution reflecting the current state, extract the correlation data between the capacitance characteristics and the change rule from the characteristic distribution to obtain the preliminary distribution of the dynamic rule;

[0042] If the preliminary distribution of the dynamic rule exceeds a preset threshold, fit the correlation data through a support vector regression algorithm to determine the corrected distribution data, analyze the change trend between the real-time impedance and the aging characteristics according to the corrected distribution data to obtain the adjusted value of the trend parameter, and perform secondary update on the parameters of the three-dimensional relationship model through the adjusted value to obtain the updated model parameter distribution;

[0043] Extract the dynamic rule data reflecting the capacitance characteristics from the updated model parameter distribution, judge the evolution direction of the aging characteristics, and perform time series correction on the characteristic distribution using the judgment result to determine the final evolution trend distribution.

[0044] Preferably, the obtaining of the dynamically adjusted decay trend parameter includes:

[0045] Update the three-dimensional model with newly collected impedance spectrum data, separate the double-layer capacitance and the Faraday capacitance, obtain the preliminary capacitance characteristic distribution, and extract the attenuation trend from the preliminary capacitance characteristic distribution;

[0046] Use the dynamic adjustment method to determine the adjusted trend parameters. For the adjusted trend parameters, verify the accuracy of real-time processing through impedance spectrum data, obtain the verified parameter distribution. If the verified parameter distribution exceeds the preset threshold, then fit the attenuation trend through the support vector regression algorithm to obtain the corrected trend data;

[0047] Update the three-dimensional model according to the corrected trend data, obtain the optimized model parameter distribution, extract the dynamically adjusted attenuation trend from the optimized model parameter distribution, judge the change direction of the capacitance characteristics, and correct the time series of data acquisition through the change direction to determine the final trend parameter distribution.

[0048] Preferably, the obtaining the updated predicted value of the remaining battery life includes:

[0049] Run the support vector regression algorithm through the attenuation trend data, obtain the preliminary predicted value of the remaining life, extract the aging mechanism characteristics according to the preliminary predicted value of the remaining life, determine the aging characteristic distribution, and analyze the change of the state of charge through the aging characteristic distribution to obtain the preliminary distribution of the coupling relationship;

[0050] If the preliminary distribution of the coupling relationship exceeds the preset threshold, then correct the change trend through the support vector regression algorithm to obtain the adjusted distribution data, update the attenuation trend according to the adjusted distribution data, judge the change direction of the aging mechanism, and correct the time series of the state of charge through the change direction to determine the final coupling relationship distribution;

[0051] Update the predicted value of the remaining life according to the final coupling relationship distribution to obtain the optimized prediction result.

[0052] Compared with the prior art, the present invention has the following advantages and technical effects:

[0053] The present invention discloses a method for estimating the cycle service life of a lithium-ion power battery based on SOC. By collecting impedance data under different states of charge and cycle numbers, separating the double-layer capacitance and the Faraday capacitance, a three-dimensional relationship model of capacitance characteristics, cycle numbers, and state of charge is established. This method uses the polynomial regression algorithm to quantify the dynamic change characteristics, determine the mathematical expression of the aging mechanism, and predicts the remaining battery life through the support vector regression algorithm. The present invention can also update the model parameters in real time and dynamically adjust the attenuation trend, so as to accurately reflect the latest aging mechanism. This method can effectively evaluate the battery performance attenuation, realize accurate life prediction, and provide an important basis for optimizing the battery management system and extending the service life. Description of the Drawings

[0054] The drawings forming a part of this application are used to provide a further understanding of this application. The schematic embodiments and descriptions thereof of this application are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0055] Figure 1 is a flowchart of the method according to an embodiment of the present invention;

[0056] Figure 2 is a schematic diagram of the method according to an embodiment of the present invention. Detailed Embodiments

[0057] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine with the embodiments to detail this application.

[0058] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0059] Embodiment 1

[0060] As Figure 1-2 shown, in this embodiment, a method for estimating the cycle service life of a lithium-ion power battery based on SOC is provided, including the following steps:

[0061] Collect impedance data of the battery at different state of charge and cycle numbers to obtain an initial data set of capacitance characteristics including double-layer capacitance and Faraday capacitance;

[0062] Separate the contributions of double-layer capacitance and Faraday capacitance according to the initial data set, fit the impedance spectrum data using an equivalent circuit model, and determine the independent variation laws of the two capacitance characteristics at different state of charge;

[0063] Conduct a statistical analysis on the correlation between the capacitance characteristics and the state of charge within each cycle number segment in the independent variation laws to obtain the attenuation trends of the two capacitances with the cycle numbers;

[0064] Establish a three-dimensional relationship model of capacitance characteristics, cycle numbers, and state of charge through the attenuation trend data;

[0065] Extract the capacitance characteristic attenuation values at specific state of charge and cycle numbers from the three-dimensional relationship model, and perform weighted processing on the differences between double-layer capacitance and Faraday capacitance to obtain a comprehensive attenuation trend parameter;

[0066] If the comprehensive attenuation trend parameter exceeds the preset threshold, the remaining battery life is predicted by the support vector regression algorithm to obtain the life estimation result based on the capacitance characteristics at the current cycle number;

[0067] Update the three-dimensional relationship model according to the life estimation result, and iteratively optimize the model parameters using real-time impedance spectrum data;

[0068] Obtain the optimized three-dimensional relationship model, and obtain the dynamically adjusted attenuation trend parameter based on the newly collected impedance spectrum data;

[0069] Rerun the support vector regression algorithm with the dynamically adjusted attenuation trend parameter to obtain the updated predicted value of the remaining battery life.

[0070] Specifically, it includes the following steps:

[0071] S101. Collect impedance data of the battery at different states of charge and cycle numbers through electrochemical impedance spectroscopy technology to obtain an initial data set. Extract capacitance characteristic data including double-layer capacitance and Faraday capacitance from the initial data set to obtain a capacitance characteristic subset.

[0072] Specifically, collecting impedance data of the battery at different states of charge and cycle numbers through electrochemical impedance spectroscopy technology is one of the core methods in battery performance analysis. Electrochemical impedance spectroscopy can reflect the electrochemical reaction characteristics inside the battery, such as charge transfer resistance, double-layer capacitance, and the dynamic changes of the diffusion process.

[0073] Exemplarily, when testing a lithium-ion battery, the states of charge can be set to 20%, 50%, and 80%, corresponding to low, medium, and high battery charge intervals respectively. At the same time, the cycle numbers are set to 0 times, 100 times, and 500 times to simulate the process of the battery from new to aged. Through an AC excitation with a frequency range from 0.01 Hz to 100 kHz, the real and imaginary part data of the impedance are obtained to form an initial data set. The advantage of this method is that it can comprehensively capture the performance degradation characteristics of the battery at different usage stages and provide data support for subsequent optimization. Extracting capacitance characteristic data including double-layer capacitance and Faraday capacitance from the initial data set is an effective way to further focus on the electrochemical behavior of the battery. The double-layer capacitance mainly reflects the charge storage ability on the electrode surface, while the Faraday capacitance is closely related to the redox reaction of the electrode material.

[0074] Specifically, the Faraday capacitance information can be extracted from the straight-line part in the low-frequency region of the Nyquist plot of the impedance spectrum, while the semi-circle in the middle-frequency region corresponds to the double-layer capacitance.

[0075] For example, in a certain experiment, the double-layer capacitance value at the initial cycle may be 50 μF / cm 2, as the number of cycles increases to 500 times, it drops to 30 μF / cm 2 , indicating a decrease in the surface activity of the electrode. This extraction process helps to quantify the impact of battery aging on capacitance characteristics and provides a basis for material improvement.

[0076] In this embodiment, after obtaining the capacitance characteristic subset, it can be analyzed in combination with the actual battery operation scenario.

[0077] For example, for power batteries used in electric vehicles, assuming that at a 50% state of charge, the double-layer capacitance of the initial cycle is relatively high, indicating a stable interface between the electrode and the electrolyte. However, after 300 cycles, the capacitance value drops by 20%, which may be caused by the thickening of the solid electrolyte interface film. This can be verified by comparing the changes in the impedance spectrum shape at different cycle numbers. For example, an increase in the diameter of the semicircle reflects an increase in the interface resistance.

[0078] Preferably, this analysis can also guide the battery management system to adjust the charge and discharge strategy and extend the service life.

[0079] It should be noted that the extraction of double-layer capacitance and Faraday capacitance is not carried out in isolation, but supports each other to form a complete description of electrochemical characteristics.

[0080] In this embodiment, if a certain type of lithium-ion battery is tested and it is found that the Faraday capacitance decreases significantly at a high state of charge, it may be related to the collapse of the structure of the positive electrode material. At this time, combining the double-layer capacitance data, if its value remains stable, it can be inferred that the problem mainly lies in the active material rather than the electrode interface. This multi-faceted analysis can improve the diagnostic accuracy.

[0081] It can be understood that this method is not only applicable to laboratory research, but also can provide technical support for quality control in industrial production.

[0082] For example, in the battery R & D stage, by analyzing the capacitance characteristics of electrodes with different formulations through the above method, materials with better cycle resistance can be screened out.

[0083] In this embodiment, for a certain electrode formulation at an 80% state of charge, after 500 cycles, the Faraday capacitance only drops by 10%, while for another formulation it drops by 25%. Obviously, the former has more advantages. The benefit of this comparative analysis is to accelerate the R & D process and reduce the trial-and-error cost. In short, by obtaining and analyzing capacitance characteristic data through electrochemical impedance spectroscopy technology, the internal mechanism of battery performance changes can be revealed from multiple dimensions, providing a solid foundation for improving battery design and application efficiency.

[0084] S102. Separate the contributions of double-layer capacitance and Faraday capacitance from the initial data set, fit the impedance spectrum data using an equivalent circuit model, and determine the independent variation laws of the two capacitance characteristics at different states of charge.

[0085] Load an equivalent circuit model with impedance spectrum data to obtain a set of fitted parameters. Extract the contributions of the double-layer capacitance and the Faraday capacitance from the fitted parameter set to obtain a separated capacitance data set. Use the separated capacitance data set to determine the characteristic parameters of the double-layer capacitance at different state of charge. Calculate the distribution of the Faraday capacitance through the characteristic parameters to obtain its independent change curve at different state of charge. If the independent change curve deviates from a preset threshold, adjust the parameters of the equivalent circuit model to obtain optimized capacitance characteristic data. Based on the optimized capacitance characteristic data, judge the dynamic response of the two capacitances when the state of charge changes to obtain the final independent change law.

[0086] Specifically, load an equivalent circuit model with impedance spectrum data to obtain a set of fitted parameters.

[0087] It can be understood that the core of this step is to convert complex impedance spectrum data into analyzable circuit element parameters. For example.

[0088] In this embodiment, an equivalent circuit model including resistors, capacitors and inductors can be selected. Input the impedance spectrum data of the battery into the model, and obtain the parameter values of each element through software fitting. For example, the resistance may be 50 milliohms and the capacitance may be 200 microfarads. The advantage of this method is that it can convert abstract spectrum data into intuitive physical meanings, which is convenient for subsequent analysis. Extract the contributions of the double-layer capacitance and the Faraday capacitance from the fitted parameter set to obtain a separated capacitance data set.

[0089] Specifically, the contributions of the two types of capacitances can be distinguished according to the behaviors of different elements in the circuit model.

[0090] Exemplarily, the double-layer capacitance is usually related to the fast charge and discharge on the electrode surface, showing a response at higher frequencies, while the Faraday capacitance is related to the electrochemical reaction and is more reflected in the low-frequency region.

[0091] In this embodiment, assume that the capacitance measured in the high-frequency region is 150 microfarads, which can be attributed to the double-layer capacitance, and the additional 50 microfarads in the low-frequency region are attributed to the Faraday capacitance. This separation helps to more clearly understand the roles of the two capacitances. Use the separated capacitance data set to determine the characteristic parameters of the double-layer capacitance at different state of charge.

[0092] It should be noted that the characteristic parameters may include the slope of the capacitance value changing with voltage or the response time.

[0093] For example, when the state of charge increases from 20% to 80%, the double-layer capacitance may increase from 140 microfarads to 160 microfarads, with a slope of 0.33 microfarads / %. This analysis can reflect the dynamic characteristics of the capacitance during battery operation and provide a basis for optimizing the design. By calculating the distribution of the Faraday capacitance through characteristic parameters, an independent variation curve thereof under different states of charge is obtained.

[0094] Preferably, a curve graph can be drawn by fitting parameters.

[0095] In this embodiment, when the state of charge is 30%, the contribution of the Faraday capacitance is 40 microfarads, and it increases to 60 microfarads at 70%, showing a non-linear growth. This distribution reveals the variation law of the Faraday capacitance with the intensity of the electrochemical reaction and helps to predict the battery performance. If the independent variation curve deviates from a preset threshold, the parameters of the equivalent circuit model are adjusted to obtain optimized capacitance characteristic data.

[0096] For example, assume that the preset threshold is that the slope of the curve is less than 0.5, while the actually measured value is 0.7. Then, an inductance element in the model can be increased or the resistance value can be adjusted, and refitting is performed until the curve meets the expectation. This adjustment can improve the accuracy of the data and ensure the reliability of the analysis results. According to the optimized capacitance characteristic data, the dynamic responses of the two capacitances during the change of the state of charge are judged to obtain the final independent variation law.

[0097] In this embodiment, it can be observed that the double-layer capacitance tends to be stable near the state of charge of 50%, while the Faraday capacitance increases significantly at high states of charge. This indicates that their response mechanisms are different. The former depends more on the surface effect, and the latter is related to the reaction depth. Obtaining this law helps to improve the accuracy of the battery management strategy.

[0098] S103. Perform a statistical analysis on the correlation between the capacitance characteristics and the state of charge in each cycle number segment of the independent variation law to obtain the attenuation trends of the two capacitances with the cycle number.

[0099] By loading the data set corresponding to the cycle number, obtain the capacitance characteristic data in each cycle number segment, extract the distribution characteristics of the state of charge from the obtained capacitance characteristic data, obtain the correlation parameter, process the correlation parameter by statistical analysis, determine the attenuation trend of the double-layer capacitance, obtain its variation law with the cycle number by calculating the dynamic response of the Faraday capacitance; if the variation law deviates from the preset threshold, adjust the parameters of the statistical analysis to obtain the optimized attenuation trend, and judge the independent distribution characteristics of the two capacitances under the state of charge according to the optimized attenuation trend, and generate a complete data set through the independent distribution characteristics to determine the final attenuation trend.

[0100] Specifically, by loading the dataset corresponding to the number of cycles, it can be understood as obtaining a set of capacitance data that evolves over time from experiments or simulations.

[0101] For example, in a battery cycling test, assuming there are 1000 cycles, the corresponding voltage and current changes are recorded for each cycle, and then the capacitance characteristic data is deduced.

[0102] Exemplarily, these data may show that the capacitance value decreased by 5% at the 200th cycle and by 20% at the 800th cycle. This kind of data provides a basis for subsequent analysis. Extract the distribution characteristics of the state of charge from the obtained capacitance characteristic data.

[0103] Specifically, it can be achieved by statistically analyzing the charge accumulation within the voltage range.

[0104] In this embodiment, assume that the state of charge is divided into three ranges: low, medium, and high, and the capacitance value distributions within each range are statistically analyzed respectively.

[0105] For example, the electric double - layer capacitance dominates in the low state of charge, while the proportion of the Faraday capacitance increases in the high state of charge. This distribution characteristic reflects the dynamic change of capacitance with the state of charge. Use statistical analysis to process the correlation parameters.

[0106] It should be noted that this step aims to quantify the relationship between the features.

[0107] For example, the relationship between the number of cycles and capacitance decay can be analyzed using the correlation coefficient. Assume that the correlation coefficient is 0.9 at the 500th cycle, indicating a significant decay trend. This method helps to extract rules from the data. Determine the decay trend of the electric double - layer capacitance.

[0108] Preferably, it can be visually presented by fitting a curve.

[0109] For example, plot the change graph of the electric double - layer capacitance with the number of cycles and find that the decay is mild within the first 300 cycles and then accelerates. This trend helps to predict the long - term performance. By calculating the dynamic response of the Faraday capacitance.

[0110] In this embodiment, its peak change in each cycle can be observed.

[0111] For example, the peak response is 50 mF at the 100th cycle and drops to 30 mF at the 900th cycle. This change rule reveals its evolution characteristics with the cycle. If the change rule deviates from the preset threshold, for example, the preset decay does not exceed 15%, but actually reaches 18%, then adjust the parameters of the statistical analysis.

[0112] In this embodiment, the weight factor can be increased for re - analysis to ensure that the results are closer to the actual situation. This optimization improves the reliability of the data. The independent distribution characteristics of the two types of capacitors under the charged state are judged according to the optimized attenuation trend.

[0113] For example, the electric double - layer capacitor remains stable in the low - charged state, while the Faraday capacitor shows obvious attenuation in the high - charged state. This independence analysis helps to understand the action mechanisms of the two. By generating a complete data set based on the independent distribution characteristics, it can be understood as integrating all analysis results to form a comprehensive view.

[0114] For example, the attenuation data of 1000 cycles are summarized in a table to clearly show the changes of the two types of capacitors. This complete data set provides support for subsequent applications. Determine the final attenuation trend.

[0115] For example, it is finally obtained that the attenuation rate of the electric double - layer capacitor is about 2% per year, and that of the Faraday capacitor is 5%. This clear trend provides a basis for optimized design and helps to extend the service life.

[0116] S104. Establish a three - dimensional relationship model of capacitance characteristics, number of cycles, and charged state through the attenuation trend data, use the polynomial regression algorithm to quantify the dynamic change characteristics in the three - dimensional relationship, and determine the mathematical expression of the aging mechanism.

[0117] Generate a preliminary mapping relationship between capacitance characteristics and number of cycles through the attenuation trend data, use the polynomial regression algorithm to process the mapping relationship, obtain the preliminary characteristics of the dynamic change, extract the distribution parameters of the charged state from the preliminary characteristics, determine the data basis of the three - dimensional relationship, adjust the parameters of the polynomial regression according to the three - dimensional relationship data, obtain the optimized dynamic change characteristics, and calculate the intermediate expression of the aging mechanism through the optimized dynamic change characteristics to obtain the mathematical expression.

[0118] Specifically, when generating a preliminary mapping relationship between capacitance characteristics and number of cycles through the attenuation trend data.

[0119] It can be understood that this process aims to establish a basic model.

[0120] For example, it can be assumed that the capacitance of a certain capacitor decays to 90% of the initial value after 1000 cycles and drops to 70% at 5000 cycles. By recording these data points, the preliminary mapping relationship can be expressed in a simple two - dimensional coordinate form, intuitively reflecting the influence of the number of cycles on the capacitance characteristics. Using the polynomial regression algorithm to process the mapping relationship is to further capture the trend of non - linear changes.

[0121] In this embodiment, a quadratic or cubic polynomial can be selected to fit the data. For example, if the initial capacitance is set to 100 microfarads and the attenuation trend shows a characteristic of being gentle first and then accelerating, a smooth curve can be obtained through regression analysis. The advantage of this method is that it can more accurately reflect the dynamic changes, rather than a simple linear assumption. When extracting the distribution parameters of the state of charge from the preliminary characteristics.

[0122] It should be noted that the state of charge is usually closely related to the actual working state of the capacitor.

[0123] Exemplarily, assume that in a certain cycle, the capacitor shows stable voltage in the fully charged state, while there are slight fluctuations in the half-charged state. By statistically analyzing the voltage distribution characteristics in multiple cycles, distribution parameters such as the mean and variance can be extracted to quantify the changes in the state of charge.

[0124] Specifically, if the mean gradually shifts with the number of cycles, it may indicate a decrease in the stability of the internal materials of the capacitor. Such distribution parameters provide a data basis for constructing the three-dimensional relationship and help to more comprehensively understand the interactive effects between variables. When adjusting the parameters of the polynomial regression according to the three-dimensional relationship data.

[0125] Preferably, fine-tuning can be performed according to the deviation degree of the actual data.

[0126] For example, if it is found that in the high cycle number segment, the difference between the predicted value and the actual attenuation value is large, the order of the polynomial can be increased or the weight coefficient can be adjusted to make the curve closer to the real trend.

[0127] In this embodiment, if the original regression predicts that the capacitance after 5000 cycles is 75 microfarads, while the actual measurement is 70 microfarads, the parameters are iteratively optimized to reduce the error. Such optimized dynamic change characteristics can more realistically reflect the law of the capacitor's evolution over time and lay a foundation for subsequent analysis. When calculating the intermediate expression of the aging mechanism through the optimized dynamic change characteristics.

[0128] It can be understood that this expression aims to reveal the physical meaning behind the attenuation.

[0129] For example, in an electric double-layer capacitor, aging may be related to the pore blockage on the electrode surface, while in a Faraday capacitor, it may be related to the loss of active substances.

[0130] Specifically, it can be assumed that for a certain capacitor, after 3000 cycles, the internal resistance increases by 20%. Through characteristic analysis, the mathematical relationship between the resistance change and the capacitance attenuation is deduced. The advantage of such an intermediate expression is that it not only provides a mathematical description of the data but also provides clues for in-depth research on the aging mechanism.

[0131] In this embodiment, by comparing the resistance increments under different state of charge, it can be verified whether the aging is related to specific working conditions, thereby providing a basis for optimizing the capacitor design or usage strategy.

[0132] In this embodiment, the series analysis of these steps can also reveal potential performance bottlenecks.

[0133] For example, if it is found that the state of charge distribution parameters are abnormally concentrated at high cycle numbers, it may indicate that the capacitor ages rapidly under certain conditions. This kind of insight has practical value for extending the capacitor life or adjusting the working environment.

[0134] Preferably, through multi-faceted verification, such as combining the trends of voltage, resistance, and capacity decay, jointly support the inference of the aging mechanism to ensure the reliability of the analysis results.

[0135] S105: Extract the capacitance characteristic decay values at specific state of charge and cycle numbers from the three-dimensional relationship model, perform weighted processing on the differences between the electric double layer capacitor and the Faraday capacitor, and obtain the comprehensive decay trend parameter.

[0136] Extract the decay values corresponding to the state of charge and cycle numbers through the three-dimensional relationship model, use weighted processing to quantify the differences between the electric double layer capacitor and the Faraday capacitor, obtain the comprehensive decay parameter, obtain the time series distribution of the decay values from the comprehensive decay parameter, perform linear interpolation processing on the change trend of the cycle numbers, determine the continuous expression of the decay values, extract the dynamic distribution characteristics of the state of charge according to the continuous expression, classify the characteristics through the random forest algorithm, and judge the contribution ratio of the electric double layer capacitor and the Faraday capacitor; if the contribution ratio exceeds the preset threshold, recalculate the comprehensive decay parameter by adjusting the parameters of the weighted processing to obtain the corrected comprehensive decay trend.

[0137] Specifically, the key to extracting the decay values corresponding to the state of charge and cycle numbers through the three-dimensional relationship model lies in separating the key variables from the multi-dimensional data. For example.

[0138] In this embodiment, the capacitor test data can be grouped according to the cycle numbers, and each group of data records the decay values of the state of charge from 100% to 20%. Suppose the decay value of a certain battery is 5% at the 50th cycle and 8% at the 100th cycle, and the preliminary contour of the decay curve is constructed through these data points. This method helps to intuitively reflect the trend of decay with the increase of cycle numbers and lays a foundation for subsequent quantification. When using weighted processing to quantify the differences between the electric double layer capacitor and the Faraday capacitor.

[0139] It should be noted that the electric double layer capacitor mainly comes from physical adsorption, while the Faraday capacitor is related to chemical reactions.

[0140] Specifically, a weighting coefficient can be set. For example, the proportion of the electric double layer capacitance is 0.6, and the proportion of the Faraday capacitance is 0.4. By decomposing the attenuation value, the comprehensive attenuation parameter is calculated.

[0141] For example, in a certain test, the total attenuation value is 10%. After weighting, the electric double layer capacitance contributes 6%, and the Faraday capacitance contributes 4%. In this way, the influences of the two mechanisms can be clearly distinguished.

[0142] Preferably, by adjusting the coefficient through multiple experiments, the actual attenuation source can be more accurately reflected. The time series distribution of the attenuation value is obtained from the comprehensive attenuation parameter.

[0143] In this embodiment, the attenuation values of each cycle can be arranged in chronological order to form a sequence, such as 5%, 5.2%, 5.5%, etc. For the change trend of the number of cycles, linear interpolation is performed, that is, these discrete points are connected into a continuous curve.

[0144] For example, between the 50th and 51st cycles, the attenuation value smoothly transitions from 5% to 5.2%. This continuous expression is convenient for capturing the subtle changes in attenuation, especially suitable for long-term trend analysis. When extracting the dynamic distribution characteristics of the state of charge according to the continuous expression.

[0145] It can be understood that the state of charge will show a non-uniform distribution over time.

[0146] Exemplarily, assume that in a certain cycle segment, the state of charge rapidly drops from 80% to 50%. Its dynamic characteristics are characterized by statistical distribution parameters such as the mean and variance. Then, the features are classified by the random forest algorithm.

[0147] In this embodiment, the features can be input into the model, and the contribution ratios of the electric double layer capacitance and the Faraday capacitance are output, such as 55% and 45%. If the ratio exceeds the preset threshold, for example, the electric double layer capacitance exceeds 60%, the weighting coefficient is adjusted and recalculated.

[0148] For example, the electric double layer coefficient is adjusted from 0.6 to 0.55, the attenuation value is re-decomposed, and finally the corrected trend is obtained, such as the attenuation is adjusted from 10% to 9.5%. This iterative optimization method can improve the accuracy of the analysis.

[0149] In this embodiment, the significance of judging the contribution ratio is to reveal the capacitance aging mechanism.

[0150] For example, an increase in the proportion of the electric double layer capacitance may indicate the deterioration of the physical structure, while an increase in the Faraday capacitance may be related to the decomposition of the electrolyte. This classification result provides data support for subsequent optimization design.

[0151] Preferably, by verifying the ratio change through multiple groups of experiments, the internal law of attenuation can be more comprehensively grasped.

[0152] S106. If the comprehensive decay trend parameter exceeds the preset threshold, the remaining battery life is predicted by the support vector regression algorithm to obtain the life estimation result based on the capacitance characteristics at the current cycle number.

[0153] The comprehensive decay trend parameter is processed by the support vector regression algorithm to obtain a preliminary prediction value of the remaining battery life. The preliminary prediction value is corrected by the cycle number for time series to determine the corrected life estimation distribution. The change characteristics of the capacitance characteristics are extracted from the corrected life estimation distribution to obtain the dynamic distribution data of the characteristics. If the dynamic distribution data exceeds the preset threshold, the characteristics are secondarily fitted by the regression algorithm to obtain the corrected prediction result. The correlation between the cycle number and the capacitance characteristics is analyzed according to the corrected prediction result, and the significance of the correlation is judged. The comprehensive decay trend parameter is weighted and adjusted according to the significance to obtain the adjusted trend parameter distribution. The life estimation value in the current state is extracted from the adjusted trend parameter distribution to determine the final prediction output.

[0154] Specifically, when the comprehensive decay trend parameter is processed by the support vector regression algorithm, it can be regarded as a mapping means to convert the multi-dimensional data of the cycle number and the capacitance characteristic decay into a preliminary prediction value of the remaining life.

[0155] In this embodiment, it is assumed that the battery cycle number is 500 times, and the comprehensive decay trend parameter shows that the capacitance characteristic has decreased by 15%. The support vector regression can train the model based on historical data and predict that the remaining life is 2000 cycles. The core of this method is to use the kernel function to capture the non-linear relationship and is applicable to the complex changes of battery decay. When the preliminary prediction value is corrected by the cycle number for time series.

[0156] Preferably, the concept of a time window can be introduced.

[0157] Specifically, based on the decay trend of the last 100 cycles, the preliminary prediction value is weighted and adjusted to reduce the influence of short-term fluctuations.

[0158] Exemplarily, if the preliminary prediction value is 2000 times and the decay has accelerated in the last 50 cycles, the corrected life estimation may be adjusted to 1800 times. This correction method can improve the dynamic adaptability of the prediction. When the change characteristics of the capacitance characteristics are extracted from the corrected life estimation distribution.

[0159] It can be understood that the characteristics may include the slope of the decay rate or the fluctuation amplitude.

[0160] In this embodiment, if the estimated distribution shows that the life gradually decreases from 1,800 times, the extracted feature may be that the slope changes from 0.02 to 0.05, indicating an acceleration of attenuation. After obtaining the dynamic distribution data, if the slope exceeds the preset threshold of 0.04, a second-order fitting is required.

[0161] For example, after refitting through a regression algorithm, the corrected prediction result may be adjusted from 1,800 times to 1,700 times, improving the prediction accuracy. When analyzing the correlation between the number of cycles and the capacitance characteristics based on the corrected prediction result.

[0162] It should be noted that the significance level can be judged through statistical tests.

[0163] In this embodiment, if the number of cycles increases from 500 times to 1,000 times, and the decrease in capacitance characteristics increases from 15% to 25%, and the calculation of the correlation coefficient shows a high significance level, then the correlation is strong. This kind of analysis helps to reveal the driving factors of attenuation. When weighted adjustment is performed on the comprehensive attenuation trend parameters according to the significance level.

[0164] Specifically, the weights can be dynamically allocated according to the magnitude of the correlation coefficient.

[0165] For example, if the contribution of the double-layer capacitance is significantly higher, the weight is adjusted from 0.4 to 0.6 to obtain the adjusted distribution of trend parameters. This adjustment can more realistically reflect the attenuation effects of different capacitance types. When extracting the life estimation value from the adjusted distribution of trend parameters.

[0166] Preferably, the final output can be determined in combination with the current cycle state.

[0167] In this embodiment, assuming that the current number of cycles is 700 times, and the adjusted parameter distribution shows that the attenuation trend tends to be stable, the final predicted output may be 1,600 times. This method ensures that the prediction result is closer to the actual situation through multi-level analysis, which helps to optimize the battery management strategy.

[0168] S107. Update the three-dimensional relationship model according to the life estimation result, and use the real-time impedance spectrum data to iteratively optimize the model parameters to determine the change law of the capacitance characteristics reflecting the latest aging mechanism.

[0169] Update the three-dimensional relationship model through real-time impedance spectrum data, adjust the model parameters using an iterative optimization method, obtain the characteristic distribution reflecting the current state, extract the correlation data between the capacitance characteristics and the change rules from the characteristic distribution, obtain the preliminary distribution of the dynamic rules. If the preliminary distribution of the dynamic rules exceeds the preset threshold, then fit the correlation data through the support vector regression algorithm to determine the corrected distribution data. Analyze the change trend between the real-time impedance and the aging characteristics based on the corrected distribution data to obtain the adjustment value of the trend parameter, and perform a secondary update on the parameters of the three-dimensional relationship model through the adjustment value to obtain the updated model parameter distribution. Extract the dynamic rule data reflecting the capacitance characteristics from the updated model parameter distribution, judge the evolution direction of the aging characteristics, and use the judgment result to perform a time-series correction on the characteristic distribution to determine the final evolution trend distribution.

[0170] Specifically, when updating the three-dimensional relationship model through real-time impedance spectrum data, it can be understood as using the dynamic impedance information during battery operation to optimize the model structure.

[0171] For example, during the charging and discharging process of the battery, the impedance spectrum data collected in real time may show variation characteristics in the frequency range of 1 Hz to 1000 Hz. These data are input into the initial three-dimensional relationship model to reflect the aging state of the electrode material.

[0172] In this embodiment, the resistance and capacitance parameters in the model can be adjusted by comparing the differences between the new and old impedance spectra to make it closer to the current battery state. When adjusting the model parameters using the iterative optimization method.

[0173] Specifically, the parameter values can be gradually corrected based on the idea of gradient descent.

[0174] Exemplarily, assume that the capacitance value predicted by the initial model is 500 F, while the real-time data indicates that the actual value is close to 480 F. The iterative optimization will gradually narrow the gap between the prediction and the actual value through multiple loop calculations, and finally obtain the characteristic distribution reflecting the current state. This distribution may show the fluctuation range of the capacitance value at different loop times, for example, it is 480 F to 490 F at 100 loops. When extracting the correlation data between the capacitance characteristics and the change rules from the characteristic distribution.

[0175] It should be noted that this step focuses on how the capacitance changes over time or the number of cycles.

[0176] In this embodiment, it is analyzed that after 200 cycles, the capacitance has decreased by 5%, while the impedance has increased by 10%, thereby extracting the preliminary correlation rule between the two. If the preliminary distribution of the dynamic rules exceeds the preset threshold, for example, a capacitance decrease rate exceeding 2% is regarded as abnormal, then the correlation data is fitted through the support vector regression algorithm.

[0177] Preferably, the algorithm trains the model based on historical data and outputs the corrected distribution data. For example, the abnormal decline rate is adjusted to a smoother 1.8%. When analyzing the change trend between the real-time impedance and the aging characteristics according to the corrected distribution data, it can be understood as paying attention to the change in the shape of the impedance spectrum curve.

[0178] For example, a significant increase in the impedance in the low-frequency band may imply electrolyte loss. By quantifying this trend, the adjusted value of the trend parameter is obtained. For example, the resistance parameter in the model is increased from 0.1 Ω to 0.15 Ω. After the parameters of the three-dimensional relationship model are updated twice by the adjusted value, the obtained model parameter distribution may show the dynamic law data of the capacitance characteristics at 300 cycles. For example, the downward trend tends to flatten. When judging the evolution direction of the aging characteristics.

[0179] Specifically, by comparing the characteristic distributions at different cycle stages, it can be inferred whether the battery enters the accelerated aging stage.

[0180] In this embodiment, if the capacitance decline rate changes from 0.5% to 1% after 400 cycles, it indicates that the aging intensifies. When performing time series correction on the characteristic distribution using the judgment result.

[0181] For example, a time weighting factor can be introduced to increase the weight of recent data by 20%, so as to determine the final evolution trend distribution. This distribution may predict that the capacitance of the battery is 450 F at 500 cycles, which helps to plan the maintenance strategy in advance.

[0182] S108. Obtain the optimized three-dimensional relationship model, and repeat the steps of separating the double-layer capacitance and the Faraday capacitance for the newly collected impedance spectrum data to obtain the dynamically adjusted attenuation trend parameters.

[0183] Update the three-dimensional model with the newly collected impedance spectrum data, separate the double-layer capacitance and the Faraday capacitance to obtain the preliminary capacitance characteristic distribution, extract the attenuation trend from the preliminary capacitance characteristic distribution, use the dynamic adjustment method to determine the adjusted trend parameters, for the adjusted trend parameters, verify the accuracy of real-time processing through the impedance spectrum data, obtain the verified parameter distribution, if the verified parameter distribution exceeds the preset threshold, then fit the attenuation trend through the support vector regression algorithm to obtain the corrected trend data, update the three-dimensional model according to the corrected trend data, obtain the optimized model parameter distribution, extract the dynamically adjusted attenuation trend from the optimized model parameter distribution, judge the change direction of the capacitance characteristics, and correct the time series of data acquisition through the change direction to determine the final trend parameter distribution.

[0184] Specifically, when updating the three-dimensional model with the newly collected impedance spectrum data.

[0185] It is understandable that impedance spectroscopy data usually contains rich frequency information and can reflect the electrochemical characteristics of materials.

[0186] Exemplarily, in the scenario of lithium battery aging analysis, assuming that the collected impedance spectroscopy data is from battery cycle tests, a three-dimensional model can be initially constructed through these data, with the horizontal axis being the frequency, the vertical axis being the impedance amplitude, and the depth axis being the phase angle.

[0187] In this embodiment, separating the double-layer capacitance and the Faraday capacitance can be accomplished through an equivalent circuit model. For example, the double-layer capacitance is regarded as the response in the high-frequency part, while the Faraday capacitance is more manifested in the low-frequency region. For instance, if the high-frequency impedance value is around 0.1 ohm and the low-frequency impedance rises to 1 ohm, it can be preliminarily judged that the attenuation of the double-layer capacitance is small and the contribution of the Faraday capacitance increases.

[0188] Specifically, when extracting the attenuation trend from the preliminary capacitance characteristic distribution, the change of the impedance spectroscopy curve at different cycle numbers can be compared.

[0189] For example, at the 50th cycle, the low-frequency impedance increases by 20%, while at the 100th cycle, it increases by 50%, indicating that the attenuation trend gradually intensifies.

[0190] Preferably, when using the dynamic adjustment method to determine the trend parameters, a time weight factor can be introduced, with higher weights assigned to recent data. For example, the data weight for the most recent 10 cycles is 0.8, and the earlier data is 0.2, so as to more accurately reflect the current state.

[0191] It should be noted that when verifying the accuracy of real-time processing through impedance spectroscopy data, the impedance value predicted by the model can be compared with the actual measured value. If the error is less than 5%, the parameter distribution is considered reliable.

[0192] In this embodiment, if the verified parameter distribution exceeds the preset threshold, for example, the attenuation trend exceeds the expected 30%, the data is fitted through the support vector regression algorithm.

[0193] For example, inputting the corresponding relationship between 100 groups of impedance values and the cycle number, and outputting a smooth attenuation curve, the corrected trend data may show that the attenuation rate is adjusted from 0.5% / cycle to 0.4% / cycle. After updating the three-dimensional model according to the corrected trend data, the optimized model parameter distribution may show that the proportion of the double-layer capacitance decreases from 60% to 50%, reflecting the change of the aging characteristics. When extracting the dynamically adjusted attenuation trend from it, it can be observed that the growth rate of the low-frequency impedance slows down, indicating that the capacitance characteristics tend to be stable.

[0194] For example, when determining the change direction of capacitance characteristics, if the high-frequency impedance decreases while the low-frequency impedance increases, it can be inferred that the double-layer capacitance weakens while the Faraday process strengthens. When correcting the time series of data acquisition through the change direction, the sampling frequency can be adjusted from every 10 cycles to every 5 cycles to capture more detailed trend changes.

[0195] In this embodiment, the final trend parameter distribution may show that the attenuation rate stabilizes at 0.3% / cycle, which helps predict the battery life and optimize the usage strategy.

[0196] It can be understood that this method can dynamically adapt to the aging process and improve the practicality of the model.

[0197] S109. Re-run the support vector regression algorithm with the dynamically adjusted attenuation trend parameters to obtain the updated predicted value of the remaining battery life, and determine whether there is a significant change in the coupling relationship between the aging mechanism and the state of charge.

[0198] Run the support vector regression algorithm through the attenuation trend data to obtain the preliminary predicted value of the remaining life. Extract the aging mechanism characteristics according to the preliminary predicted value of the remaining life, determine the aging characteristic distribution, analyze the change of the state of charge through the aging characteristic distribution, and obtain the preliminary distribution of the coupling relationship. If the preliminary distribution of the coupling relationship exceeds the preset threshold, correct the change trend through the support vector regression algorithm to obtain the adjusted distribution data, update the attenuation trend according to the adjusted distribution data, judge the change direction of the aging mechanism, correct the time series of the state of charge through the change direction, determine the final coupling relationship distribution, update the predicted value of the remaining life according to the final coupling relationship distribution, and obtain the optimized prediction result.

[0199] Specifically, run the support vector regression algorithm through the attenuation trend data to obtain the preliminary predicted value of the remaining life.

[0200] It can be understood that support vector regression is used here to capture the non-linear characteristics in the attenuation trend.

[0201] For example, in the battery aging scenario, assume that the capacity attenuation data of a certain battery shows an exponential decline trend with the number of cycles. The initial capacity is 1000 mAh and it drops to 800 mAh after 500 cycles. Through the algorithm, its remaining life can be predicted to be about 300 cycles. Extract the aging mechanism characteristics according to the preliminary predicted value of the remaining life.

[0202] Specifically, the aging mechanism may involve crack propagation of the electrode material or electrolyte decomposition.

[0203] In this embodiment, when extracting features, attention can be paid to the relationship between the capacity attenuation rate and the increase in internal resistance. For example, when the internal resistance rises from 50 mΩ to 80 mΩ, it indicates that the solid electrolyte interface film thickens, and this feature can be used as a sign of the aging mechanism. When determining the aging feature distribution.

[0204] Exemplarily, the distribution laws of the internal resistance and temperature can be observed by statistically analyzing the data of multiple experiments.

[0205] For example, in two environments of 25 °C and 45 °C, the increase amplitudes of the internal resistance are 20 mΩ and 35 mΩ respectively, showing the influence of temperature on the aging feature distribution. Analyze the change of the state of charge through the aging feature distribution.

[0206] It should be noted that the state of charge is usually related to the open-circuit voltage of the battery.

[0207] In this embodiment, if the voltage of a certain battery drops from 3.7 V to 3.65 V at a 50% state of charge, it may reflect the capacity decline caused by the loss of active substances. After obtaining the preliminary distribution of the coupling relationship, if it exceeds the preset threshold, for example, the voltage change exceeds 0.1 V, the support vector regression algorithm is used to correct the change trend.

[0208] Preferably, environmental factors can be introduced during correction. For example, when the humidity is 60%, the attenuation accelerates, and the predicted life is corrected from 300 cycles to 280 cycles after adjustment. Update the attenuation trend according to the adjusted distribution data.

[0209] For example, combining temperature and the number of cycles, the attenuation curve changes from a single exponential form to a piecewise function form, reflecting the aging rate at different stages. When judging the change direction of the aging mechanism.

[0210] In this embodiment, if the rate of increase in the internal resistance slows down, it may indicate that the electrolyte decomposition tends to be stable. The time series of the state of charge can be corrected through the change direction.

[0211] For example, the state of charge at 100 cycles in the original data is corrected from 60% to 58%, improving the data consistency. When determining the final coupling relationship distribution.

[0212] Specifically, it can be verified through multiple groups of experiments. For example, at different charging rates, the difference in the coupling distribution between 0.5 C and 1 C is reflected in that the predicted life values differ by 50 cycles. Update the remaining life prediction value according to the final coupling relationship distribution.

[0213] For example, after comprehensive adjustment, the obtained life is 290 cycles, which is closer to the actual situation than the preliminary prediction. When obtaining the optimized prediction result.

[0214] Preferably, the accuracy can be verified by comparing with the measured data. For example, the error between the predicted value and the actual life of 295 cycles is only 5 cycles. This method can effectively improve the reliability of the prediction and provide a more accurate reference for battery management.

[0215] The above are only the preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the technical field of the present application within the technical scope disclosed by the present application should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for estimating the cycle life of a lithium-ion power battery based on SOC, characterized in that: The following steps are involved: Collect impedance data of the battery at different states of charge and cycle numbers to obtain an initial data set of capacitance characteristics including double-layer capacitance and Faraday capacitance; The contributions of double-layer capacitance and Faraday capacitance were separated based on the initial data set, and the impedance spectrum data were fitted with an equivalent circuit model to determine the independent variation of the two capacitance characteristics under different charge states. The correlation between the capacitance characteristics and the state of charge in each cycle number section in the independent variation law was statistically analyzed to obtain the attenuation trend of the two capacitances with the cycle number. A three-dimensional relationship model between capacitance characteristics, cycle times and state of charge is established through attenuation trend data; The capacitance characteristic attenuation value under specific state of charge and number of cycles is extracted from the three-dimensional relationship model, and the difference between double-layer capacitance and Faraday capacitance is weighted to obtain the comprehensive attenuation trend parameter; If the comprehensive attenuation trend parameter exceeds the preset threshold, the remaining battery life is predicted by the support vector regression algorithm to obtain the life estimation result based on the capacitance characteristics under the current number of cycles; Update the three-dimensional relationship model based on the life estimation results, and iteratively optimize the model parameters using real-time impedance spectrum data; Obtaining an optimized three-dimensional relationship model, and obtaining dynamically adjusted attenuation trend parameters according to newly collected impedance spectrum data; The support vector regression algorithm is re-run with the dynamically adjusted attenuation trend parameters to obtain an updated battery remaining life prediction value.

2. The method according to claim 1, characterized in that Obtaining the initial data set of capacitance characteristics includes: The impedance data of the battery at different states of charge and cycle numbers are collected by electrochemical impedance spectroscopy technology to obtain an initial data set; Capacitance characteristic data including double-layer capacitance and Faraday capacitance are extracted from the initial data set to obtain a capacitance characteristic data subset.

3. The method according to claim 1, characterized in that Determining the independent variation rules of the two capacitance characteristics under different charge states includes: Load the equivalent circuit model through the impedance spectrum data to obtain the fitted parameter set; Extract the contributions of double-layer capacitance and Faraday capacitance from the fitted parameter set to obtain a separated capacitance data set; Using the separated capacitance data set, the characteristic parameters of the double layer capacitance at different states of charge are determined; The distribution of Faraday capacitance is calculated by characteristic parameters, and its independent variation curves under different charge states are obtained; If the independent variation curve deviates from the preset threshold, the optimized capacitance characteristic data is obtained by adjusting the parameters of the equivalent circuit model; According to the optimized capacitor characteristic data, the dynamic response of the two capacitors when the charge state changes is determined, and the final independent change law is obtained.

4. The method according to claim 1, characterized in that: The attenuation trends of the two capacitances with the number of cycles include: By loading the data set corresponding to the number of cycles, the capacitance characteristic data in each cycle number section is obtained, and the distribution characteristics of the state of charge are extracted from the obtained capacitance characteristic data to obtain the correlation parameters. The correlation parameters are processed by statistical analysis to determine the attenuation trend of the double-layer capacitance. By calculating the dynamic response of the Faraday capacitance, its variation law with the number of cycles is obtained. If the change pattern deviates from the preset threshold, the parameters of the statistical analysis are adjusted to obtain the optimized attenuation trend. Based on the optimized attenuation trend, the independent distribution characteristics of the two capacitors under the charged state are determined. A complete data set is generated through the independent distribution characteristics to obtain the final attenuation trend.

5. The method according to claim 1, characterized in that The three-dimensional relationship model of capacitance characteristics, cycle times and state of charge is established by using attenuation trend data, including: The initial mapping relationship between capacitance characteristics and cycle times is generated through attenuation trend data, and the mapping relationship is processed using a polynomial regression algorithm to obtain the initial characteristics of dynamic changes; The distribution parameters of the state of charge are extracted from the preliminary characteristics, the data basis of the three-dimensional relationship is determined according to the distribution parameters, the parameters of the polynomial regression are adjusted according to the three-dimensional relationship data, the optimized dynamic change characteristics are obtained, the intermediate expression of the aging mechanism is calculated through the optimized dynamic change characteristics, and the three-dimensional relationship model of mathematical expression is obtained.

6. The method according to claim 1, characterized in that The comprehensive attenuation trend parameters obtained include: The attenuation value corresponding to the state of charge and the number of cycles is extracted through a three-dimensional relationship model, and the difference between the double-layer capacitance and the Faraday capacitance is quantified using weighted processing to obtain a comprehensive attenuation parameter. The time series distribution of the attenuation value is obtained from the comprehensive attenuation parameters. The linear interpolation processing is performed on the changing trend of the number of cycles to determine the continuous expression of the attenuation value. The dynamic distribution characteristics of the state of charge are extracted based on the continuous expression. The characteristics are classified by the random forest algorithm to determine the contribution ratio of the double-layer capacitance and the Faraday capacitance. If the contribution ratio exceeds a preset threshold, the comprehensive attenuation parameters are recalculated by adjusting the weighted processing parameters to obtain a corrected comprehensive attenuation trend.

7. The method according to claim 1, characterized in that The obtaining of the life estimation result based on the capacitance characteristics at the current number of cycles includes: The comprehensive attenuation trend parameters are processed by the support vector regression algorithm to obtain the preliminary prediction value of the remaining battery life. The preliminary prediction value is corrected by the number of cycles to determine the corrected life estimation distribution. Extract the change characteristics of the capacitance characteristics from the corrected life estimation distribution, and obtain the dynamic distribution data of the characteristics. If the dynamic distribution data exceeds the preset threshold, perform a quadratic fit on the characteristics through a regression algorithm to obtain a corrected prediction result. Analyze the correlation between the number of cycles and the capacitance characteristics according to the revised prediction results, determine the significance of the correlation, perform weighted adjustment on the comprehensive attenuation trend parameters according to the significance, and obtain the adjusted trend parameter distribution; The lifetime estimate for the current state is extracted from the adjusted trend parameter distribution to determine the final forecast output.

8. The method according to claim 1, characterized in that The determination of the change law of the capacitance characteristic reflecting the latest aging mechanism includes: The three-dimensional relationship model is updated through real-time impedance spectrum data, and the model parameters are adjusted by iterative optimization method to obtain the characteristic distribution reflecting the current state. The correlation data of capacitance characteristics and change laws are extracted from the characteristic distribution to obtain the preliminary distribution of dynamic laws. If the preliminary distribution of the dynamic law exceeds the preset threshold, the associated data is fitted by the support vector regression algorithm to determine the corrected distribution data, and the change trend between the real-time impedance and the aging characteristics is analyzed according to the corrected distribution data to obtain the adjustment value of the trend parameter. The parameters of the three-dimensional relationship model are updated for the second time by the adjustment value to obtain the updated model parameter distribution; The dynamic law data reflecting the capacitor characteristics are extracted from the updated model parameter distribution to determine the evolution direction of the aging characteristics. The judgment result is used to perform time series correction on the characteristic distribution to determine the final evolution trend distribution.

9. The method according to claim 1, characterized in that: The dynamically adjusted attenuation trend parameters include: The three-dimensional model is updated through the newly collected impedance spectrum data, the double-layer capacitance and the Faraday capacitance are separated, a preliminary capacitance characteristic distribution is obtained, and the attenuation trend is extracted from the preliminary capacitance characteristic distribution; A dynamic adjustment method is used to determine the adjusted trend parameters. For the adjusted trend parameters, the accuracy of real-time processing is verified through impedance spectrum data to obtain the verified parameter distribution. If the verified parameter distribution exceeds the preset threshold, the attenuation trend is fitted through the support vector regression algorithm to obtain the corrected trend data. The three-dimensional model is updated according to the corrected trend data to obtain the optimized model parameter distribution, the dynamically adjusted attenuation trend is extracted from the optimized model parameter distribution, the change direction of the capacitance characteristics is determined, the time series of data acquisition is corrected according to the change direction, and the final trend parameter distribution is determined.

10. The method according to claim 1, characterized in that The obtaining of the updated remaining battery life prediction value comprises: Run the support vector regression algorithm through the attenuation trend data to obtain the preliminary remaining life prediction value, extract the aging mechanism characteristics based on the preliminary remaining life prediction value, determine the aging characteristic distribution, analyze the charge state change through the aging characteristic distribution, and obtain the preliminary distribution of the coupling relationship; If the preliminary distribution of the coupling relationship exceeds the preset threshold, the change trend is corrected by the support vector regression algorithm to obtain the adjusted distribution data. The attenuation trend is updated according to the adjusted distribution data to determine the change direction of the aging mechanism. The time series of the state of charge is corrected according to the change direction to determine the final coupling relationship distribution. The remaining life prediction value is updated according to the final coupling relationship distribution to obtain the optimized prediction result.

Images