A method and system for health assessment of an energy storage lithium battery system

Abstract

The present application relates to the technical field of lithium battery, and more particularly to a health evaluation method and system of an energy storage lithium battery system. The method comprises the following steps: collecting wide frequency domain complex impedance response data of the energy storage lithium battery system in a floating state, and performing frequency domain segmentation to obtain lithium battery impedance characteristic data; determining a calendar-microcycle aging distribution weight based on the lithium battery impedance characteristic data; performing microcalorimetry detection on the energy storage lithium battery system to obtain real-time battery heat flow time series data; performing DSC testing on the energy storage lithium battery system based on the real-time battery heat flow time series data to obtain a DSC heat flow curve; and performing side reaction product evolution prediction on the energy storage lithium battery system based on the DSC heat flow curve to obtain a positive electrode interface side reaction parameter set. The present application realizes real-time and accurate evaluation and life prediction of the health state of the energy storage lithium battery system.

Description

Technical Field

[0001] This invention relates to the field of lithium battery technology, and in particular to a health assessment method and system for energy storage lithium battery systems. Background Technology

[0002] In energy storage applications for communication base stations, lithium battery systems serve as backup power sources during mains power outages, playing a crucial role in ensuring the continuous operation of base stations. These scenarios exhibit distinct characteristics: batteries operate at high float charging voltages (typically 53.5-54V) for extended periods, undergoing only extremely shallow charge-discharge cycles daily (depth of discharge often less than 5%), and must operate continuously within high-temperature, sealed cabinets. This complex operating condition of "high-voltage constant-voltage float charging + micro-circulation + high-temperature environment" significantly differs from the deep-cycle mode of electric vehicles or grid-scale energy storage, causing traditional health assessment models to frequently fail in this scenario.

[0003] Existing technologies often overlook the nonlinear coupling effect between calendar aging and microcirculation aging. During the long service life of base station batteries, which can last for several years, 99% of the time is spent in a fully charged float charge state, where high temperatures accelerate electrolyte decomposition and SEI thickening; the remaining 1% of the time is spent handling instantaneous loads (such as equipment start-up and shutdown) or short-term discharges (such as mains power interruptions). Experiments show that the by-reaction products at the positive electrode interface induced by high-voltage float charging can cross-catalyze lithium plating at the negative electrode caused by microcirculation, resulting in a step-like increase in the capacity decay rate during the later stages of service. Summary of the Invention

[0004] Therefore, the present invention needs to provide a health assessment method and system for energy storage lithium battery systems to solve at least one of the above-mentioned technical problems.

[0005] To achieve the above objectives, a health assessment method for an energy storage lithium battery system includes the following steps:

[0006] Step S1: Collect wide-frequency domain complex impedance response data of the energy storage lithium battery system under float charging state, and perform frequency domain segmentation to obtain lithium battery impedance characteristic data; determine calendar-micro-cycle aging distribution weights based on lithium battery impedance characteristic data.

[0007] Step S2: Perform microcalorimetry on the energy storage lithium battery system to obtain real-time battery heat flow time-series data; perform DSC test on the energy storage lithium battery system based on the real-time battery heat flow time-series data to obtain DSC heat flow curve; predict the evolution of side reaction products of the energy storage lithium battery system based on the DSC heat flow curve to obtain the positive electrode interface side reaction parameter set.

[0008] Step S3: Configure a potential step relaxation test device for the energy storage lithium battery system, and quantify the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test device to obtain lithium battery lithium plating degree data; predict the capacity decay of the lithium battery based on the lithium battery lithium plating degree data to obtain capacity drop inflection point prediction data.

[0009] Step S4: Construct a lithium battery health state transition equation based on calendar-microcirculation aging distribution weights; evaluate the lithium battery health state based on the lithium battery health state transition equation to obtain key health indicators of the lithium battery.

[0010] Step S5: Based on the key health indicators of lithium batteries and the predicted inflection point of capacity drop, assess the remaining life of the energy storage lithium battery system to obtain the corrected life prediction data of the lithium battery.

[0011] This invention, by acquiring wide-frequency domain complex impedance response data and performing frequency domain segmentation, can capture the impedance characteristics of lithium batteries at different frequencies from an electrochemical perspective, thereby accurately extracting impedance feature data. This not only considers the charge transport characteristics of the electrode materials inside the battery but also incorporates the electrochemical reaction dynamics at the electrolyte-electrode interface, effectively distinguishing the nonlinear effects of calendar aging and micro-cycle aging on battery performance and accurately determining the aging distribution weights. This impedance feature-based weight determination method fundamentally solves the problem of insufficient understanding of aging mechanisms in traditional methods. By combining microcalorimetry and differential scanning calorimetry (DSC) testing, the evolution of by-reaction products at the cathode interface is predicted from a thermodynamic perspective. This can accurately capture the heat flow changes of by-reactions inside the battery, and by analyzing the heat flow curves, reveal key parameters such as the enthalpy change characteristics, reaction rate constants, and product concentrations of the by-reactions. This not only allows for real-time monitoring of the dynamics of by-reactions inside the battery but also predicts the potential degradation risk of battery performance through the evolution trend of by-reaction products, thus providing a more comprehensive basis for health status assessment. The degree of lithium plating is quantified using a potential step relaxation test device, and the capacity drop inflection point is predicted. From an electrochemical kinetics perspective, this accurately assesses the impact of lithium plating on battery performance by monitoring the negative electrode potential relaxation characteristic parameters. Combining nuclear magnetic resonance spectroscopy (NMR) and scanning electron microscopy (SEM), the degree of lithium plating is further quantified at the atomic scale and microstructure level, thus achieving accurate prediction of battery capacity decay. This effectively captures changes in the battery's internal microstructure, providing more precise quantitative indicators for health status assessment. Through a health status transition equation constructed based on aging distribution weights, and a remaining lifetime assessment combining key health indicators and capacity drop inflection point prediction data, real-time and accurate assessment of the health status and lifetime prediction of energy storage lithium battery systems is achieved. This not only provides early warning of battery health problems but also provides a scientific basis for battery maintenance and replacement by accurately predicting capacity drop inflection points, thereby effectively extending battery life, reducing the operation and maintenance costs of communication base station energy storage systems, and improving system reliability and economy.

[0012] Preferably, step S1 involves acquiring wide-frequency domain complex impedance response data of the energy storage lithium battery system in float charging state, and performing frequency domain segmentation, including:

[0013] Configure a multi-channel data acquisition interface for the energy storage lithium battery system to obtain the electrochemical impedance spectroscopy test parameter configuration;

[0014] Based on the configuration of electrochemical impedance spectroscopy test parameters, a wide frequency domain scanning range was set for the energy storage lithium battery system to obtain wide frequency domain impedance scanning parameters. The wide frequency domain setting frequency scanning range is 0.01Hz-10kHz, a logarithmic distribution method is adopted, 10 test points are set for each frequency range, the perturbation signal amplitude is set to 5mV, the phase accuracy is ±0.1°, the impedance accuracy is ±0.1%, and the single scan time is controlled within 30 minutes.

[0015] Impedance spectrum acquisition of energy storage lithium battery system under float charging state was performed based on wide frequency domain impedance scanning parameters to obtain raw complex impedance response data;

[0016] Noise filtering was applied to the original complex impedance response data to obtain the purified impedance spectrum data of the lithium battery.

[0017] Normalize the lithium battery purification impedance spectrum data to generate a standard lithium battery impedance spectrum dataset.

[0018] According to the preset frequency domain segmentation rules, the standard lithium battery impedance spectrum dataset is segmented in the frequency domain to obtain lithium battery impedance characteristic data. The frequency domain is divided into three intervals: high frequency, mid frequency, and low frequency. For each frequency band, four characteristic parameters are extracted: real part, imaginary part, magnitude, and phase angle, forming a 12-dimensional feature vector.

[0019] Preferably, the determination of calendar-microcycle aging distribution weights based on lithium battery impedance characteristic data in step S1 includes:

[0020] The wavelet basis function is selected based on the impedance characteristic data of lithium battery, and the selection parameters of the wavelet basis function are obtained.

[0021] Based on the selection parameters of wavelet basis function, continuous wavelet transform is performed on the impedance data of each frequency point in the impedance characteristic data of lithium battery to obtain the time-frequency domain lithium battery impedance coefficient set.

[0022] The wavelet coefficient energy at each scale of the time-frequency domain lithium battery impedance coefficient set is calculated to obtain the energy distribution data at each scale.

[0023] Based on the energy distribution data at each scale, the top 8 main scales with a concentrated energy proportion greater than 5% in the time-frequency domain lithium battery impedance coefficient are selected, and the real part, imaginary part, modulus and phase information of the corresponding wavelet coefficients are extracted to form a time-frequency domain impedance feature vector set.

[0024] The Hilbert-Huang transform is performed on the time-frequency domain impedance eigenvector set to obtain the impedance empirical mode decomposition parameters;

[0025] Based on the impedance empirical mode decomposition parameters, each dimension of the eigenvector set of the impedance eigenvector set in the time-frequency domain is decomposed into EMD one by one to obtain the initial set of intrinsic mode functions.

[0026] The initial set of intrinsic mode functions is subjected to Hilbert transform to obtain the instantaneous frequency and amplitude data of the modes; frequency characteristic statistics are performed on the instantaneous frequency and amplitude data of the modes to obtain a multi-band intrinsic mode function set;

[0027] The weights of aging mechanisms are calculated based on the multi-band intrinsic mode function set to obtain the calendar-microcycle aging distribution weights.

[0028] Preferably, step S2 includes the following steps:

[0029] Step S21: Configure a micro-calorimetry measurement device for the energy storage lithium battery system to obtain the parameters of the battery heat flow detection system;

[0030] Step S22: Monitor the heat flow of the energy storage lithium battery system under float charging state based on the parameters of the battery heat flow detection system to obtain the original battery heat flow time series data;

[0031] Step S23: Perform baseline drift correction on the original battery heat flow time series data to obtain purified battery heat flow time series data;

[0032] Step S24: Identify the heat flow peak value of the purified battery heat flow time series data to obtain the side reaction heat flow characteristic points;

[0033] Step S25: Based on the side reaction heat flow characteristic points, pre-configure the DSC test parameters of the energy storage lithium battery system to obtain the DSC test parameter configuration;

[0034] Step S26: According to the DSC test parameters, program the temperature of the battery sample of the energy storage lithium battery system and record the relationship curve between heat flow and temperature to obtain the DSC heat flow curve.

[0035] Step S27: Based on the DSC heat flow curve, predict the evolution of by-reaction products of the energy storage lithium battery system to obtain the parameter set of by-reactions at the positive electrode interface.

[0036] Preferably, step S27 includes the following steps:

[0037] Step S271: Perform peak separation and fitting on the DSC heat flow curve to obtain the enthalpy change characteristic data of the side reaction, wherein the enthalpy change characteristic data of the side reaction includes peak temperature, peak intensity, peak width and peak area;

[0038] Step S272: Repeated DSC tests were performed on the battery samples of the energy storage lithium battery system at different temperature points at each peak temperature, and the reaction rate constant at different temperatures was calculated to obtain the battery reaction rate constant parameters.

[0039] Step S273: Prepare for X-ray photoelectron spectroscopy (XPS) detection of the energy storage lithium battery system to obtain XPS test condition parameters;

[0040] Step S274: Detect the surface composition of the positive electrode of the energy storage lithium battery system according to the XPS test conditions and parameters to obtain the lithium battery chemical state spectrum data;

[0041] Step S275: Perform peak decomposition and quantification of by-reaction products on the lithium battery chemical state spectrum data to obtain by-reaction product concentration data;

[0042] Step S276: Construct a side reaction model for the energy storage lithium battery system based on the battery reaction rate constant parameter and the side reaction product concentration data to obtain the lithium battery side reaction model;

[0043] Step S277: Based on the preset prediction period and lithium battery side reaction model, predict the evolution of side reaction products of the energy storage lithium battery system to obtain positive electrode interface degradation trend data; extract and integrate parameters from the positive electrode interface degradation trend data to obtain a positive electrode interface side reaction parameter set.

[0044] Preferably, step S3, which involves configuring a potential step relaxation test device for the energy storage lithium battery system and quantifying the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test device, includes:

[0045] The potential step relaxation test equipment for the energy storage lithium battery system was configured to obtain the parameters of the electrochemical workstation;

[0046] Based on the parameters of the electrochemical workstation, the micro-circulation potential of the energy storage lithium battery system was monitored to obtain the raw data of the negative electrode potential response.

[0047] Noise filtering is applied to the raw negative electrode potential response data to obtain purified lithium battery potential relaxation data;

[0048] Relaxation curves were fitted to the potential relaxation data of purified lithium batteries to obtain potential relaxation characteristic parameters.

[0049] Lithium plating criteria are established based on potential relaxation characteristic parameters to obtain the critical lithium plating potential threshold.

[0050] The degree of lithium plating in the energy storage lithium battery system is quantified based on the critical potential threshold for lithium plating and the potential relaxation data of purified lithium batteries, thus obtaining lithium battery lithium plating degree data.

[0051] Preferably, step S3, which predicts lithium battery capacity degradation based on lithium battery lithium plating level data, includes:

[0052] The energy storage lithium battery system was prepared for nuclear magnetic resonance spectroscopy detection to obtain NMR test parameter configuration; based on the NMR test parameter configuration, the energy storage lithium battery system was quantitatively detected for lithium metal to obtain 7Li NMR spectrum data;

[0053] Peak decomposition and lithium deposition quantification were performed on the 7Li NMR spectrum data to obtain lithium deposition data of the lithium battery anode.

[0054] Based on lithium plating data from lithium battery anodes, scanning electron microscope (SEM) configurations were performed on energy storage lithium battery systems to obtain SEM detection parameters.

[0055] The surface morphology of the negative electrode of the energy storage lithium battery system was scanned based on the scanning electron microscope detection parameters to obtain lithium deposition morphology characteristic data.

[0056] A simulation model of the energy storage lithium battery system was constructed to obtain the MD simulation parameter configuration; based on the MD simulation parameter configuration and the positive electrode interface side reaction parameter set, the cross-catalytic effect of the energy storage lithium battery system was investigated to obtain intermolecular interaction energy data.

[0057] Catalytic reaction simulation of energy storage lithium battery system was performed based on intermolecular interaction energy data to obtain cross-catalytic reaction activation energy data;

[0058] The coupling effect was evaluated based on lithium battery lithium plating degree data, lithium battery anode lithium plating data, and cross-catalytic reaction activation energy data to obtain positive and negative electrode interface coupling parameters.

[0059] Based on the coupling parameters of the positive and negative electrode interfaces, capacity decay prediction of energy storage lithium battery systems is performed to obtain prediction data of the inflection point of capacity drop.

[0060] Preferably, step S4 includes the following steps:

[0061] Step S41: Data synchronously acquires data from the energy storage lithium battery system according to the preset three-level time scale acquisition system to obtain lithium battery time-series synchronous acquisition parameters;

[0062] Step S42: Align the degradation information time of the lithium battery time-series synchronous acquisition parameters according to the calendar-micro-cycle aging distribution weight to obtain the lithium battery synchronous degradation data table;

[0063] Step S43: Extract degradation feature vectors from the lithium battery synchronous degradation data table to obtain lithium battery degradation feature data;

[0064] Step S44: Establish a state-space model of lithium battery based on lithium battery degradation characteristic data;

[0065] Step S45: Solve the state covariance of the lithium battery state space model based on the lithium battery degradation characteristic data to obtain the lithium battery state covariance data;

[0066] Step S46: Perform covariance correction on the lithium battery state covariance data to obtain the Kalman filter parameter set;

[0067] Step S47: Model the state space of the lithium battery based on the Kalman filter parameter set to obtain the lithium battery health state transition equation;

[0068] Step S48: Evaluate the health status of the lithium battery based on the lithium battery health status transition equation to obtain key health indicators of the lithium battery.

[0069] Preferably, step S5 includes the following steps:

[0070] Step S51: Configure the support vector machine classifier parameters for the key health indicators of lithium batteries to obtain the SVM classifier parameter configuration;

[0071] Step S52: Construct training samples for key health indicators of lithium batteries according to the SVM classifier parameter configuration to obtain the SVM classification training dataset;

[0072] Step S53: Train the SVM model based on the SVM classification training dataset to obtain the SVM classifier model; perform cross-validation evaluation on the SVM classifier model to obtain SVM classifier performance evaluation data.

[0073] Step S54: Classify the health status of the energy storage lithium battery system based on the SVM classifier performance evaluation data and key health indicators of the lithium battery to obtain health status level classification data.

[0074] Step S55: Based on the health status level classification data, perform historical data statistics on the energy storage lithium battery system to obtain the health threshold adjustment benchmark data;

[0075] Step S56: Based on the health threshold adjustment benchmark data, perform adaptive threshold calculation on the energy storage lithium battery system to obtain the dynamic health threshold parameters of the lithium battery;

[0076] Step S57: Based on the dynamic threshold parameters of lithium battery health and the classification data of health status level, set early warning rules for the energy storage lithium battery system to obtain lithium battery health early warning rules;

[0077] Step S58: Based on the lithium battery health early warning rules and capacity drop inflection point prediction data, assess the remaining life of the energy storage lithium battery system to obtain lithium battery corrected life prediction data.

[0078] Preferably, the present invention also provides a health assessment system for an energy storage lithium battery system, used to perform the health assessment method for the energy storage lithium battery system as described above, the health assessment system for the energy storage lithium battery system comprising:

[0079] Impedance analysis module is used to collect wide-frequency domain complex impedance response data of energy storage lithium battery system in float charge state, and perform frequency domain segmentation to obtain lithium battery impedance characteristic data; based on lithium battery impedance characteristic data, calendar-micro-cycle aging distribution weights are determined.

[0080] The heat flow monitoring module is used to perform micro-caloric measurement and detection on the energy storage lithium battery system to obtain real-time battery heat flow time-series data; based on the real-time battery heat flow time-series data, the energy storage lithium battery system is subjected to DSC test to obtain DSC heat flow curve; based on the DSC heat flow curve, the evolution of side reaction products of the energy storage lithium battery system is predicted to obtain the positive electrode interface side reaction parameter set.

[0081] The lithium plating quantification module is used to configure the potential step relaxation test equipment for the energy storage lithium battery system, and to quantify the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test equipment to obtain lithium battery lithium plating degree data; based on the lithium battery lithium plating degree data, the lithium battery capacity decay is predicted to obtain capacity drop inflection point prediction data.

[0082] The health assessment module is used to construct a lithium battery health state transition equation based on calendar-microcirculation aging distribution weights; and to assess the health state of the lithium battery based on the lithium battery health state transition equation to obtain key health indicators of the lithium battery.

[0083] The life prediction module is used to assess the remaining life of energy storage lithium battery systems based on key health indicators and capacity drop inflection point prediction data of lithium batteries, and obtain corrected life prediction data of lithium batteries.

[0084] This invention utilizes multi-dimensional data fusion to accurately extract the impedance characteristics of lithium batteries from electrochemical, thermodynamic, and electrochemical kinetic perspectives, monitor the dynamics of side reactions, and quantify lithium plating phenomena. This comprehensively reflects the aging characteristics of lithium batteries under complex operating conditions, providing a rich information foundation for health status assessment. Based on impedance characteristic data, this invention determines the calendar-microcycle aging distribution weights, accurately reflecting the impact of different aging factors on battery performance and addressing the shortcomings of traditional methods in understanding aging mechanisms. Through a heat flow monitoring module and DSC testing, the system can monitor the evolution trend of side reaction products at the cathode interface in real time, providing early warnings of battery health problems and offering a scientific basis for maintenance optimization. The lithium plating quantification module quantifies the degree of lithium plating at the atomic scale and microscopic morphology level, accurately predicting capacity drop inflection points and further improving the accuracy of health status assessment. The health assessment module constructs a health status transition equation based on aging distribution weights, combining key health indicators and capacity drop inflection point prediction data to achieve real-time, accurate assessment and lifespan prediction of lithium battery health status. This timely reflects battery performance change trends, provides a scientific basis for battery maintenance and replacement, effectively extends battery life, and reduces operation and maintenance costs. This invention provides early warnings of battery health problems through accurate health status assessment and lifespan prediction, preventing communication base stations from being interrupted due to battery failure, and significantly improving the reliability and economy of the system. Attached Figure Description

[0085] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings:

[0086] Figure 1 A flowchart illustrating the steps of a health assessment method for an energy storage lithium battery system according to an embodiment is shown.

[0087] Figure 2 A detailed flowchart of step S27 of one embodiment is shown. Detailed Implementation

[0088] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0089] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0090] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0091] To achieve the above objectives, please refer to Figures 1 to 2 This invention provides a health assessment method for an energy storage lithium battery system, comprising the following steps:

[0092] Step S1: Collect wide-frequency domain complex impedance response data of the energy storage lithium battery system under float charging state, and perform frequency domain segmentation to obtain lithium battery impedance characteristic data; determine calendar-micro-cycle aging distribution weights based on lithium battery impedance characteristic data.

[0093] Step S2: Perform microcalorimetry on the energy storage lithium battery system to obtain real-time battery heat flow time-series data; perform DSC test on the energy storage lithium battery system based on the real-time battery heat flow time-series data to obtain DSC heat flow curve; predict the evolution of side reaction products of the energy storage lithium battery system based on the DSC heat flow curve to obtain the positive electrode interface side reaction parameter set.

[0094] Step S3: Configure a potential step relaxation test device for the energy storage lithium battery system, and quantify the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test device to obtain lithium battery lithium plating degree data; predict the capacity decay of the lithium battery based on the lithium battery lithium plating degree data to obtain capacity drop inflection point prediction data.

[0095] Step S4: Construct a lithium battery health state transition equation based on calendar-microcirculation aging distribution weights; evaluate the lithium battery health state based on the lithium battery health state transition equation to obtain key health indicators of the lithium battery.

[0096] Step S5: Based on the key health indicators of lithium batteries and the predicted inflection point of capacity drop, assess the remaining life of the energy storage lithium battery system to obtain the corrected life prediction data of the lithium battery.

[0097] Preferably, step S1 involves acquiring wide-frequency domain complex impedance response data of the energy storage lithium battery system in float charging state, and performing frequency domain segmentation, including:

[0098] Configure a multi-channel data acquisition interface for the energy storage lithium battery system to obtain the electrochemical impedance spectroscopy test parameter configuration;

[0099] Based on the configuration of electrochemical impedance spectroscopy test parameters, a wide frequency domain scanning range was set for the energy storage lithium battery system to obtain wide frequency domain impedance scanning parameters. The wide frequency domain setting frequency scanning range is 0.01Hz-10kHz, a logarithmic distribution method is adopted, 10 test points are set for each frequency range, the perturbation signal amplitude is set to 5mV, the phase accuracy is ±0.1°, the impedance accuracy is ±0.1%, and the single scan time is controlled within 30 minutes.

[0100] Impedance spectrum acquisition of energy storage lithium battery system under float charging state was performed based on wide frequency domain impedance scanning parameters to obtain raw complex impedance response data;

[0101] Noise filtering was applied to the original complex impedance response data to obtain the purified impedance spectrum data of the lithium battery.

[0102] Normalize the lithium battery purification impedance spectrum data to generate a standard lithium battery impedance spectrum dataset.

[0103] According to the preset frequency domain segmentation rules, the standard lithium battery impedance spectrum dataset is segmented in the frequency domain to obtain lithium battery impedance characteristic data. The frequency domain is divided into three intervals: high frequency, mid frequency, and low frequency. For each frequency band, four characteristic parameters are extracted: real part, imaginary part, magnitude, and phase angle, forming a 12-dimensional feature vector.

[0104] In this embodiment, a Gamry Reference 3000 electrochemical workstation was used as a multi-channel data acquisition device in a laboratory environment. This device has multiple channels, enabling simultaneous connection of multiple lithium battery samples for batch testing. The data acquisition interface was configured via the Gamry software interface, and the sampling frequency was set to 1000Hz. Based on the testing requirements, the potentiostat mode was selected, and the potential control accuracy was set to 0.01mV. After completing the interface configuration, the electrochemical impedance spectroscopy (EIS) test parameters were recorded, including the sampling frequency, operating mode, and accuracy settings. Based on the Gamry Reference 3000 EIS parameter configuration, the wideband scanning range was set to 0.01Hz to 10kHz. A logarithmic distribution was used, with 10 test points set at each frequency range to ensure sufficient data points to reflect impedance characteristics across different frequency intervals. The perturbation signal amplitude was set to 5mV, an experimentally verified suitable amplitude that effectively stimulates the electrochemical reaction inside the battery without causing irreversible damage. The phase accuracy was set to ±0.1°, and the impedance accuracy to ±0.1%. To improve testing efficiency and reduce environmental interference during the testing process, the single scan time was controlled within 30 minutes. After setting the above parameters, complete wideband impedance scanning parameters were obtained. The lithium battery system under test was connected to the GamryReference 3000 electrochemical workstation, and impedance spectrum acquisition under float charge conditions was performed according to the set wideband impedance scanning parameters. During the acquisition process, the battery was kept in float charge state to simulate the high float charge voltage state in actual use scenarios. The acquisition process was monitored in real time through the software of the electrochemical workstation. After the acquisition was completed, raw complex impedance response data was obtained, which included information such as the real part, imaginary part, magnitude, and phase angle of the battery impedance at different frequencies. The raw complex impedance response data was processed using MATLAB software. The raw data file was loaded, and noise filtering was performed on the data using MATLAB's built-in filtering functions, such as medfilt1 (median filtering) or movmean (moving average filtering). Taking median filtering as an example, an appropriate window size was selected, such as a window size of 5, and the median was calculated for each data point and its two adjacent data points, thereby removing random noise from the data. After noise filtering, purified impedance spectrum data of lithium batteries were obtained. MATLAB software was used to normalize the purified impedance spectrum data. The max-min normalization method was selected to normalize the real part, imaginary part, magnitude, and phase angle of the impedance at each frequency point to a range of 0 to 1. Specifically, for each characteristic parameter, its maximum and minimum values ​​in the entire dataset were calculated, and the value of each data point was subtracted from the minimum value and then divided by the difference between the maximum and minimum values. In this way, a standard lithium battery impedance spectrum dataset was generated.Based on preset frequency domain segmentation rules, the standard lithium battery impedance spectrum dataset is divided into three intervals: high frequency (1kHz-10kHz), mid frequency (0.1Hz-1kHz), and low frequency (0.01Hz-0.1Hz). For each frequency band, four characteristic parameters are extracted: the real part, imaginary part, magnitude, and phase angle of the impedance. For example, in MATLAB, data points for each frequency band are filtered using logical indexes or loop statements, and the average or eigenvalue of each characteristic parameter within each band is calculated. Finally, the 12 characteristic parameters from the three frequency bands are combined into a 12-dimensional feature vector, which comprehensively reflects the impedance characteristics of the lithium battery in different frequency domains.

[0105] Preferably, the determination of calendar-microcycle aging distribution weights based on lithium battery impedance characteristic data in step S1 includes:

[0106] The wavelet basis function is selected based on the impedance characteristic data of lithium battery, and the selection parameters of the wavelet basis function are obtained.

[0107] Based on the selection parameters of wavelet basis function, continuous wavelet transform is performed on the impedance data of each frequency point in the impedance characteristic data of lithium battery to obtain the time-frequency domain lithium battery impedance coefficient set.

[0108] The wavelet coefficient energy at each scale of the time-frequency domain lithium battery impedance coefficient set is calculated to obtain the energy distribution data at each scale.

[0109] Based on the energy distribution data at each scale, the top 8 main scales with a concentrated energy proportion greater than 5% in the time-frequency domain lithium battery impedance coefficient are selected, and the real part, imaginary part, modulus and phase information of the corresponding wavelet coefficients are extracted to form a time-frequency domain impedance feature vector set.

[0110] The Hilbert-Huang transform is performed on the time-frequency domain impedance eigenvector set to obtain the impedance empirical mode decomposition parameters;

[0111] Based on the impedance empirical mode decomposition parameters, each dimension of the eigenvector set of the impedance eigenvector set in the time-frequency domain is decomposed into EMD one by one to obtain the initial set of intrinsic mode functions.

[0112] The initial set of intrinsic mode functions is subjected to Hilbert transform to obtain the instantaneous frequency and amplitude data of the modes; frequency characteristic statistics are performed on the instantaneous frequency and amplitude data of the modes to obtain a multi-band intrinsic mode function set;

[0113] The weights of aging mechanisms are calculated based on the multi-band intrinsic mode function set to obtain the calendar-microcycle aging distribution weights.

[0114] In this embodiment, MATLAB software is used to perform wavelet analysis on lithium battery impedance characteristic data. The Morlet wavelet is selected as the basis function. In MATLAB, the Morlet function is used to generate the Morlet wavelet basis function, and its parameters are set, such as a center frequency of 5Hz and a bandwidth of 1.5Hz. These parameters are selected based on a preliminary analysis of the spectral characteristics of the lithium battery impedance data, ultimately yielding the wavelet basis function selection parameters. The cwt function in MATLAB is used to perform continuous wavelet transform on the lithium battery impedance characteristic data. The previously selected Morlet wavelet basis function and its parameters are passed to the cwt function to transform the impedance data at each frequency point. During the transformation, the cwt function calculates the wavelet coefficients at different scales, thus obtaining a time-frequency domain lithium battery impedance coefficient set. These coefficient sets contain characteristic information of the impedance data at different time scales and frequencies. By using visualization tools (such as MATLAB's imagesc function) to plot the time-frequency graph of the wavelet coefficients, the changing trend of the impedance data in different frequency bands can be observed intuitively. In MATLAB, energy calculation is performed on the time-frequency domain lithium battery impedance coefficient set. Specifically, the modulus of the wavelet coefficients at each scale is squared to obtain the energy value at that scale. By traversing all scales, the total energy at each scale is calculated and normalized to the range of 0 to 1, thus obtaining energy distribution data for each scale. This data reflects the energy proportion of the impedance signal at different scales, helping to identify key scales that have a significant impact on impedance characteristics. For example, the modulus of wavelet coefficients is calculated using the `abs` function in MATLAB, and the energy value is calculated using the `.^2` operation. Finally, the energy distribution curve is plotted using the `plot` function to visually display the energy distribution at each scale. Based on the obtained energy distribution data, the top 8 main scales with an energy proportion greater than 5% are selected. In MATLAB, these scales are located using logical indices, and the real, imaginary, modulus, and phase information of the corresponding wavelet coefficients are extracted. For example, the `real`, `imag`, `abs`, and `angle` functions are used to extract the real, imaginary, modulus, and phase of the wavelet coefficients, respectively. These features are combined into a feature vector, forming a time-frequency domain impedance feature vector set. Each feature vector contains key features of the impedance signal at the corresponding scale. The time-frequency domain impedance feature vector set is processed using the Hilbert-Huang transform toolbox in MATLAB. Empirical Mode Decomposition (EMD) is performed on each eigenvector. Using MATLAB's `emd` function, the impedance eigenvector is decomposed into several intrinsic mode functions (IMFs), and the decomposition parameters for each IMF are recorded, such as the number of IMFs and the energy percentage of each IMF. These parameters serve as the impedance empirical mode decomposition parameters. The waveform of each IMF is plotted using visualization tools (such as the `plot` function), allowing for a direct observation of the impedance signal's characteristics under different modes.Based on the obtained impedance empirical mode decomposition parameters, EMD decomposition is performed dimension-by-dimensionally on each eigenvector in the time-frequency domain impedance eigenvector set. In MATLAB, the `emd` function is used to decompose each eigenvector, obtaining the corresponding intrinsic mode functions (IMFs). For example, the real part eigenvector of impedance is decomposed into several IMFs using the `emd` function; similarly, EMD decomposition is performed on the imaginary part, magnitude, and phase eigenvectors respectively. The IMFs obtained from all eigenvector decompositions are combined into a set to form the initial set of intrinsic mode functions. The `surf` function is used to plot the three-dimensional time-frequency graph of the IMFs, which can intuitively show the distribution characteristics of each IMF in time and frequency. Hilbert transform is performed on each IMF in the initial set of intrinsic mode functions. In MATLAB, the `hilbert` function is used to transform each IMF to obtain a complex analytic signal. By calculating the instantaneous frequency and amplitude of the complex analytic signal, the instantaneous frequency and amplitude data of the modes are obtained. For example, the `abs` function is used to calculate the instantaneous amplitude, the `angle` function is used to calculate the instantaneous phase, and the instantaneous frequency is obtained by differentiation. Frequency characteristic statistics are performed on these instantaneous frequency and amplitude data, such as calculating the average frequency, frequency range, and amplitude distribution of each IMF. These characteristic data are then organized into a multi-band intrinsic mode function set. Based on the characteristic data of the multi-band IMF set and combined with the aging mechanism of lithium batteries, calendar-microcycle aging distribution weights are calculated, correlating the IMF characteristics of different frequency bands with the aging mechanism. For example, high-frequency IMF characteristics are related to microcycle aging, while low-frequency IMF characteristics are related to calendar aging. The calendar-microcycle aging distribution weights are obtained by calculating the degree of influence of each frequency band IMF characteristic on the overall impedance characteristics. Specifically, a weighted summation of the IMF characteristics of each frequency band can be performed, with the weight determined based on its correlation with the aging mechanism. This yields the calendar-microcycle aging distribution weights, which reflect the degree of influence of different aging mechanisms on the health status of lithium batteries.

[0115] Preferably, step S2 includes the following steps:

[0116] Step S21: Configure a micro-calorimetry measurement device for the energy storage lithium battery system to obtain the parameters of the battery heat flow detection system;

[0117] Step S22: Monitor the heat flow of the energy storage lithium battery system under float charging state based on the parameters of the battery heat flow detection system to obtain the original battery heat flow time series data;

[0118] Step S23: Perform baseline drift correction on the original battery heat flow time series data to obtain purified battery heat flow time series data;

[0119] Step S24: Identify the heat flow peak value of the purified battery heat flow time series data to obtain the side reaction heat flow characteristic points;

[0120] Step S25: Based on the side reaction heat flow characteristic points, pre-configure the DSC test parameters of the energy storage lithium battery system to obtain the DSC test parameter configuration;

[0121] Step S26: According to the DSC test parameters, program the temperature of the battery sample of the energy storage lithium battery system and record the relationship curve between heat flow and temperature to obtain the DSC heat flow curve.

[0122] Step S27: Based on the DSC heat flow curve, predict the evolution of by-reaction products of the energy storage lithium battery system to obtain the parameter set of by-reactions at the positive electrode interface.

[0123] In this embodiment, a MicroCal PEAQ-DSC micro calorimeter from TA Instruments was used for battery thermal flow detection. The micro calorimeter was connected to a computer and configured using its accompanying Origin software. The temperature range was set to 25°C to 60°C in the software to cover the operating temperature range of the battery in float charge mode. The heating rate was set to 5°C / min. The sample cell atmosphere was set to nitrogen with a flow rate of 50 mL / min to prevent oxidation of the battery during testing. After completing the device configuration, the battery thermal flow detection system parameters, including temperature range, heating rate, and atmosphere conditions, were recorded. The lithium-ion battery system under test was placed in the sample cell of the micro calorimeter, ensuring the battery was in float charge mode. Based on the previously set battery thermal flow detection system parameters, the micro calorimeter was started for thermal flow monitoring. During monitoring, the micro calorimeter recorded the changes in battery thermal flow at different temperatures in real time and transmitted the data to the Origin software on the computer. After monitoring was completed, the raw battery thermal flow time-series data was exported from the software. This data contained the thermal flow curves of the battery in float charge mode over time. Baseline drift correction was performed on the raw battery thermal flow time series data using Origin software. The "Baseline Correction" module was selected, and a polynomial fitting method was used to correct the baseline. Specifically, the start and end points of the data curve were selected as reference points for baseline correction, and the polynomial order was set to 3 to ensure accurate fitting of the baseline drift. The software automatically calculated the polynomial fitting curve for the baseline drift and subtracted it from the original data to obtain the purified battery thermal flow time series data. In Origin software, thermal flow peak identification was performed on the purified battery thermal flow time series data. The "Peak Find" function was used, with the peak detection sensitivity set to 0.1 mW. The minimum peak width was set to 1 minute to avoid misidentifying noise as a peak. The software automatically scanned the data curve, identified the thermal flow peaks, and marked the location and size of the side reaction thermal flow characteristic points. These characteristic points correspond to the side reaction thermal flow changes that occur in the battery under float charging conditions. Based on the side reaction thermal flow characteristic points identified in step S24, the DSC test parameters for the energy storage lithium battery system were pre-configured. In the Origin software of the Micro Cal PEAQ-DSC calorimeter, adjust the DSC test temperature range according to the temperature range of the side reaction heat flux characteristic points. For example, if the side reaction heat flux characteristic points are mainly concentrated between 40℃ and 50℃, set the DSC test temperature range to 35℃ to 55℃ to ensure more accurate capture of side reaction heat flux changes. Maintain a heating rate of 5℃ / min, a nitrogen atmosphere, and a flow rate of 50mL / min. After completing the parameter adjustment, record the DSC test parameter configuration. Place the battery sample of the energy storage lithium battery system in the sample cell of the DSC calorimeter, ensuring good thermal contact between the sample and the reference cell.According to the DSC test parameter configuration set in step S25, start the micro calorimeter to perform a programmed temperature rise test. During the test, the micro calorimeter heats the battery sample according to the set heating rate and records the heat flow versus temperature curve in real time. After the test is completed, export the DSC heat flow curve from the Origin software. This curve clearly shows the heat flow changes of the battery at different temperatures, reflecting the thermal effects of the internal side reactions of the battery. For detailed implementation procedures of step S27, please refer to the sub-steps of step S27.

[0124] Preferably, step S27 includes the following steps:

[0125] Step S271: Perform peak separation and fitting on the DSC heat flow curve to obtain the enthalpy change characteristic data of the side reaction, wherein the enthalpy change characteristic data of the side reaction includes peak temperature, peak intensity, peak width and peak area;

[0126] Step S272: Repeated DSC tests were performed on the battery samples of the energy storage lithium battery system at different temperature points at each peak temperature, and the reaction rate constant at different temperatures was calculated to obtain the battery reaction rate constant parameters.

[0127] Step S273: Prepare for X-ray photoelectron spectroscopy (XPS) detection of the energy storage lithium battery system to obtain XPS test condition parameters;

[0128] Step S274: Detect the surface composition of the positive electrode of the energy storage lithium battery system according to the XPS test conditions and parameters to obtain the lithium battery chemical state spectrum data;

[0129] Step S275: Perform peak decomposition and quantification of by-reaction products on the lithium battery chemical state spectrum data to obtain by-reaction product concentration data;

[0130] Step S276: Construct a side reaction model for the energy storage lithium battery system based on the battery reaction rate constant parameter and the side reaction product concentration data to obtain the lithium battery side reaction model;

[0131] Step S277: Based on the preset prediction period and lithium battery side reaction model, predict the evolution of side reaction products of the energy storage lithium battery system to obtain positive electrode interface degradation trend data; extract and integrate parameters from the positive electrode interface degradation trend data to obtain a positive electrode interface side reaction parameter set.

[0132] In this embodiment, DSC heat flux data obtained using a MicroCal PEAQ-DSC microcalorimeter from TA Instruments were used for peak separation and fitting via Origin software. In Origin, the "Peak Analysis" tool was selected from the "Analyze" menu, and the DSC heat flux data was loaded. The peak detection sensitivity was set to 0.05 mW. A Gaussian function was used to fit each detected peak, and the fitting parameters were adjusted to ensure a goodness of fit of over 95% between the fitted curve and the original data. Based on the fitting results, the peak temperature, peak intensity, peak width, and peak area of ​​each peak were extracted; these data constitute the enthalpy change characteristic data of the side reaction. For example, a typical side reaction peak has the characteristics of a peak temperature of 45 °C, a peak intensity of 0.2 mW, a peak width of 5 °C, and a peak area of ​​1.5 mW·℃. For each side reaction peak temperature obtained in step S271, for example, 45℃, multiple temperature points (e.g., 40℃, 42.5℃, 45℃, 47.5℃, 50℃) were selected near this temperature to perform repeated DSC tests on the energy storage lithium battery system. A TA Instruments MicroCal PEAQ-DSC micro calorimeter was used, maintaining a heating rate of 2℃ / min, a nitrogen atmosphere, and a flow rate of 50 mL / min. The test was repeated three times at each temperature point. After the test, the heat flow data at different temperatures were analyzed using the Arrhenius equation to calculate the reaction rate constant. For example, at 45℃, the reaction rate constant was 1.2 × 10⁻⁶. -3 s -1 At 50℃, it is 2.5×10 -3 s -1 Before performing X-ray photoelectron spectroscopy (XPS) detection, the surface of the energy storage lithium battery system needs to be treated. An ion beam etching instrument is used to clean the surface of the battery's positive electrode to remove the surface oxide layer. A ThermoFisher Scientific K-Alpha XPS system is selected for detection. XPS test parameters are set, including an Al Kα X-ray source (1486.6 eV), a beam diameter of 400 μm, and a vacuum level better than 1 × 10⁻⁶. -8Torr. Set the acquisition angle of the electron energy analyzer to 0° and the energy resolution to 0.5 eV. Place the surface-treated lithium-ion battery cathode sample on the XPS system's sample stage. Start the instrument to detect the cathode surface composition according to the set XPS test parameters. The instrument automatically acquires XPS spectrum data, including core energy spectra such as C1s, O1s, and Li1s. For example, the C1s spectrum shows a CC peak at 284.8 eV and a CO peak at 286.5 eV, while the O1s spectrum shows an O2- peak at 530.0 eV. Use XPS Peak Fit software to decompose the obtained XPS spectrum data. Taking the C1s spectrum as an example, select the Gaussian-Lorentz mixture function to fit the spectrum, decomposing it into peaks of different chemical states such as CC, CO, and C=O. Calculate the area ratio of each peak to obtain the relative concentration of different chemical states. For example, the CO peak accounts for 30% of the area, and the C=O peak accounts for 20%. Based on the total carbon content of the cathode material, the absolute concentration of by-reaction products (such as lithium carbonate) can be calculated. Assuming the total carbon content of the cathode material is 10%, the concentration of lithium carbonate is 2%.

[0133] By combining the obtained reaction rate constant and by-product concentration data, a lithium battery by-product model was constructed using MATLAB software. The model, based on the Arrhenius equation and the law of conservation of mass, describes the relationship between the formation rate of by-products and temperature and concentration. For example, it is assumed that the exponential relationship between the formation rate of lithium carbonate and temperature is k(T) = k0exp(-E). a / RT), where k0 is the frequency factor, E a Let R be the activation energy, R be the gas constant, and T be the absolute temperature. The model parameters k0 and E are determined by fitting experimental data. a This yields a complete lithium battery side reaction model. Based on the constructed model, a prediction period of 6 months is set. MATLAB software is used to predict the evolution of side reaction products in the energy storage lithium battery system within this period. The concentration change of side reaction products at each time point is calculated using numerical integration. For example, the prediction results show that the concentration of lithium carbonate increases from 2% to 5% within 6 months. Parameters are extracted from the prediction results, including the concentration change rate and the maximum concentration, and integrated into a set of side reaction parameters at the cathode interface.

[0134] Preferably, step S3, which involves configuring a potential step relaxation test device for the energy storage lithium battery system and quantifying the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test device, includes:

[0135] The potential step relaxation test equipment for the energy storage lithium battery system was configured to obtain the parameters of the electrochemical workstation;

[0136] Based on the parameters of the electrochemical workstation, the micro-circulation potential of the energy storage lithium battery system was monitored to obtain the raw data of the negative electrode potential response.

[0137] Noise filtering is applied to the raw negative electrode potential response data to obtain purified lithium battery potential relaxation data;

[0138] Relaxation curves were fitted to the potential relaxation data of purified lithium batteries to obtain potential relaxation characteristic parameters.

[0139] Lithium plating criteria are established based on potential relaxation characteristic parameters to obtain the critical lithium plating potential threshold.

[0140] The degree of lithium plating in the energy storage lithium battery system is quantified based on the critical potential threshold for lithium plating and the potential relaxation data of purified lithium batteries, thus obtaining lithium battery lithium plating degree data.

[0141] In this embodiment, a Gamry Reference 3000 electrochemical workstation was used for potential step relaxation testing. The energy storage lithium battery system was connected to the test port of the electrochemical workstation. In the Gamry software interface, test parameters were configured, including setting the potential step amplitude to 10mV, the step duration to 10 seconds, and the sampling frequency to 10Hz. After configuration, the parameter settings of the electrochemical workstation were recorded. Microcirculation potential monitoring was initiated according to the configured electrochemical workstation parameters. During monitoring, the electrochemical workstation applied a potential step to the battery according to the set step amplitude and duration, and recorded the negative electrode potential response in real time. During monitoring, the battery was kept in a float charge state to simulate a real-world usage scenario. After monitoring, the raw negative electrode potential response data was exported from the Gamry software. This data included the relaxation process after the potential step. MATLAB software was used to perform noise filtering on the raw negative electrode potential response data. In MATLAB, the raw data file was loaded, and the movmean function was used for moving average filtering. Choosing an appropriate window size, such as 5 data points, the average of each data point and its two adjacent data points is calculated to remove random noise from the data. After filtering, purified lithium battery potential relaxation data is obtained. In MATLAB, relaxation curves are fitted to the purified lithium battery potential relaxation data. The `fit` function is used to fit the data to an exponential decay model (such as a single-exponential or double-exponential model). For example, for a single-exponential decay model, the fitting formula is V(t) = V0 + Aexp(-t / τ), where V0 is the initial potential, A is the amplitude, and τ is the relaxation time constant. Through fitting, potential relaxation characteristic parameters are obtained, including the relaxation time constant τ and the amplitude A. These parameters reflect the speed and magnitude of the potential relaxation process. Based on the obtained potential relaxation characteristic parameters and existing lithium plating research data, a lithium plating criterion is established. By analyzing the relaxation time constant τ and amplitude A at different potentials, the critical potential threshold for lithium plating is determined. For example, experiments have shown that lithium plating occurs in the battery when the relaxation time constant τ is less than 0.5 seconds and the amplitude A is greater than 10 mV. Therefore, the critical potential threshold for lithium plating is set to V. threshold=V0-10mV, where V0 is the initial potential. This threshold effectively distinguishes between normal potential relaxation and potential changes caused by lithium plating. Using MATLAB software, the degree of lithium plating in the energy storage lithium battery system is quantified by combining the obtained critical potential threshold for lithium plating with the potential relaxation data of the purified lithium battery. Specifically, the purified potential relaxation data is compared with the critical potential threshold for lithium plating, and the number of data points exceeding the threshold and their duration are counted. For example, if the proportion of data points exceeding the threshold reaches 10% and the duration exceeds 1 second, it is determined that the battery has obvious lithium plating. Based on the degree of exceeding the threshold, the degree of lithium plating is quantified, for example, the degree of lithium plating is divided into three levels: mild (1%-5%), moderate (5%-10%), and severe (>10%), ultimately obtaining the lithium battery lithium plating degree data.

[0142] Preferably, step S3, which predicts lithium battery capacity degradation based on lithium battery lithium plating level data, includes:

[0143] The energy storage lithium battery system was prepared for nuclear magnetic resonance spectroscopy detection to obtain NMR test parameter configuration; based on the NMR test parameter configuration, the energy storage lithium battery system was quantitatively detected for lithium metal to obtain 7Li NMR spectrum data;

[0144] Peak decomposition and lithium deposition quantification were performed on the 7Li NMR spectrum data to obtain lithium deposition data of the lithium battery anode.

[0145] Based on lithium plating data from lithium battery anodes, scanning electron microscope (SEM) configurations were performed on energy storage lithium battery systems to obtain SEM detection parameters.

[0146] The surface morphology of the negative electrode of the energy storage lithium battery system was scanned based on the scanning electron microscope detection parameters to obtain lithium deposition morphology characteristic data.

[0147] A simulation model of the energy storage lithium battery system was constructed to obtain the MD simulation parameter configuration; based on the MD simulation parameter configuration and the positive electrode interface side reaction parameter set, the cross-catalytic effect of the energy storage lithium battery system was investigated to obtain intermolecular interaction energy data.

[0148] Catalytic reaction simulation of energy storage lithium battery system was performed based on intermolecular interaction energy data to obtain cross-catalytic reaction activation energy data;

[0149] The coupling effect was evaluated based on lithium battery lithium plating degree data, lithium battery anode lithium plating data, and cross-catalytic reaction activation energy data to obtain positive and negative electrode interface coupling parameters.

[0150] Based on the coupling parameters of the positive and negative electrode interfaces, capacity decay prediction of energy storage lithium battery systems is performed to obtain prediction data of the inflection point of capacity drop.

[0151] In this embodiment, a Bruker AVANCE III 400MHz NMR spectrometer was used for quantitative detection of lithium metal. The negative electrode material from the energy storage lithium battery system was removed and sample prepared. NMR test parameters were configured in the Bruker TopSpin software, including setting the magnetic field strength to 400MHz, selecting the standard spin-echo sequence (CPMG) pulse sequence, a repetition time of 2 seconds, an echo time of 0.5 milliseconds, and 128 scans. These parameters were chosen based on optimal detection performance for the lithium metal signal. After configuration, the NMR test parameter settings were recorded. The prepared negative electrode sample was placed in the sample tube of the NMR spectrometer, and the Bruker AVANCE III 400MHz spectrometer was started. Quantitative detection of lithium metal was performed according to the set test parameters. During the detection process, the spectrometer automatically acquired the 7Li NMR signal and generated an NMR spectrum. After detection, the 7Li NMR spectrum data, containing the characteristic signals of lithium metal, was exported from the TopSpin software. Peak decomposition of the 7Li NMR spectrum data was performed using Origin software. In Origin, select the "Peak Analysis" tool from the "Analyze" menu and load the NMR spectrum data. Set the peak detection sensitivity to 0.01. Use the Lorentz function to fit each detected peak, adjusting the fitting parameters to ensure the goodness of fit between the fitted curve and the original data is above 95%. Extract the peak area of ​​the lithium metal signal from the fitting results, and calculate the lithium metal content in the negative electrode material by combining the sample weight and the molar mass of lithium. For example, assuming a peak area of ​​1000, a sample weight of 1 gram, and a lithium molar mass of 6.94 g / mol, the lithium metal content is 15%. These data constitute the lithium deposition data for the lithium-ion battery negative electrode. Based on the obtained lithium deposition data for the lithium-ion battery negative electrode, select a suitable scanning electron microscope (SEM) to scan the surface morphology of the negative electrode. Use a Zeiss Supra 55 field emission scanning electron microscope, setting the accelerating voltage to 5 kV, the working distance to 10 mm, and the magnification to 1000x. These parameters are chosen based on the optimal observation effect on the negative electrode surface morphology. Select the secondary electron imaging mode to obtain a clear image of the surface morphology. After completing the configuration, record the scanning electron microscope (SEM) detection parameters. Place the processed negative electrode sample on the sample stage of the Zeiss Supra 55 SEM, start the SEM, and scan the surface morphology of the negative electrode according to the set detection parameters. During the scanning process, the SEM automatically acquires secondary electron signals and generates morphological images of the negative electrode surface. After the scan is completed, export the morphological images from the SEM's control software. These images clearly show the lithium plating morphology characteristics of the negative electrode surface, including the size, distribution, and morphology of the lithium plating particles. For example, the image shows that the lithium plating particles are distributed in a dendritic pattern with an average diameter of 1 micrometer. These data are the lithium plating morphology characteristic data.Molecular dynamics (MD) simulation models of the energy storage lithium battery system were constructed using Materials Studio software. In Materials Studio, a suitable force field (such as the COMPASS force field) and model building module (such as the AmorphousCell module) were selected to construct a molecular model of the anode material. The model size was set to 10 nm × 10 nm × 10 nm, containing approximately 1000 lithium atoms and corresponding electrolyte molecules. Based on the side reaction parameter set at the cathode interface, the chemical bond parameters and reactive sites in the model were adjusted. Simulation parameters were set, including a temperature of 300 K, a pressure of 1 atm, and a simulation time of 10 nanoseconds. After completing the model construction and parameter configuration, the MD simulation parameter configuration was recorded. Materials Studio software was launched, and the molecular dynamics simulation was run according to the MD simulation parameter configuration set in step 6. During the simulation, the software calculated the intermolecular interaction energies in real time, including van der Waals forces, Coulomb interactions, and hydrogen bonds. After the simulation was completed, the intermolecular interaction energy data were extracted from the analysis module of Materials Studio. For example, the interaction energy between lithium atoms and electrolyte molecules in the anode material is -0.5 eV. These data reflect the interaction strength between the negative electrode material and the electrolyte. Catalytic reaction simulations of the energy storage lithium battery system were performed using Gaussian 16 software. In Gaussian 16, the obtained intermolecular interaction energy data were input, and a suitable theoretical method (such as B3LYP / 6-31G*) was selected for reaction path scanning. The activation energy data of the cross-catalytic reaction was obtained by calculating the energy changes along the reaction path. For example, the simulation results showed that the activation energy of the cross-catalytic reaction between lithium atoms in the negative electrode material and electrolyte molecules was 0.8 eV. Combining the obtained lithium battery lithium plating degree data, lithium battery negative electrode lithium plating data, and cross-catalytic reaction activation energy data, the coupling effect was evaluated using MATLAB software. In MATLAB, a coupling model was established that included the lithium plating degree, negative electrode lithium plating content, and cross-catalytic reaction activation energy. Through model calculations, the coupling parameters of the positive and negative electrode interfaces were obtained, such as a coupling strength coefficient of 0.6 and a coupling reaction rate constant of 1.2 × 10⁻³. s-1. These parameters reflect the interaction strength and reaction rate between the positive and negative electrode interfaces. Using MATLAB software, a capacity decay prediction model was constructed based on the obtained positive and negative electrode interface coupling parameters. The model, based on electrochemical kinetic equations and coupling parameters, describes the change in battery capacity over time. Through numerical simulation, the capacity decay at different usage cycles was predicted. For example, the prediction results show that at the 100th cycle, the battery capacity will decrease to 80% of the initial capacity, and at the 200th cycle, the capacity will decrease to 60%. Based on the prediction results, the capacity drop inflection point was determined to be at the 150th cycle, at which point the rate of capacity decay accelerates significantly. These data constitute the capacity drop inflection point prediction data.

[0152] Preferably, step S4 includes the following steps:

[0153] Step S41: Data synchronously acquires data from the energy storage lithium battery system according to the preset three-level time scale acquisition system to obtain lithium battery time-series synchronous acquisition parameters;

[0154] Step S42: Align the degradation information time of the lithium battery time-series synchronous acquisition parameters according to the calendar-micro-cycle aging distribution weight to obtain the lithium battery synchronous degradation data table;

[0155] Step S43: Extract degradation feature vectors from the lithium battery synchronous degradation data table to obtain lithium battery degradation feature data;

[0156] Step S44: Establish a state-space model of lithium battery based on lithium battery degradation characteristic data;

[0157] Step S45: Solve the state covariance of the lithium battery state space model based on the lithium battery degradation characteristic data to obtain the lithium battery state covariance data;

[0158] Step S46: Perform covariance correction on the lithium battery state covariance data to obtain the Kalman filter parameter set;

[0159] Step S47: Model the state space of the lithium battery based on the Kalman filter parameter set to obtain the lithium battery health state transition equation;

[0160] Step S48: Evaluate the health status of the lithium battery based on the lithium battery health status transition equation to obtain key health indicators of the lithium battery.

[0161] In this embodiment, a three-level timescale acquisition system is constructed using an NI (National Instruments) data acquisition card and LabVIEW software. The three timescales are: short-term (1 minute), medium-term (1 hour), and long-term (1 day). Short-term data is used to capture rapid dynamic changes, medium-term data is used to monitor short-term cyclic effects, and long-term data is used to assess the overall aging trend. The data acquisition card is configured using LabVIEW software, with sampling frequencies set to 10Hz for short-term, 1Hz for medium-term, and 0.1Hz for long-term. Voltage, current, and temperature sensors of the energy storage lithium battery system are connected to the data acquisition card to synchronously acquire these parameters. After acquisition, the lithium battery time-series synchronous acquisition parameters are obtained, including voltage, current, and temperature data at different timescales. The obtained lithium battery time-series synchronous acquisition parameters are processed using MATLAB software. Based on the determined calendar-micro-cycle aging distribution weights, the data at different timescales are time-aligned. Specifically, the long-term data is used as a baseline, and the time axes of the medium-term and short-term data are adjusted according to the aging weights. For example, if the calendar aging weight is 0.7 and the microcirculation aging weight is 0.3, then during time alignment, the time axis of the medium-term data will converge 70% towards the long-term data, and the time axis of the short-term data will converge 30% towards the long-term data. After time alignment, the aligned data is organized into a table, namely the lithium battery synchronous degradation data table, which includes time-aligned voltage, current, and temperature data. In MATLAB, feature extraction is performed on the obtained lithium battery synchronous degradation data table. Principal component analysis (PCA) is used to extract degradation feature vectors. The data table is standardized so that the mean of each feature is 0 and the standard deviation is 1. The covariance matrix of the data is calculated and eigenvalue decomposition is performed to obtain the principal components and their corresponding eigenvalues. Principal components with a cumulative contribution rate of 95% are selected as degradation feature vectors. For example, assuming that after PCA analysis, the cumulative contribution rate of the first three principal components is 95%, these three principal components are extracted as degradation feature vectors. These feature vectors can effectively reflect the key characteristics of battery degradation, obtaining lithium battery degradation feature data. The lithium battery state-space model is built using MATLAB's System Identification Toolbox. Using the obtained lithium battery degradation characteristic data as input data, the model recognition function in the toolbox is used to select an appropriate model structure (such as ARX, ARMAX, or a state-space model). For example, a state-space model structure is selected, and the model order is set to 3 (based on the number of feature vectors). The toolbox will automatically fit the model parameters based on the input data to obtain a lithium battery state-space model. This model can describe the dynamic changes in the battery state. In MATLAB, the state covariance of the established lithium battery state-space model is solved. The lyap function is used to solve the continuous-time Lyapunov equations to obtain the state covariance matrix.The specific operation involves inputting the system matrix (A matrix) and noise covariance matrix (Q matrix) of the state-space model into the `lyap` function to calculate the state covariance matrix P. For example, assuming the system matrix A is a 3×3 matrix and the noise covariance matrix Q is an identity matrix, the state covariance matrix P is calculated using the `lyap(A,Q)` function. The state covariance matrix P reflects the uncertainty of the battery state. MATLAB is used to perform covariance correction on the obtained lithium battery state covariance data. The state covariance matrix P is adjusted to adapt to the actual measurement data through the design of a Kalman filter. Specifically, based on the measurement noise covariance matrix R and the process noise covariance matrix Q, the Kalman gain K and the updated state covariance matrix P are calculated using the `kalman` function. For example, assuming the measurement noise covariance matrix R is a 3×3 diagonal matrix with diagonal elements of 0.1, 0.2, and 0.3, the Kalman gain K and the updated state covariance matrix P are calculated using the `kalman(A,C,Q,R)` function. The updated state covariance matrix P and Kalman gain K constitute the Kalman filter parameter set. Based on the obtained Kalman filter parameter set, the lithium battery state-space model is reconstructed using the MATLAB `ss` function. The updated state covariance matrix P and Kalman gain K are used as model parameter inputs to obtain the corrected state-space model. Through this model, the lithium battery health state transition equation can be obtained, describing the change of the battery health state over time. For example, suppose the system matrix A, input matrix B, output matrix C, and direct transfer matrix D are respectively: If C = 1 0 0 and D = 0, then the state-space model is constructed using the ss(A,B,C,D) function, and the lithium battery health state transition equation is obtained. For detailed implementation of step S48, please refer to the sub-steps of step S48.

[0162] Of particular importance, step S48 also includes the following steps:

[0163] Step S481: Perform variational Bayesian inference based on the lithium battery health state transition equation to obtain the variational Bayesian prior distribution parameters;

[0164] Step S482: Perform variational inference on the energy storage lithium battery system based on the variational Bayesian prior distribution parameters and the lithium battery health state transition equation to obtain approximate data of the posterior distribution of the health state.

[0165] Step S483: Perform Markov chain Monte Carlo sampling on the approximate data of the posterior distribution of health status to obtain the posterior distribution data of health status;

[0166] Step S484: Based on the posterior distribution data of the health status and the parameter set of the side reaction at the positive electrode interface, feature fusion is performed to obtain a comprehensive lithium battery health feature vector;

[0167] Step S485: Based on the comprehensive lithium battery health feature vector and capacity drop inflection point prediction data, feature importance assessment is performed to obtain lithium battery feature weight distribution data;

[0168] Step S486: Perform principal component dimensionality reduction on the comprehensive lithium battery health feature vector based on the lithium battery feature weight distribution data to obtain the dimensionality-reduced lithium battery feature covariance data;

[0169] Step S487: Perform principal component projection transformation on the dimensionality-reduced lithium battery feature covariance data to obtain the lithium battery feature principal component score; collect key health indicators of the energy storage lithium battery system based on the lithium battery feature principal component score to obtain the key health indicators of the lithium battery.

[0170] In this embodiment, variational Bayesian inference is performed using MATLAB software. Based on the obtained lithium battery health state transition equation, the system's state variables and observation variables are defined. State variables are defined as battery health state indicators (e.g., capacity, internal resistance), and observation variables are defined as the actual measured battery performance data (e.g., voltage, current). In MATLAB, the `fitdist` function is used to fit the distribution of the observation data, assuming the data follows a Gaussian distribution, to obtain the mean and variance. According to variational Bayesian theory, an appropriate prior distribution (e.g., Gaussian distribution) is selected and the prior distribution parameters are initialized. For example, the mean of the prior distribution is assumed to be 0.5, and the variance to be 0.1. These parameters serve as the variational Bayesian prior distribution parameters. In MATLAB, the variational Bayesian inference method is used to infer the health state of the energy storage lithium battery system. Based on the obtained variational Bayesian prior distribution parameters and the health state transition equation, a variational Bayesian model is constructed. The parameters of the posterior distribution are updated using an iterative optimization algorithm (e.g., coordinate ascent method) until convergence. The specific operation involves iterative calculation using the `vbem` function (Variational Bayesian Expectation-Maximization algorithm) in MATLAB. The initial iteration error is assumed to be 0.01, and the maximum number of iterations is 100. After multiple iterations, approximate posterior distribution data of the healthy state is obtained, including the mean and variance of the posterior distribution. Markov chain Monte Carlo (MCMC) sampling is then performed using MATLAB's Statistics and Machine Learning Toolbox. Based on the obtained approximate posterior distribution data of the healthy state, initial parameters for MCMC sampling are set, including the sampling step size and the number of samples. For example, the Metropolis-Hastings algorithm is selected for sampling, with a sampling step size of 0.05 and a sampling count of 10,000. Sampling is then performed using the `mhsample` function to obtain the posterior distribution data of the healthy state. In MATLAB, the obtained posterior distribution data of the healthy state is fused with the positive electrode interface side reaction parameter set. The positive electrode interface side reaction parameter set includes data such as the concentration of side reaction products and the reaction rate constant. By using feature concatenation, posterior distribution data (such as mean and variance) and the set of side reaction parameters are combined into a comprehensive feature vector. For example, assuming the posterior distribution has a mean of 0.6, a variance of 0.05, a side reaction product concentration of 0.1, and a reaction rate constant of 0.01, the comprehensive feature vector would be [0.6, 0.05, 0.1, 0.01]. The machine learning toolbox in MATLAB is used to evaluate the feature importance of the comprehensive lithium battery health feature vector. A feature importance evaluation model is constructed by combining the obtained capacity drop inflection point prediction data. For example, a random forest algorithm is used for feature importance evaluation. In MATLAB, the Tree Bagger function is used to create a random forest model, setting the number of trees to 100. The importance score for each feature is obtained through the model's OOBPredict or Importance attributes.Assuming the importance scores of the four features in the comprehensive feature vector are 0.4, 0.3, 0.2, and 0.1, the lithium battery feature weight distribution data is [0.4, 0.3, 0.2, 0.1]. These weight data reflect the degree of influence of different features on the battery health status. In MATLAB, Principal Component Analysis (PCA) is used to reduce the dimensionality of the comprehensive lithium battery health feature vector. Based on the obtained feature weight distribution data, the two features with the largest weights are selected for principal component analysis. The pca function is used to reduce the dimensionality of the comprehensive feature vector to obtain the dimensionality-reduced feature covariance data. For example, assuming the comprehensive feature vector is [0.6, 0.05, 0.1, 0.01] and the feature weight distribution is [0.4, 0.3, 0.2, 0.1], the first two features are selected for PCA analysis. In MATLAB, principal component projection transformation is performed on the obtained dimensionality-reduced lithium battery feature covariance data. The score of each principal component is obtained through the output of the pca function. For example, assuming the dimensionality-reduced feature covariance data is [0.6, 0.05], the score after principal component projection transformation is [0.8, 0.2]. These scores reflect the main characteristics of the battery's health status. Based on the principal component scores, the principal components with higher scores are selected as key health indicators. For example, selecting the first principal component as the key health indicator, its score of 0.8 represents the main trend of battery health status changes, ultimately yielding the key health indicators for lithium batteries.

[0171] Preferably, step S5 includes the following steps:

[0172] Step S51: Configure the support vector machine classifier parameters for the key health indicators of lithium batteries to obtain the SVM classifier parameter configuration;

[0173] Step S52: Construct training samples for key health indicators of lithium batteries according to the SVM classifier parameter configuration to obtain the SVM classification training dataset;

[0174] Step S53: Train the SVM model based on the SVM classification training dataset to obtain the SVM classifier model; perform cross-validation evaluation on the SVM classifier model to obtain SVM classifier performance evaluation data.

[0175] Step S54: Classify the health status of the energy storage lithium battery system based on the SVM classifier performance evaluation data and key health indicators of the lithium battery to obtain health status level classification data.

[0176] Step S55: Based on the health status level classification data, perform historical data statistics on the energy storage lithium battery system to obtain the health threshold adjustment benchmark data;

[0177] Step S56: Based on the health threshold adjustment benchmark data, perform adaptive threshold calculation on the energy storage lithium battery system to obtain the dynamic health threshold parameters of the lithium battery;

[0178] Step S57: Based on the dynamic threshold parameters of lithium battery health and the classification data of health status level, set early warning rules for the energy storage lithium battery system to obtain lithium battery health early warning rules;

[0179] Step S58: Based on the lithium battery health early warning rules and capacity drop inflection point prediction data, assess the remaining life of the energy storage lithium battery system to obtain lithium battery corrected life prediction data.

[0180] In this embodiment, the Support Vector Machine (SVM) classifier parameters are configured using MATLAB's Machine Learning Toolbox. Based on the obtained key health indicators of the lithium battery, a suitable kernel function is selected. For example, the Radial Basis Function (RBF) is chosen as the kernel function because it is suitable for nonlinear classification problems. In MATLAB, the `fitcsvm` function is used to configure the SVM classifier. The kernel function parameters are set, such as the penalty parameter C and the kernel width σ. For example, C = 1 and σ = 0.5. These parameter selections are based on preliminary analysis and experience with battery health status data. After configuration, the SVM classifier parameter configuration is recorded. Based on the configured SVM classifier parameters, MATLAB is used to construct training samples for the key health indicators of the lithium battery. Battery samples with known health states are selected from historical data, including normal, mildly aged, and severely aged states. It is assumed that there are 100 samples for each state, and each sample contains two key health indicators (such as principal component scores). These sample data are divided into a training set and a test set, with the training set accounting for 70% and the test set accounting for 30%. In MATLAB, the `cvpartition` function is used to partition the data. After partitioning, an SVM classification training dataset is obtained. The model is trained in MATLAB using this dataset. The `fitcsvm` function is used to train the SVM model according to the configured parameters (e.g., C=1, σ=0.5). After training, the SVM classifier model is obtained. To evaluate the model's performance, cross-validation is used. In MATLAB, the `crossval` function is used to perform 10-fold cross-validation. The SVM classifier performance evaluation data is obtained by calculating the cross-validation accuracy, recall, and F1 score. For example, assuming an average cross-validation accuracy of 90%, recall of 85%, and F1 score of 87%. Based on the obtained SVM classifier performance evaluation data and key health indicators of lithium batteries, MATLAB is used to classify the health status of energy storage lithium battery systems. The key health indicators from the test set are input into the trained SVM classifier model, and the model outputs the health status category for each sample (e.g., normal, mildly aged, severely aged). Based on the model's output, the health status levels of the energy storage lithium battery system are classified. For example, suppose a test set contains 30 samples classified as normal, 20 samples as mildly aged, and 10 samples as severely aged. These classification results constitute the health status level classification data. In MATLAB, historical data statistics are performed on the obtained health status level classification data. The distribution of each health status (normal, mildly aged, severely aged) in the historical data is analyzed. For example, the normal status samples account for 60% of the total samples, the mildly aged status accounts for 30%, and the severely aged status accounts for 10%. The mean and standard deviation of the key health indicators for each status are calculated.For example, the mean of key health indicators in the normal state is 0.8, and the standard deviation is 0.1; the mean is 0.6, and the standard deviation is 0.2; and the mean is 0.4, and the standard deviation is 0.3. These statistical results serve as benchmark data for adjusting health thresholds. Based on the obtained benchmark data for health threshold adjustment, MATLAB is used to perform adaptive threshold calculations on the energy storage lithium battery system. Dynamic thresholds are calculated according to the mean and standard deviation of key health indicators for each health state. For example, assuming the threshold for the normal state is the mean minus one standard deviation (0.8 - 0.1 = 0.7), the threshold for the mildly aged state is 0.6 - 0.2 = 0.4, and the threshold for the severely aged state is 0.4 - 0.3 = 0.1. These thresholds can be dynamically adjusted according to the distribution of historical data to adapt to changes in different health states, and the final lithium battery health dynamic threshold parameters are [0.7, 0.4, 0.1]. In MATLAB, lithium battery health early warning rules are set based on the obtained lithium battery health dynamic threshold parameters and health state level classification data. For example, a mild aging warning is issued when the key health indicator is below 0.7; a severe aging warning is issued when it is below 0.4; and an emergency replacement warning is issued when it is below 0.1. These warning rules can be adjusted in real time according to dynamic threshold parameters to adapt to changes in battery health status. The final lithium battery health warning rules are: Mild aging warning: Key health indicator < 0.7; Severe aging warning: Key health indicator < 0.4; Emergency replacement warning: Key health indicator < 0.1. For detailed implementation procedures of step S58, please refer to the sub-steps of step S58.

[0181] Of particular importance, step S58 also includes the following steps:

[0182] Step S581: Initialize the particle filter algorithm parameters for the lithium battery health warning rules to obtain particle filter configuration data;

[0183] Step S582: Initialize the particles according to the particle filter configuration data and the lithium battery health warning rules to obtain the initial lithium battery particle set;

[0184] Step S583: Based on the initial lithium battery particle set, predict the particle state to obtain predicted lithium battery particle state data;

[0185] Step S584: Update and resample the predicted lithium battery particle state data to obtain the posterior lithium battery particle distribution data.

[0186] Step S585: Based on the posterior lithium battery particle distribution data, perform a probability prediction of the remaining lifetime of the energy storage lithium battery system to obtain lithium battery lifetime prediction trajectory data.

[0187] Step S586: Based on the lithium battery life prediction trajectory data and the capacity drop inflection point prediction data, perform coupled aging correction to obtain the lithium battery corrected life prediction data.

[0188] In this embodiment, MATLAB software is used to initialize the parameters of the particle filter algorithm. Based on the established lithium battery health warning rules, the basic parameters of the particle filter algorithm are determined. The total number of particles is set to 1000, representing different values ​​of the battery health state. The particle state distribution is initialized, assuming the initial state follows a Gaussian distribution with a mean of 0.8 (close to a healthy state) and a standard deviation of 0.1. The resampling threshold for the particle filter is set to 0.5; a resampling operation is triggered when the number of effective particles falls below this threshold. After these parameters are configured, the initial configuration data of the particle filter is obtained. Based on the obtained particle filter configuration data, MATLAB is used for particle initialization. According to the initialized particle state distribution (Gaussian distribution, mean 0.8, standard deviation 0.1), 1000 particles are randomly generated, each representing a battery health state. The initial weights of these particles are set to be equal, i.e., each particle has a weight of 1 / 1000. These particles and their weights are organized into a set, i.e., the initial lithium battery particle set. This set contains the initial estimate of the battery health state. In MATLAB, the state prediction is performed on the obtained initial lithium battery particle set. Based on a dynamic model of battery health (e.g., assuming health degrades at a certain rate), the state of each particle is updated. Assuming a degradation rate of 0.01 per month and a prediction period of one month, the new state of each particle is its initial state minus 0.01. For example, a particle with an initial state of 0.82 has a predicted state of 0.81. After updating the states of all particles, the predicted lithium battery particle state data is obtained. In MATLAB, the weights of the obtained predicted lithium battery particle state data are updated. Based on new measurement data (e.g., new health indicator measurements), the weights of each particle are updated using a likelihood function. Assuming a measurement value of 0.8 and a measurement error following a Gaussian distribution with a standard deviation of 0.05, the likelihood value of each particle is calculated according to the Gaussian distribution, and its weight is updated. After the weight update, it is checked whether the number of valid particles is below the resampling threshold (0.5). If it is below the threshold, a resampling operation is performed, retaining particles with high weights, discarding particles with low weights, and redistributing weights to finally obtain the posterior lithium battery particle distribution data. In MATLAB, based on the obtained posterior lithium battery particle distribution data, the remaining lifetime of an energy storage lithium battery system is probabilistically predicted. The remaining lifetime for each particle is calculated, assuming the battery life ends when the health state drops to 0.2. The remaining lifetime of each particle is predicted based on its health state and degradation rate. For example, a particle with a health state of 0.81 is expected to have a remaining lifetime of 60 months ((0.81-0.2) / 0.01). Statistical analysis of the remaining lifetimes of all particles yields the probability distribution of the remaining lifetimes, ultimately providing the predicted trajectory data for the lithium battery lifetime.In MATLAB, coupled aging correction is performed by combining the obtained lithium battery lifetime prediction trajectory data and the obtained capacity drop inflection point prediction data. It is assumed that the capacity drop inflection point is predicted to be at the 50th month, at which point the battery capacity will rapidly decline. Based on this information, the lifetime prediction trajectory is adjusted to make the predicted lifetime more conservative before the drop inflection point. For example, for particles with a predicted lifetime exceeding 50 months, their remaining lifetime is shortened proportionally. After the correction is completed, the corrected lithium battery lifetime prediction data is obtained.

[0189] Preferably, the present invention also provides a health assessment system for an energy storage lithium battery system, used to perform the health assessment method for the energy storage lithium battery system as described above, the health assessment system for the energy storage lithium battery system comprising:

[0190] Impedance analysis module is used to collect wide-frequency domain complex impedance response data of energy storage lithium battery system in float charge state, and perform frequency domain segmentation to obtain lithium battery impedance characteristic data; based on lithium battery impedance characteristic data, calendar-micro-cycle aging distribution weights are determined.

[0191] The heat flow monitoring module is used to perform micro-caloric measurement and detection on the energy storage lithium battery system to obtain real-time battery heat flow time-series data; based on the real-time battery heat flow time-series data, the energy storage lithium battery system is subjected to DSC test to obtain DSC heat flow curve; based on the DSC heat flow curve, the evolution of side reaction products of the energy storage lithium battery system is predicted to obtain the positive electrode interface side reaction parameter set.

[0192] The lithium plating quantification module is used to configure the potential step relaxation test equipment for the energy storage lithium battery system, and to quantify the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test equipment to obtain lithium battery lithium plating degree data; based on the lithium battery lithium plating degree data, the lithium battery capacity decay is predicted to obtain capacity drop inflection point prediction data.

[0193] The health assessment module is used to construct a lithium battery health state transition equation based on calendar-microcirculation aging distribution weights; and to assess the health state of the lithium battery based on the lithium battery health state transition equation to obtain key health indicators of the lithium battery.

[0194] The life prediction module is used to assess the remaining life of energy storage lithium battery systems based on key health indicators and capacity drop inflection point prediction data of lithium batteries, and obtain corrected life prediction data of lithium batteries.

[0195] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0196] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A health assessment method for an energy storage lithium battery system, characterized in that, Includes the following steps: Step S1: Collect wide-frequency domain complex impedance response data of the energy storage lithium battery system in float charging state, and perform frequency domain segmentation to obtain lithium battery impedance characteristic data; Determining the calendar-microcycle aging distribution weights based on lithium battery impedance characteristic data; wherein, step S1, determining the calendar-microcycle aging distribution weights based on lithium battery impedance characteristic data, includes: The wavelet basis function is selected based on the impedance characteristic data of lithium battery, and the selection parameters of the wavelet basis function are obtained. Based on the selection parameters of wavelet basis function, continuous wavelet transform is performed on the impedance data of each frequency point in the impedance characteristic data of lithium battery to obtain the time-frequency domain lithium battery impedance coefficient set. The wavelet coefficient energy at each scale of the time-frequency domain lithium battery impedance coefficient set is calculated to obtain the energy distribution data at each scale. Based on the energy distribution data at each scale, the top 8 main scales with a concentrated energy proportion greater than 5% in the time-frequency domain lithium battery impedance coefficient are selected, and the real part, imaginary part, modulus and phase information of the corresponding wavelet coefficients are extracted to form a time-frequency domain impedance feature vector set. The Hilbert-Huang transform is performed on the time-frequency domain impedance eigenvector set to obtain the impedance empirical mode decomposition parameters; Based on the impedance empirical mode decomposition parameters, each dimension of the eigenvector set of the impedance eigenvector set in the time-frequency domain is decomposed into EMD one by one to obtain the initial set of intrinsic mode functions. The initial set of intrinsic mode functions is subjected to Hilbert transform to obtain the instantaneous frequency and amplitude data of the modes; frequency characteristic statistics are performed on the instantaneous frequency and amplitude data of the modes to obtain a multi-band intrinsic mode function set; The weight of aging mechanism is calculated based on the multi-band intrinsic mode function set to obtain the calendar-micro-cycle aging distribution weight. Step S2: Perform microcalorimetry on the energy storage lithium battery system to obtain real-time battery heat flow time-series data; perform DSC test on the energy storage lithium battery system based on the real-time battery heat flow time-series data to obtain DSC heat flow curve; predict the evolution of side reaction products of the energy storage lithium battery system based on the DSC heat flow curve to obtain the positive electrode interface side reaction parameter set. Step S3: Configure a potential step relaxation test device for the energy storage lithium battery system, and quantify the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test device to obtain lithium battery lithium plating degree data; predict the capacity decay of the lithium battery based on the lithium battery lithium plating degree data to obtain capacity drop inflection point prediction data. Step S4: Construct a lithium battery health state transition equation based on calendar-microcycle aging distribution weights; assess the lithium battery health state based on the lithium battery health state transition equation to obtain key health indicators of the lithium battery; Step S4 includes the following steps: Step S41: Data synchronously acquires data from the energy storage lithium battery system according to the preset three-level time scale acquisition system to obtain lithium battery time-series synchronous acquisition parameters; Step S42: Align the degradation information time of the lithium battery time-series synchronous acquisition parameters according to the calendar-micro-cycle aging distribution weight to obtain the lithium battery synchronous degradation data table; Step S43: Extract degradation feature vectors from the lithium battery synchronous degradation data table to obtain lithium battery degradation feature data; Step S44: Establish a state-space model of lithium battery based on lithium battery degradation characteristic data; Step S45: Solve the state covariance of the lithium battery state space model based on the lithium battery degradation characteristic data to obtain the lithium battery state covariance data; Step S46: Perform covariance correction on the lithium battery state covariance data to obtain the Kalman filter parameter set; Step S47: Model the state space of the lithium battery based on the Kalman filter parameter set to obtain the lithium battery health state transition equation; Step S48: Evaluate the health status of the lithium battery based on the lithium battery health status transition equation to obtain key health indicators of the lithium battery; Step S5: Based on the key health indicators of lithium batteries and the predicted inflection point of capacity drop, assess the remaining life of the energy storage lithium battery system to obtain the corrected life prediction data of the lithium battery.

2. The health assessment method for an energy storage lithium battery system according to claim 1, characterized in that, Step S1 involves collecting wide-frequency domain complex impedance response data of the energy storage lithium battery system under float charging conditions and performing frequency domain segmentation, including: Configure a multi-channel data acquisition interface for the energy storage lithium battery system to obtain the electrochemical impedance spectroscopy test parameter configuration; Based on the configuration of electrochemical impedance spectroscopy test parameters, a wide frequency domain scanning range was set for the energy storage lithium battery system to obtain wide frequency domain impedance scanning parameters. The wide frequency domain setting frequency scanning range is 0.01Hz-10kHz, a logarithmic distribution method is adopted, 10 test points are set for each frequency range, the perturbation signal amplitude is set to 5mV, the phase accuracy is ±0.1°, the impedance accuracy is ±0.1%, and the single scan time is controlled within 30 minutes. Impedance spectrum acquisition of energy storage lithium battery system under float charging state was performed based on wide frequency domain impedance scanning parameters to obtain raw complex impedance response data; Noise filtering was applied to the original complex impedance response data to obtain the purified impedance spectrum data of the lithium battery. Normalize the lithium battery purification impedance spectrum data to generate a standard lithium battery impedance spectrum dataset. According to the preset frequency domain segmentation rules, the standard lithium battery impedance spectrum dataset is segmented in the frequency domain to obtain lithium battery impedance characteristic data. The frequency domain is divided into three intervals: high frequency, mid frequency, and low frequency. For each frequency band, four characteristic parameters are extracted: real part, imaginary part, magnitude, and phase angle, forming a 12-dimensional feature vector.

3. The health assessment method for an energy storage lithium battery system according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Configure a micro-calorimetry measurement device for the energy storage lithium battery system to obtain the parameters of the battery heat flow detection system; Step S22: Monitor the heat flow of the energy storage lithium battery system under float charging state based on the parameters of the battery heat flow detection system to obtain the original battery heat flow time series data; Step S23: Perform baseline drift correction on the original battery heat flow time series data to obtain purified battery heat flow time series data; Step S24: Identify the heat flux peak value of the purified battery heat flux time series data to obtain the side reaction heat flux characteristic points; Step S25: Based on the side reaction heat flow characteristic points, pre-configure the DSC test parameters of the energy storage lithium battery system to obtain the DSC test parameter configuration; Step S26: According to the DSC test parameters, program the temperature of the battery sample of the energy storage lithium battery system and record the relationship curve between heat flow and temperature to obtain the DSC heat flow curve. Step S27: Based on the DSC heat flow curve, predict the evolution of by-reaction products of the energy storage lithium battery system to obtain the parameter set of by-reactions at the positive electrode interface.

4. The health assessment method for an energy storage lithium battery system according to claim 3, characterized in that, Step S27 includes the following steps: Step S271: Perform peak separation and fitting on the DSC heat flow curve to obtain the enthalpy change characteristic data of the side reaction, wherein the enthalpy change characteristic data of the side reaction includes peak temperature, peak intensity, peak width and peak area; Step S272: Repeated DSC tests were performed on the battery samples of the energy storage lithium battery system at different temperature points at each peak temperature, and the reaction rate constant at different temperatures was calculated to obtain the battery reaction rate constant parameters. Step S273: Prepare for X-ray photoelectron spectroscopy (XPS) detection of the energy storage lithium battery system to obtain XPS test condition parameters; Step S274: Detect the surface composition of the positive electrode of the energy storage lithium battery system according to the XPS test conditions and parameters to obtain the lithium battery chemical state spectrum data; Step S275: Perform peak decomposition and quantification of by-reaction products on the lithium battery chemical state spectrum data to obtain by-reaction product concentration data; Step S276: Construct a side reaction model for the energy storage lithium battery system based on the battery reaction rate constant parameter and the side reaction product concentration data to obtain the lithium battery side reaction model; Step S277: Based on the preset prediction period and lithium battery side reaction model, predict the evolution of side reaction products of the energy storage lithium battery system to obtain positive electrode interface degradation trend data; extract and integrate parameters from the positive electrode interface degradation trend data to obtain a positive electrode interface side reaction parameter set.

5. The health assessment method for an energy storage lithium battery system according to claim 1, characterized in that, Step S3 involves configuring a potential step relaxation test device for the energy storage lithium battery system and quantifying the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test device, including: The potential step relaxation test equipment for the energy storage lithium battery system was configured to obtain the parameters of the electrochemical workstation; Micro-circulation potential monitoring of the energy storage lithium battery system was performed based on electrochemical workstation parameters to obtain raw data of the negative electrode potential response. Noise filtering is applied to the raw negative electrode potential response data to obtain purified lithium battery potential relaxation data; Relaxation curves were fitted to the potential relaxation data of purified lithium batteries to obtain potential relaxation characteristic parameters. Lithium plating criteria are established based on potential relaxation characteristic parameters to obtain the critical lithium plating potential threshold. The degree of lithium plating in the energy storage lithium battery system is quantified based on the critical potential threshold for lithium plating and the potential relaxation data of purified lithium batteries, thus obtaining lithium battery lithium plating degree data.

6. The health assessment method for an energy storage lithium battery system according to claim 1, characterized in that, Step S3, which predicts lithium battery capacity degradation based on lithium battery lithium plating level data, includes: The energy storage lithium battery system was prepared for nuclear magnetic resonance spectroscopy detection to obtain NMR test parameter configuration; based on the NMR test parameter configuration, the energy storage lithium battery system was quantitatively detected for lithium metal to obtain 7Li NMR spectrum data; Peak decomposition and lithium deposition quantification were performed on the 7Li NMR spectrum data to obtain lithium deposition data of lithium battery anode. Based on lithium plating data from lithium battery anodes, scanning electron microscope (SEM) configurations were performed on energy storage lithium battery systems to obtain SEM detection parameters. The surface morphology of the negative electrode of the energy storage lithium battery system was scanned based on the scanning electron microscope detection parameters to obtain lithium deposition morphology characteristic data. A simulation model of the energy storage lithium battery system was constructed to obtain the MD simulation parameter configuration; based on the MD simulation parameter configuration and the positive electrode interface side reaction parameter set, the cross-catalytic effect of the energy storage lithium battery system was investigated to obtain intermolecular interaction energy data. Catalytic reaction simulation of energy storage lithium battery system was performed based on intermolecular interaction energy data to obtain cross-catalytic reaction activation energy data; The coupling effect was evaluated based on lithium battery lithium plating degree data, lithium battery anode lithium plating data, and cross-catalytic reaction activation energy data to obtain positive and negative electrode interface coupling parameters. Based on the coupling parameters of the positive and negative electrode interfaces, capacity decay prediction of energy storage lithium battery systems is performed to obtain prediction data of the inflection point of capacity drop.

7. The health assessment method for an energy storage lithium battery system according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: Configure the support vector machine classifier parameters for the key health indicators of lithium batteries to obtain the SVM classifier parameter configuration; Step S52: Construct training samples for key health indicators of lithium batteries according to the SVM classifier parameter configuration to obtain the SVM classification training dataset; Step S53: Train the SVM model based on the SVM classification training dataset to obtain the SVM classifier model; perform cross-validation evaluation on the SVM classifier model to obtain SVM classifier performance evaluation data. Step S54: Classify the health status of the energy storage lithium battery system based on the SVM classifier performance evaluation data and key health indicators of the lithium battery to obtain health status level classification data. Step S55: Based on the health status level classification data, perform historical data statistics on the energy storage lithium battery system to obtain the health threshold adjustment benchmark data; Step S56: Based on the health threshold adjustment benchmark data, perform adaptive threshold calculation on the energy storage lithium battery system to obtain the dynamic health threshold parameters of the lithium battery; Step S57: Based on the dynamic threshold parameters of lithium battery health and the classification data of health status level, set early warning rules for the energy storage lithium battery system to obtain lithium battery health early warning rules; Step S58: Based on the lithium battery health early warning rules and capacity drop inflection point prediction data, assess the remaining life of the energy storage lithium battery system to obtain lithium battery corrected life prediction data.

8. A health assessment system for an energy storage lithium battery system, characterized in that, For performing the health assessment method of the energy storage lithium battery system as described in claim 1, the health assessment system of the energy storage lithium battery system includes: Impedance analysis module is used to collect wide-frequency domain complex impedance response data of energy storage lithium battery system in float charge state, and perform frequency domain segmentation to obtain lithium battery impedance characteristic data; based on lithium battery impedance characteristic data, calendar-micro-cycle aging distribution weights are determined. The heat flow monitoring module is used to perform micro-caloric measurement and detection on the energy storage lithium battery system to obtain real-time battery heat flow time-series data; based on the real-time battery heat flow time-series data, the energy storage lithium battery system is subjected to DSC test to obtain DSC heat flow curve; based on the DSC heat flow curve, the evolution of side reaction products of the energy storage lithium battery system is predicted to obtain the positive electrode interface side reaction parameter set. The lithium plating quantification module is used to configure the potential step relaxation test equipment for the energy storage lithium battery system, and to quantify the degree of lithium plating in the energy storage lithium battery system based on the potential step relaxation test equipment to obtain lithium battery lithium plating degree data; based on the lithium battery lithium plating degree data, the lithium battery capacity decay is predicted to obtain capacity drop inflection point prediction data. The health assessment module is used to construct a lithium battery health state transition equation based on calendar-microcirculation aging distribution weights; and to assess the health state of the lithium battery based on the lithium battery health state transition equation to obtain key health indicators of the lithium battery. The life prediction module is used to assess the remaining life of energy storage lithium battery systems based on key health indicators and capacity drop inflection point prediction data of lithium batteries, and obtain corrected life prediction data of lithium batteries.

Images