Lithium battery health comprehensive evaluation method and system, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a lithium battery health comprehensive evaluation method and system, electronic equipment and a storage medium, and belongs to the technical field of lithium batteries. According to the scheme, battery characteristic data of a lithium battery are obtained, and the battery characteristic data comprise effective capacity, discharge depth, temperature variation, charging time, current variation and voltage variation; feature extraction is carried out on the battery feature data to obtain a candidate feature set, and the candidate feature set comprises a plurality of candidate features; performing correlation analysis on the candidate feature set to obtain health features; on the basis of the health features, a comprehensive evaluation model is used for battery health state estimation and remaining service life prediction, a comprehensive evaluation result is obtained, the comprehensive evaluation model is obtained by training a back propagation neural network through a particle swarm optimization, and the scheme can improve the accuracy and robustness of the lithium battery health evaluation result.

Description

Technical Field

[0001] This application relates to the technical field of lithium batteries, and in particular to a comprehensive evaluation method, system, electronic device and storage medium for the health of lithium batteries. Background Art

[0002] Lithium-ion batteries are widely used in fields such as electric vehicles, energy storage systems, and portable electronic devices due to their advantages such as high energy density, long cycle life, and low self-discharge rate. However, as the battery is used, its performance will gradually degrade, manifested as phenomena such as capacity attenuation and internal resistance increase, ultimately leading to battery failure. The State of Health (SOH) of the battery is a key indicator to measure the degree of battery performance degradation, directly affecting the safety and economy of the battery. Therefore, accurately estimating the SOH of lithium-ion batteries is of great significance for the optimization of battery management systems, the extension of battery life, and the prevention of safety risks.

[0003] Currently, most of the related technologies estimate the battery based on a single dimension, which is difficult to comprehensively reflect the complex mechanism of battery aging, resulting in insufficient accuracy and reliability of the estimation results. Summary of the Invention

[0004] The main purpose of the embodiments of this application is to propose a comprehensive evaluation method, system, electronic device and storage medium for the health of lithium batteries, aiming to improve the accuracy and robustness of the lithium battery health evaluation results.

[0005] To achieve the above object, on the one hand, an embodiment of this application proposes a comprehensive evaluation method for the health of lithium batteries, and the method includes:

[0006] Obtain battery characteristic data of the lithium battery, where the battery characteristic data includes effective capacity, depth of discharge, temperature change amount, charging time, current change amount, and voltage change amount;

[0007] Extract features from the battery characteristic data to obtain a candidate feature set, where the candidate feature set includes multiple candidate features;

[0008] Perform correlation analysis on the candidate feature set to obtain health features;

[0009] Based on the health features, use a comprehensive evaluation model to estimate the battery health state and predict the remaining service life, and obtain a comprehensive evaluation result, where the comprehensive evaluation model is obtained by training a backpropagation neural network through a particle swarm algorithm.

[0010] In some embodiments, the extracting features from the battery characteristic data to obtain a candidate feature set includes the following steps:

[0011] Determine the constant current charging duration and the constant voltage charging duration according to the charging time, and perform a ratio process on the constant current charging duration and the constant voltage charging duration to obtain a charging duration ratio;

[0012] Perform an averaging process on the temperature change amount to obtain an average constant current charging temperature and an average constant voltage charging temperature;

[0013] Perform an integration process on the temperature change amount to obtain an integral of the constant current charging temperature curve and an integral of the constant voltage charging temperature curve;

[0014] Obtain a candidate feature set according to the constant current charging duration, the constant voltage charging duration, the charging duration ratio, the average constant current charging temperature, the average constant voltage charging temperature, the integral of the constant current charging temperature curve, and the integral of the constant voltage charging temperature curve.

[0015] In some embodiments, before the step of extracting features from the battery characteristic data to obtain a candidate feature set, the method further includes the following steps:

[0016] Determine abnormal data in the battery characteristic data through a quartile box plot algorithm;

[0017] Adopt an interpolation method to perform data correction processing on the abnormal data to obtain corrected battery characteristic data.

[0018] In some embodiments, the determining abnormal data in the battery characteristic data through the quartile box plot algorithm includes the following steps:

[0019] Determine the first quartile, the second quartile, and the third quartile according to the battery characteristic data;

[0020] Determine the interquartile range according to the first quartile and the third quartile;

[0021] Determine the lower limit value of the abnormal boundary according to the first quartile and the interquartile range;

[0022] Determine the upper limit value of the abnormal boundary according to the third quartile and the interquartile range;

[0023] Determine abnormal data according to the lower limit value and the upper limit value.

[0024] In some embodiments, the performing a correlation analysis on the candidate feature set to obtain a health feature includes the following steps:

[0025] Perform a correlation analysis on each of the candidate features to obtain the absolute value of the correlation coefficient;

[0026] Determine the candidate features with the absolute value of the correlation coefficient greater than a preset correlation threshold as health features.

[0027] In some embodiments, before the step of estimating the battery health state and predicting the remaining useful life by using a comprehensive evaluation model based on the health characteristics to obtain a comprehensive evaluation result, the method further includes the following steps:

[0028] Normalize the health characteristics by using the deviation normalization method to obtain the normalized health characteristics.

[0029] In some embodiments, the method further includes the following steps:

[0030] Perform a difference operation on the current change amounts at different stages to obtain a current difference;

[0031] Perform a difference operation on the voltage change amounts at different stages to obtain a voltage difference;

[0032] Determine the current difference and the voltage difference as supplementary features;

[0033] The step of estimating the battery health state and predicting the remaining useful life by using a comprehensive evaluation model based on the health characteristics to obtain a comprehensive evaluation result includes the following steps:

[0034] Input the health characteristics and the supplementary features into the comprehensive evaluation model for capacity prediction to obtain the battery health state;

[0035] Input the battery health state into the comprehensive evaluation model for mapping processing to obtain the remaining useful life;

[0036] Perform uncertainty quantification processing on the battery health state and the remaining useful life to obtain a comprehensive evaluation result.

[0037] To achieve the above object, on the other hand, an embodiment of the present application proposes a comprehensive evaluation system for lithium battery health, and the system includes:

[0038] A first module, configured to obtain battery characteristic data of a lithium battery, where the battery characteristic data includes effective capacity, depth of discharge, temperature change amount, charging time, current change amount, and voltage change amount;

[0039] A second module, configured to perform feature extraction on the battery characteristic data to obtain a candidate feature set, where the candidate feature set includes a plurality of candidate features;

[0040] A third module, configured to perform correlation analysis on the candidate feature set to obtain health characteristics;

[0041] The fourth module is used to estimate the battery health state and predict the remaining useful life based on the health features by using a comprehensive evaluation model to obtain a comprehensive evaluation result, wherein the comprehensive evaluation model is obtained by training a backpropagation neural network through a particle swarm algorithm.

[0042] To achieve the above object, on the other hand, an embodiment of the present application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above method is implemented.

[0043] To achieve the above object, on the other hand, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above method is implemented.

[0044] The embodiments of the present application at least include the following beneficial effects: The present application provides a comprehensive evaluation method, system, electronic device and storage medium for lithium battery health. The solution obtains battery characteristic data of the lithium battery, wherein the battery characteristic data includes effective capacity, discharge depth, temperature change amount, charging time, current change amount and voltage change amount; extracts features from the battery characteristic data to obtain a candidate feature set, wherein the candidate feature set includes multiple candidate features; performs correlation analysis on the candidate feature set to obtain health features; based on the health features, uses a comprehensive evaluation model to estimate the battery health state and predict the remaining useful life to obtain a comprehensive evaluation result, wherein the comprehensive evaluation model is obtained by training a backpropagation neural network through a particle swarm algorithm. This solution can improve the accuracy and robustness of the lithium battery health evaluation result. Description of the Drawings

[0045] Figure 1 is a flowchart of the comprehensive evaluation method for lithium battery health provided by the embodiment of the present application;

[0046] Figure 2 is a flowchart of the PSO-BP neural network training provided by the embodiment of the present application;

[0047] Figure 3 is a schematic diagram of the quartile box plot provided by the embodiment of the present application;

[0048] Figure 4 is a flowchart of the comprehensive evaluation method for lithium battery health provided by another embodiment of the present application;

[0049] Figure 5 is a schematic diagram of the structure of the comprehensive evaluation system for lithium battery health provided by the embodiment of the present application;

[0050] Figure 6It is a schematic diagram of the hardware structure of the electronic device provided by an embodiment of the present application. Detailed implementation manners

[0051] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application. When the following description refers to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the embodiments of the present application. They are only examples of devices and methods consistent with some aspects of the embodiments of the present application detailed in the appended claims.

[0052] It can be understood that the terms "first", "second", etc. used in the present application may be used in this document to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if", "when" used herein may be interpreted as "when...", "when...", or "in response to determining".

[0053] The terms "at least one", "a plurality of", "each", "any one", etc. used in the present application, at least one includes one, two or more than two, a plurality includes two or more than two, each refers to each of the corresponding plurality, and any one refers to any one of the plurality.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application, and are not intended to limit the present application.

[0055] The comprehensive evaluation method for lithium battery health provided by the embodiments of the present application relates to the technical field of lithium batteries. The comprehensive evaluation method for lithium battery health provided by the embodiments of the present application can be applied to a terminal, a server, or software running on a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, a vehicle-mounted terminal, etc., but is not limited thereto; the server side can be configured as an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application implementing the comprehensive evaluation method for lithium battery health, etc., but is not limited to the above forms.

[0056] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet-type devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0057] Figure 1 is an optional flowchart of the comprehensive evaluation method for lithium battery health provided by the embodiments of the present application, Figure 1 The method in may include but is not limited to steps S101 to S104.

[0058] Step S101, obtain battery characteristic data of the lithium battery, where the battery characteristic data includes effective capacity, depth of discharge, temperature change amount, charging time, current change amount, and voltage change amount.

[0059] Step S102, perform feature extraction on the battery characteristic data to obtain a candidate feature set, where the candidate feature set includes multiple candidate features.

[0060] Step S103, perform correlation analysis on the candidate feature set to obtain health features.

[0061] Step S104: Based on the health characteristics, use the comprehensive evaluation model to estimate the battery health state and predict the remaining useful life, and obtain the comprehensive evaluation result. Among them, the comprehensive evaluation model is obtained by training the backpropagation neural network through the particle swarm optimization algorithm.

[0062] In this embodiment, to break through the limitations of traditional single estimation methods and the lack of interpretability of prediction models based on machine learning, if we want to perform lithium battery health modeling and health prediction, we first need to extract battery characteristic data with electrochemical and physical characteristics of the lithium battery. Among them, the battery characteristic data includes the effective capacity, depth of discharge, temperature change, charging time, current change, and voltage change.

[0063] The effective capacity refers to the amount of electricity that the battery can actually store and release in the current state. It is usually measured through a discharge experiment. The specific method is to discharge the battery until the state of charge (SOC) is 0%, and integrate the discharge current over the discharge time to obtain the effective capacity Q. act Divide the effective capacity Q act by the scalar capacity Q nom to obtain the SOH value. The calculation formula is as follows:

[0064]

[0065] In the charge / discharge cycle, it can be found that as the number of charge / discharge cycles increases, the value of Q act continually decreases, resulting in a continuous decrease in the SOH value as the battery is used. To reflect the discharge rate of the battery in actual use, a parameter depth of discharge (DoD) is introduced to characterize the discharge rate. Then, formula (1) can be expressed as:

[0066]

[0067] When the number of battery usage times reaches more than 28 times, the accuracy can reach 1%. As the battery ages, the constant current charging time shortens, and the current change rate in the constant voltage charging stage gradually decreases, resulting in an increase in the constant voltage stage charging time. Therefore, the charging time can also provide key information about the battery health state.

[0068] Temperature is an important indicator of battery degradation. The battery generates heat during operation, and the health state of the battery can be indirectly evaluated by monitoring the temperature change of the battery.

[0069] The electro-chemical reaction rate and internal resistance can also reflect battery degradation. For example, the charging current difference may reflect the change in battery internal resistance, and the charging voltage difference may reflect the polarization reaction. Therefore, the current change amount and voltage change amount at different stages during the charge and discharge process can be dynamically extracted.

[0070] Furthermore, all the features that can reflect the battery health state are extracted from the battery characteristic data to obtain a candidate feature set. Correlation analysis is performed on the candidate feature set to calculate the correlation between each candidate feature and the SOH, and the features highly correlated with the battery health are screened out to obtain the health features. It can be understood that the health features will be dynamically adjusted according to the correlation analysis results of the actual data, for example, they may vary due to different battery types, operating conditions, or environmental conditions.

[0071] The health features verified to be highly correlated with the battery health through correlation analysis are used as key features and input into the comprehensive evaluation model for battery health state estimation and remaining useful life prediction. The state of health (SOH) of the battery, as an important indicator to measure the degree of battery degradation, is usually defined using the capacity ratio, that is, the ratio of the current cycle capacity to the initial capacity. Since the SOH of lithium batteries is closely related to their performance, and the remaining useful life (RUL) of the battery can more intuitively reflect the degree of battery performance degradation, there is a certain mapping relationship between the SOH and the RUL. Therefore, the SOH can be used as the basis for RUL prediction.

[0072] It should be noted that the comprehensive evaluation model can use the particle swarm optimization (PSO) algorithm to change the weights and thresholds of the backpropagation (BP) neural network to achieve the purpose of optimization. Figure 2 Fig. is the flowchart of the PSO-BP neural network training. The optimization process is to first determine the network structure according to the data, then use the PSO algorithm to optimize the parameter values of the BP neural network structure to obtain the optimal weights and thresholds, and finally assign the obtained results to the network and use the BP neural network for training and prediction.

[0073] The BP neural network generally adopts a three-layer structure of an input layer, a hidden layer, and an output layer. Therefore, when estimating the SOH, the health features are used as the input of the model, and when predicting the RUL, the health features and the estimated SOH value are used as the input of the model. Let Xm be the input variable of the input layer in the model, and Yn be the output value of the output layer in the model. W ij and W jk are the connection weights from the input layer to the hidden layer and from the hidden layer to the output layer respectively. There is the following corresponding relationship between the unit outputs of each layer:

[0074]

[0075] In Equation (3), σ is the activation function of the hidden layer, and the logsig function or tansig function is generally selected. γ is the activation function of the output layer, and the linear purelin function is generally selected.

[0076] When designing the PSO algorithm, for any particle i, its position vector is denoted as e i = e i1 , e i2 , …, e in , and the velocity vector is v i = v i1 , v i2 , …, v in , the optimal solution vector experienced by a single particle is l i = l i1 , l i2 , …, l in , the optimal solution vector experienced by all particles in the population is lq = lq1, lq2, …, lqn, and the iterative relationship is:

[0077]

[0078] In Equation (4), ω represents the inertia weight coefficient, u = 1, 2, …, n, n is the spatial dimension, h is the number of iterations, i = 1, 2, …, s, s is the sample number of the population, p1 and p2 are random numbers between 0 and 1, k1 and k2 are constants, v iu ∈ -vmax, vmax, e iu ∈ -emax, emax is usually determined by the actual situation, and generally vmax = kemax.

[0079] In some embodiments, step S102 may include but is not limited to steps S201 to S204.

[0080] Step S201: Determine the constant current charging duration and the constant voltage charging duration according to the charging time, and perform a ratio process on the constant current charging duration and the constant voltage charging duration to obtain a charging duration ratio.

[0081] Step S202: Perform an averaging process according to the temperature change amount to obtain the average constant current charging temperature and the average constant voltage charging temperature.

[0082] Step S203: Perform an integration process according to the temperature change amount to obtain the integral of the constant current charging temperature curve and the integral of the constant voltage charging temperature curve.

[0083] Step S204: Obtain a candidate feature set according to the constant current charging duration, the constant voltage charging duration, the charging duration ratio, the average constant current charging temperature, the average constant voltage charging temperature, the integral of the constant current charging temperature curve, and the integral of the constant voltage charging temperature curve.

[0084] In this embodiment, the charging process of the lithium battery is generally divided into two stages: the constant current charging stage and the constant voltage charging stage.

[0085] In the constant current charging stage, the charging current remains constant, and the battery voltage gradually rises until the set charging cut-off voltage is reached; by analyzing the curve of the charging time, the start time and end time of the constant current charging stage can be determined, and the constant current charging duration (TCC) can be obtained.

[0086] In the constant voltage charging stage, the charging voltage remains constant, and the charging current gradually decreases until the current drops to the set cut-off current, and the charging process ends; by analyzing the curve of the charging time, the start time and end time of the constant voltage charging stage can be determined, and the constant voltage charging duration (TCV) can be obtained.

[0087] The charging duration ratio TCC / TCV is obtained by calculating the ratio of TCC to TCV.

[0088] Temperature monitoring in different charging stages is very important for both battery thermal management and SOH estimation. By averaging the temperature change, the average constant current charging temperature (ACC) and the average constant voltage charging temperature (ACV) are extracted, which can be used as characteristics reflecting the thermal behavior of the battery.

[0089] At the same time, by integrating the temperature change, the integral of the constant current charging temperature curve (FCC) and the integral of the constant voltage charging temperature curve (FCV) are extracted, which can be used as characteristics reflecting the thermal accumulation effect of the battery.

[0090] A candidate feature set is formed based on the features extracted from the battery characteristic data that can reflect the health state of the battery.

[0091] In some embodiments, before step S102, the comprehensive assessment of the lithium battery health may further include, but is not limited to, steps S301 to S302.

[0092] Step S301, determining the abnormal data in the battery characteristic data through the quartile box plot algorithm.

[0093] Step S302, performing data correction processing on the abnormal data using the interpolation method to obtain the corrected battery characteristic data.

[0094] In this embodiment, considering that equipment noise during the measurement process may cause data anomalies, it is necessary to process the data, and the quartile box plot (Boxplot) algorithm is used to screen for outliers.

[0095] After identifying abnormal data in the battery characteristic data, interpolation method is used to correct the abnormal data, covering the abnormal data, realizing the correction of the abnormal data, and improving the data quality. The calculation formula is as follows:

[0096]

[0097] In formula (5), damend is the corrected value of abnormal data; dt-1 and dt+1 are the battery characteristic data at the previous moment and the next moment of the t moment respectively.

[0098] In some embodiments, step S301 may include but is not limited to steps S401 to S405.

[0099] Step S401, determine the first quartile, the second quartile and the third quartile according to the battery characteristic data.

[0100] Step S402, determine the interquartile range according to the first quartile and the third quartile.

[0101] Step S403, determine the lower limit value of the abnormal boundary according to the first quartile and the interquartile range.

[0102] Step S404, determine the upper limit value of the abnormal boundary according to the third quartile and the interquartile range.

[0103] Step S405, determine the abnormal data according to the lower limit value and the upper limit value.

[0104] In this embodiment, referring to Figure 3 , the box plot is a visualization tool based on data distribution. The quartile is a kind of quantile in statistics, that is, all numerical values are arranged from small to large and divided into four equal parts, and the numerical values at the three segmentation points are the quartiles.

[0105] Exemplarily, the boundary of abnormal data is defined by the quartile and the interquartile range (IQR). Among them, the quartile includes the first quartile Q1, the second quartile Q2 and the third quartile Q3. The first quartile Q1 is also called the "lower quartile", which is equal to the 25% number after all numerical values in the battery characteristic data are arranged from small to large. The second quartile Q2 is also called the "median", which is equal to the 50% number after all numerical values in the sample are arranged from small to large. The third quartile Q3 is also called the "upper quartile", which is equal to the 75% number after all numerical values in the sample are arranged from small to large.

[0106] The difference between the third quartile Q1 and the first quartile Q3 is the interquartile range. If the data less than Q1 - 1.5IQR and greater than Q3 + 1.5IQR are set as abnormal data, the abnormal boundary is as follows:

[0107] [d lower ,d upper = [Q1 - 1.5I QR , Q3 + 1.5I QR (6);

[0108] In Equation (6), d lower is the lower limit value of the abnormal boundary, and d upper is the upper limit value of the abnormal boundary.

[0109] In some embodiments, step S103 may include but is not limited to steps S501 to S502.

[0110] Step S501: Perform a correlation analysis on each candidate feature to obtain the absolute value of the correlation coefficient.

[0111] Step S502: Determine the candidate features with the absolute value of the correlation coefficient greater than the preset correlation threshold as healthy features.

[0112] In this embodiment, the Spearman correlation coefficient is used to calculate the correlation between each feature and the SOH, and the absolute value of the correlation coefficient of each candidate feature is obtained. The calculation formula is as follows:

[0113]

[0114] In Equation (7), Spearmax represents the correlation coefficient. A correlation coefficient greater than 0 indicates that the candidate feature is positively correlated with the capacity, and a correlation coefficient less than 0 indicates that the candidate feature is negatively correlated with the capacity. The larger the absolute value of the correlation coefficient, the higher the correlation between the two.

[0115] Optionally, the preset correlation threshold is 0.85. The candidate features with the absolute value of the correlation coefficient greater than 0.85 are selected as healthy features and used as input parameters for subsequent SOH prediction.

[0116] In some embodiments, before step S104, the comprehensive evaluation method for the health of lithium batteries may further include but is not limited to step S601.

[0117] Step S601: Normalize the healthy features using the deviation normalization method to obtain the normalized healthy features.

[0118] In this embodiment, since the dimensions of the battery's various health features are different, in order to improve the convergence speed of the comprehensive evaluation model, the deviation (min-max) normalization method can be used to normalize each health feature. The normalization formula is as follows:

[0119]

[0120] In Equation (8), x* is the normalized health feature, and x min and x max are the minimum and maximum values of the health feature, respectively.

[0121] In some embodiments, the comprehensive evaluation method for the health of a lithium battery may further include, but is not limited to, steps S701 to S703.

[0122] Step S701: Perform a difference operation on the current change amounts at different stages to obtain a current difference.

[0123] Step S702: Perform a difference operation on the voltage change amounts at different stages to obtain a voltage difference.

[0124] Step S703: Determine the current difference and the voltage difference as supplementary features.

[0125] In this embodiment, considering that the current difference and the voltage difference during the charge / discharge process do not pass the correlation analysis, but are related to capacity degradation based on domain knowledge or experimental verification, they can still be used as supplementary features to enhance the robustness of the model. The difference refers to the difference in certain statistics between stages during the same charging process (or discharging process), such as the average current / voltage, or the integral value of the change within a stage.

[0126] Exemplarily, taking the charging process as an example, the current in the constant-current charging stage is constant, while the current in the constant-voltage charging stage gradually decreases. Since their change trends are different, the difference in the current change amounts between the two stages can be calculated as the charging current difference.

[0127] Specifically, the current in the constant-current charging stage remains unchanged, and the change amount is zero. The change amount of the current in the constant-voltage charging stage is the difference from the initial value to the cut-off value. Therefore, the difference is the difference in the current change amounts between the two stages. The difference in the current is calculated based on the current change amounts in the constant-current charging stage and the constant-voltage charging stage to obtain the charging current difference.

[0128] Similarly, the voltage in the constant-current charging stage gradually increases, while the voltage in the constant-voltage charging stage is constant. The difference in the voltage is calculated based on the voltage change amounts in the constant-current charging stage and the constant-voltage charging stage to obtain the charging voltage difference.

[0129] The calculation methods for the current difference and the voltage difference during the discharging process are similar to those during the charging process. By performing a difference operation on the current and voltage change amounts at different stages during the discharging process, the discharging current difference and the discharging voltage difference are calculated respectively.

[0130] Taking the current difference and the voltage difference as supplementary features can be input into the model together with the health features to enhance the accuracy of the estimation of the battery health state and the prediction of the remaining useful life.

[0131] Optionally, the charging current difference and charging voltage difference calculated during the charging process can be used as supplementary features, or the discharging current difference and discharging voltage difference calculated during the discharging process can be used as supplementary features.

[0132] Furthermore, if it is necessary to improve the prediction accuracy of the model, it can also be considered to use the charging current difference, charging voltage difference, discharging current difference, and discharging voltage difference calculated during the charging process and discharging process as supplementary features.

[0133] It should be noted that the supplementary features can be not only the current difference and voltage difference between different stages during the charging / discharging process, but also other features of the stages, as long as they can be related to capacity degradation based on domain knowledge or experimental verification. The embodiments of the present application do not make specific limitations.

[0134] In some embodiments, step S104 may include but is not limited to steps S801 to S803.

[0135] Step S801: Input the health features and supplementary features into the comprehensive evaluation model for capacity prediction to obtain the battery health state.

[0136] Step S802: Input the battery health state into the comprehensive evaluation model for mapping processing to obtain the remaining useful life.

[0137] Step S803: Perform uncertainty quantification processing on the battery health state and the remaining useful life to obtain the comprehensive evaluation result.

[0138] In this embodiment, through the PSO-BP neural network algorithm, combined with the health features (TCC, TCV, TCC / TCV, FCC, FCV, etc.) screened by the aforementioned correlation analysis and supplementary features such as charging current difference and charging voltage difference, SOH estimation and RUL prediction are performed.

[0139] Specifically, to solve the problem of difficult online measurement of capacity, the SOH estimation in this embodiment estimates the capacity according to the input features through the comprehensive evaluation model to replace the actual capacity measurement, and determines the SOH value according to the estimated initial capacity and the estimated current cycle capacity.

[0140] Exemplarily, based on the depth of discharge DOD during the charging and discharging process, the capacity is dynamically estimated through the comprehensive evaluation model in combination with the health features and supplementary features. The SOH value in the j-th cycle can be denoted as:

[0141]

[0142] In Equation (9), sohj represents the SOH value in the j-th cycle, is the initial capacity estimated for the battery, The current cycle capacity estimated for the battery in the j-th cycle.

[0143] After obtaining the estimated SOH value, based on the estimated SOH value, combined with the health characteristics and supplementary characteristics, the remaining useful life is predicted through a comprehensive evaluation model. The prediction of the remaining useful life is based on the mapping relationship between SOH and battery performance degradation. When the capacity loss reaches a certain threshold, the battery is regarded as failed.

[0144] Exemplarily, when the capacity loss of a lithium battery reaches 30% of its rated capacity, the performance of the lithium battery is considered to have failed. Taking the No. 5 battery as an example, the end of life of the No. 5 battery can be predicted at the 124th cycle according to the SOH value at the time of failure.

[0145] Finally, through techniques such as Bayesian method and Monte Carlo simulation, the uncertainty of the comprehensive evaluation result is quantified, which can provide more comprehensive information support for battery management.

[0146] Next, combined with specific application examples, the solutions of the embodiments of the present invention will be introduced and described in detail.

[0147] The state of health (SOH) trajectory of a lithium-ion battery exhibits significant non-linearity and volatility. For the early warning of battery thermal runaway, its critical conditions are mainly monitored. By monitoring characteristic parameters, battery thermal runaway can be effectively warned, thus avoiding economic losses.

[0148] Currently, the estimation methods for battery health can be divided into estimation methods based on electrochemical characteristics, estimation methods based on physical characteristics, and estimation methods based on data-driven.

[0149] I. Estimation methods based on electrochemical characteristics:

[0150] Incremental capacity / differential voltage (IC / DV) analysis: This is a non-destructive characterization method that has been widely used in the analysis of aging mechanisms. By differentiating, the voltage plateau in the charge and discharge curves is converted into the characteristic peaks of the IC curve and the characteristic valleys of the DV curve. Analyzing the changes of these characteristics with battery aging can clarify the degradation mechanism. IC / DV analysis is also an effective means for online SOH estimation. It only needs to obtain two parameters, voltage and charge / discharge capacity, and is applicable to different types of lithium-ion batteries.

[0151] Differential thermal voltammetry (DTV): Combining the method of IC analysis with temperature measurement to obtain thermodynamic information about the electrode material. The characteristic peak parameters (such as position, peak intensity, peak width) of the DTV curve can be used to study battery aging, such as capacity degradation, resistance increase, and non-uniform electrode performance. DTV analysis is easy to perform on a parallel battery pack in terms of pre-experiment operations, and the test current rate can be relatively high, but it is easily affected by the ambient temperature.

[0152] Electrochemical Impedance Spectroscopy (EIS): By measuring the impedance of the battery at different frequencies, changes in parameters such as the internal resistance and capacitance of the battery can be obtained, and then the health state of the battery can be evaluated. EIS analysis has an in-depth understanding of the aging mechanism and performance degradation of the battery, but requires specialized equipment and complex analysis processes.

[0153] II. Estimation methods based on physical characteristics:

[0154] Differential Mechanical Parameter Method (DMP): By modeling the mechanical behavior of the battery, such as changes in the strain and stress of a single cell, SOH estimation is achieved. The insertion / extraction of lithium ions in the electrode active material is related to its volume change, expansion and contraction, and the changes in these mechanical parameters are linearly correlated with the SOH. However, the DMP method requires additional equipment to measure mechanical parameters and is difficult to implement in actual battery packs.

[0155] Temperature monitoring: The battery generates heat during operation. By monitoring the temperature change of the battery, the health state of the battery can be indirectly evaluated. High temperature may lead to a decline in battery performance and a shortening of its life, so temperature monitoring is very important for both battery thermal management and SOH estimation.

[0156] III. Estimation methods based on data-driven:

[0157] Machine learning: In recent years, significant progress has been made in the prediction of the health state of lithium-ion batteries based on machine learning methods. Machine learning algorithms can process a large amount of battery operation data (such as voltage, current, temperature, etc.), learn the complex relationship between the battery state and performance, and build a prediction model. Common machine learning models include Artificial Neural Network (ANN), Support Vector Machine (SVM), Gaussian Process Regression (GPR), etc.

[0158] Data preprocessing and feature extraction: Before machine learning, the original data needs to be preprocessed and features extracted. Feature extraction is a key step, which directly affects the effect of SOH estimation. Common features include voltage gradient curve (dV / dt), IC / DV curve, etc.

[0159] This embodiment combines the above methods to form a comprehensive evaluation method for analyzing the health state of lithium batteries. Considering the complexity and non-linear characteristics of the health condition of lithium-ion batteries, by combining electrochemical characteristic analysis, physical characteristic monitoring and data-driven methods, a comprehensive and accurate evaluation of the battery health state is achieved.

[0160] Specifically, by adopting a method that combines electrochemistry and physical characteristics, the electrochemical parameters obtained from IC / DV analysis and EIS analysis are combined with physical parameters such as the DMP method and temperature monitoring. Utilizing the complementarity of these parameters, a multi-dimensional and multi-level health state estimation system is constructed, which can more comprehensively reflect the changes in the internal state of the battery and obtain an equivalent physical model of the battery.

[0161] Then, the physical model is fused with data-driven methods by combining the machine learning model with the physical model, improving the interpretability and reliability of the model while maintaining the prediction accuracy.

[0162] Furthermore, the machine learning model can utilize deep learning models such as particle swarm optimization (PSO) and algorithm optimization (BP) to improve the accuracy of SOH estimation.

[0163] There are various sources of uncertainty in the battery SOH estimation process, such as measurement noise and model error. Through techniques such as Bayesian methods and Monte Carlo simulations, the uncertainty of the estimation results is quantified.

[0164] Exemplarily, referring to Figure 4 , Figure 4 is a comprehensive evaluation method for the health of lithium batteries provided by another embodiment of the present application. First, data such as battery capacity, current change amount, and voltage change amount are extracted, and health features (such as TCC, TCV, TCCTCV, FCC, FCV, etc.) and supplementary features related to degradation based on domain knowledge or experimental verification (such as charging current difference, charging voltage difference, etc.) are screened from the candidate features through correlation analysis as input features of the comprehensive evaluation model to characterize the battery degradation state.

[0165] Then, by combining the battery's historical charge and discharge data (such as discharge depth DOD, temperature curve, etc.) and the capacity estimation value (the capacity predicted by a linear regression model or a PSO-BP neural network), with SOH (calculated based on the capacity ratio) or RUL (predicted based on the capacity loss threshold) as the target variable, the neural network model learns the mapping relationship between capacity degradation and features, and prediction models for SOH estimation and RUL prediction are obtained respectively.

[0166] During the model training process, the mean absolute percentage error (MAPE), root mean square error (RMSE), mean absolute error (MAE), and absolute error (AE) can be used as evaluation criteria to analyze the performance of the model.

[0167] Exemplarily, the mean absolute percentage error (MAPE) and root mean square error (RMSE) are used as evaluation criteria for SOH estimation, and the mean absolute error (MAE) and absolute error (AE) are used as evaluation criteria for RUL prediction for performance analysis. The formulas are as follows:

[0168]

[0169] In Formulas (10) to (13), j = k + 1, k + 2, …, n, where n is the total number of cycles, k is the number of cycles for training, sohj and are the actual value and the estimated value of SOH in the j-th cycle respectively, rulj and are the actual value and the predicted value of RUL in the j-th cycle respectively. represents the predicted remaining service life cycle number, and rul represents the actual remaining service life cycle number.

[0170] Please refer to Figure 5 , the embodiment of the present application also provides a comprehensive evaluation system for lithium battery health, which can implement the above-mentioned comprehensive evaluation method for lithium battery health. The system includes:

[0171] The first module is used to obtain battery characteristic data of the lithium battery, where the battery characteristic data includes effective capacity, depth of discharge, temperature change, charging time, current change, and voltage change.

[0172] The second module is used to extract features from the battery characteristic data to obtain a candidate feature set, where the candidate feature set includes multiple candidate features.

[0173] The third module is used to perform correlation analysis on the candidate feature set to obtain health features.

[0174] The fourth module is used to estimate the battery health state and predict the remaining service life based on the health features by using a comprehensive evaluation model to obtain a comprehensive evaluation result, where the comprehensive evaluation model is obtained by training a backpropagation neural network through a particle swarm algorithm.

[0175] It can be understood that the content in the above method embodiments is applicable to the system embodiments of the present application. The functions specifically implemented by the system embodiments of the present application are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0176] The embodiment of the present application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the above-mentioned comprehensive evaluation method for lithium battery health. The electronic device can be any intelligent terminal including a tablet computer, an in-vehicle computer, etc.

[0177] It can be understood that the content in the above method embodiments is applicable to the device embodiments of the present application. The functions specifically implemented by the device embodiments of the present application are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those of the above method embodiments.

[0178] Please refer to Figure 6 , Figure 6 which schematically shows the hardware structure of an electronic device according to another embodiment. The electronic device includes:

[0179] A processor 901, which can be implemented in the form of a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.

[0180] A memory 902, which can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 902 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 902 and are called by the processor 901 to execute the comprehensive evaluation method for the health of lithium batteries in the embodiments of the present application.

[0181] An input / output interface 903, which is used to implement information input and output.

[0182] A communication interface 904, which is used to implement communication and interaction between this device and other devices, and can implement communication through a wired method (such as USB, network cable, etc.) or through a wireless method (such as mobile network, WIFI, Bluetooth, etc.).

[0183] A bus 905, which transmits information between various components of the device (such as the processor 901, the memory 902, the input / output interface 903, and the communication interface 904).

[0184] Among them, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are communicatively connected to each other inside the device through the bus 905.

[0185] The embodiments of the present application also provide a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements the above-mentioned comprehensive evaluation method for the health of lithium batteries.

[0186] It can be understood that the content in the above method embodiments is applicable to the embodiments of this storage medium. The functions specifically implemented by the embodiments of this storage medium are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those in the above method embodiments.

[0187] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0188] The comprehensive evaluation method for lithium battery health, the comprehensive evaluation system for lithium battery health, the electronic device, and the storage medium provided by the embodiments of the present application can capture multi-dimensional information of battery aging more comprehensively by combining multiple methods of electrochemical characteristics, physical characteristics, and data-driven, thereby improving the accuracy and reliability of SOH estimation. This multi-source information fusion method can better adapt to the complex degradation mechanism of the battery than a single method, capture various characteristics of battery aging, and improve the accuracy and reliability of estimation.

[0189] Different estimation methods may exhibit different performances under different conditions. For example, Electrochemical Impedance Spectroscopy (EIS) has high precision under laboratory conditions, but may be greatly interfered with in an actual complex environment; while data-driven methods can automatically learn and adapt to environmental changes from a large amount of operation data. The comprehensive method of the embodiments of the present application can effectively make up for the deficiencies of a single method and enhance the overall robustness of the system.

[0190] When applying the data-driven method, the embodiments of the present application take into account the importance of data preprocessing and feature extraction, and through the adoption of advanced algorithms and technical means, effectively preprocess and extract features from the original battery operation data, and extract key information that can reflect the battery health state, providing strong data support for the construction of subsequent machine learning models.

[0191] The embodiments of the present application can not only improve the accuracy of estimating the health state of lithium-ion batteries, but also provide strong support for the intelligent upgrade of battery management systems and improve the intelligent level. Through the comprehensive evaluation method provided by the embodiments of the present application, the battery management system can more accurately evaluate the health state of the battery, optimize charge and discharge strategies, thermal management schemes, etc., thereby extending the service life of the battery and improving energy utilization efficiency.

[0192] In addition, by monitoring various parameters of the battery in real time, including voltage, current, temperature, etc., and combining with the prediction ability of the machine learning model, the embodiments of the present application can also achieve early warning of the battery health state. This helps to timely detect the trend of battery performance degradation and take measures in advance to prevent safety accidents such as battery failure or thermal runaway.

[0193] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0194] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or combine certain steps, or different steps.

[0195] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0196] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices can be implemented as software, firmware, hardware and their appropriate combinations.

[0197] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0198] The preferred embodiments of the embodiments of the present application have been described above with reference to the drawings, and thus do not limit the scope of rights of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of rights of the embodiments of the present application.

Claims

1. A comprehensive evaluation method for lithium battery health, characterized in that: The method comprises the following steps: Acquire battery characteristic data of the lithium battery, wherein the battery characteristic data includes effective capacity, discharge depth, temperature change, charging time, current change and voltage change; Extracting features from the battery characteristic data to obtain a candidate feature set, wherein the candidate feature set includes a plurality of candidate features; Performing correlation analysis on the candidate feature set to obtain health features; Based on the health characteristics, a comprehensive evaluation model is used to estimate the battery health state and predict the remaining service life to obtain a comprehensive evaluation result, wherein the comprehensive evaluation model is obtained by training a back propagation neural network through a particle swarm algorithm.

2. The method according to claim 1, characterized in that The step of extracting features from the battery characteristic data to obtain a candidate feature set includes the following steps: Determining a constant current charging time and a constant voltage charging time according to the charging time, and performing ratio processing on the constant current charging time and the constant voltage charging time to obtain a charging time ratio; Performing average processing according to the temperature variation to obtain a constant current charging temperature average value and a constant voltage charging temperature average value; Performing integration processing according to the temperature change amount to obtain a constant current charging temperature curve integral and a constant voltage charging temperature curve integral; A candidate feature set is obtained according to the constant current charging time, the constant voltage charging time, the charging time ratio, the constant current charging temperature average value, the constant voltage charging temperature average value, the constant current charging temperature curve integral and the constant voltage charging temperature curve integral.

3. The method according to claim 1, characterized in that Before the step of extracting features from the battery characteristic data to obtain a candidate feature set, the method further includes the following steps: Determining abnormal data in the battery characteristic data by using a quartile box plot algorithm; The abnormal data is corrected by using an interpolation method to obtain corrected battery characteristic data.

4. The method according to claim 3, characterized in that The method of determining abnormal data in the battery characteristic data by using a quartile box plot algorithm comprises the following steps: Determine a first quartile, a second quartile, and a third quartile according to the battery characteristic data; Determine the interquartile range based on the first quartile and the third quartile; Determine a lower limit value of an abnormal boundary according to the first quartile and the interquartile range; Determine an upper limit value of an abnormal boundary according to the third quartile and the interquartile range; Abnormal data is determined based on the lower limit value and the upper limit value.

5. The method according to claim 1, characterized in that The step of performing correlation analysis on the candidate feature set to obtain health features comprises the following steps: Performing correlation analysis on each of the candidate features to obtain an absolute value of the correlation coefficient; The candidate feature whose absolute value of the correlation coefficient is greater than a preset correlation threshold is determined as a healthy feature.

6. The method according to claim 1, characterized in that Before the step of estimating the battery health state and predicting the remaining service life by using a comprehensive evaluation model based on the health characteristics to obtain a comprehensive evaluation result, the method further includes the following steps: The health characteristics are normalized by using a deviation normalization method to obtain normalized health characteristics.

7. The method according to claim 1, characterized in that The method further comprises the following steps: Performing difference processing on the current variation at different stages to obtain a current difference; Performing difference processing on the voltage variation at different stages to obtain a voltage difference; determining the current difference and the voltage difference as complementary features; The method of using a comprehensive evaluation model to estimate the battery health status and predict the remaining service life based on the health characteristics to obtain a comprehensive evaluation result includes the following steps: Inputting the health feature and the supplementary feature into a comprehensive evaluation model to perform capacity prediction to obtain a battery health state; Inputting the battery health status into a comprehensive evaluation model for mapping processing to obtain a remaining service life; Uncertainty quantification is performed on the battery health status and the remaining service life to obtain a comprehensive evaluation result.

8. A comprehensive evaluation system for lithium battery health, characterized in that: The system comprises: The first module is used to obtain battery characteristic data of the lithium battery, wherein the battery characteristic data includes effective capacity, discharge depth, temperature change, charging time, current change and voltage change; A second module is used to extract features from the battery characteristic data to obtain a candidate feature set, wherein the candidate feature set includes multiple candidate features; The third module is used to perform correlation analysis on the candidate feature set to obtain health features; The fourth module is used to estimate the battery health status and predict the remaining service life based on the health characteristics using a comprehensive evaluation model to obtain a comprehensive evaluation result, wherein the comprehensive evaluation model is obtained by training a back propagation neural network using a particle swarm algorithm.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images