A method and system for estimating battery health status supporting any charging interval

Abstract

The present invention provides a battery health state estimation method and system supporting any charging interval in the field of battery detection technology, the method comprising: step S10, obtaining a large amount of charging data of lithium batteries; step S20, preprocessing each charging data; step S30, obtaining the SOH true value corresponding to each charging data, and constructing an online charging data set based on the preprocessed charging data and the SOH true value; step S40, using the time series feature extraction tool tsfresh to extract features of the online charging data set to obtain 64-dimensional time series features; step S50, sorting and screening each time series feature by importance to obtain 12-dimensional time series features; step S60, creating a health state estimation model, using the 12-dimensional time series features to train the health state estimation model, and using the trained health state estimation model to perform online health state estimation of lithium batteries. The advantages of the present invention are: it greatly expands the scope of application of online health state estimation of lithium batteries.

Description

Technical Field

[0001] The present invention relates to the technical field of battery detection, and in particular to a method and system for estimating the health status of a battery supporting any charging interval. Background Art

[0002] The state of health (SOH) of a lithium battery is a key indicator, defined as the percentage of the current available capacity of a lithium battery to its initial capacity. During the charge and discharge cycle of a lithium battery, some irreversible chemical reactions will occur inside the battery, causing the aging of the lithium battery and inevitably leading to some safety hazards. Therefore, it is very important to estimate the health status of a lithium battery.

[0003] Since charging conditions are generally more controllable and stable than discharging conditions, online testing of lithium batteries is generally performed under charging conditions. When charging lithium batteries, especially new energy vehicles, there are situations where the charging SOC range, charging SOC starting point, and charging rate are not fixed.

[0004] Traditional lithium battery health status estimation methods manually extract relevant features from the collected voltage, current, temperature and other signals, such as constant current charging time, maximum value, minimum value, mean value, etc., or extract the peak and trough position amplitude of the capacity increment curve IC-curve to deduce the aging state of the lithium battery. However, the traditional method has the following disadvantages: when applied to online battery health status estimation, there will be a situation where the SOC interval cannot cover the feature calculation area, resulting in the inability to estimate the battery health status in some charging SOC intervals; for example, the second peak position of the capacity increment curve is generally located in the SOC interval of 40%-70%. If the SOC interval of a certain charge does not cover the SOC interval of 40%-70%, it may lead to the inability to perform online battery health status estimation.

[0005] Therefore, how to provide a battery health status estimation method and system that supports any charging interval to expand the scope of application of online health status estimation of lithium batteries has become a technical problem that needs to be solved urgently. Summary of the invention

[0006] The technical problem to be solved by the present invention is to provide a battery health status estimation method and system supporting any charging interval, so as to expand the application scope of online health status estimation of lithium batteries.

[0007] In a first aspect, the present invention provides a method for estimating a battery health state supporting any charging interval, comprising the following steps:

[0008] Step S10, obtaining a large amount of charging data of the lithium battery;

[0009] Step S20, pre-processing each of the charging data;

[0010] Step S30, obtaining the SOH true value corresponding to each of the charging data, and constructing an online charging data set based on the preprocessed charging data and the SOH true value;

[0011] Step S40, using the time series feature extraction tool tsfresh to extract features from the online charging data set to obtain 64-dimensional time series features;

[0012] Step S50: sorting and screening the time series features by importance to obtain the 12-dimensional time series features;

[0013] Step S60: create a health state estimation model, use the 12-dimensional time series features to train the health state estimation model, and use the trained health state estimation model to perform online health state estimation of the lithium battery.

[0014] Further, in the step S10, the charging data includes voltage time series data, current time series data and SOC value in different SOC intervals;

[0015] The step S20 is specifically as follows:

[0016] A deviation ratio threshold is set, and the voltage or current in the charging data whose deviation from the mean exceeds the deviation ratio threshold is eliminated, and the missing voltage or current is filled based on the time adjacent value.

[0017] Further, in step S30, the online charging data set includes input X and output Y;

[0018] X = {V, I, SOC};

[0019] Y = {SOH};

[0020] Among them, V represents the charging voltage; I represents the charging current; SOC represents the SOC value under the current charging voltage and charging current; SOH represents the actual value of the current health status of the lithium battery;

[0021] In step S40, the feature extraction is specifically as follows:

[0022] Convert the input X = {V, I, SOC} to X' = {tsfresh(V) 1-64 ,tsfresh(I) 1-64 ,SOC min ,SOC max};

[0023] Where X' represents the time series feature set; tsfresh(V)1-64 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 64-dimensional features; tsfresh(I) 1-64 Indicates the feature extraction of the charging current tsfresh time series to obtain 64-dimensional features; SOC min Indicates the starting value of SOC during charging; SOC max Indicates the end value of SOC during charging.

[0024] Furthermore, the step S50 specifically includes:

[0025] Step S51, after arranging the 64-dimensional time series features in random order, merge them with the initial time series features;

[0026] Step S52, calculating the importance of each of the time series features by using a random forest algorithm, and determining whether the number of repetitions is greater than a preset threshold value, if so, proceeding to step S53; if not, proceeding to step S51;

[0027] Step S53: Select the 12-dimensional time series features with changed importance:

[0028] X'={tsfresh(V) 1-12 ,tsfresh(I) 1-12 ,SOC min ,SOC max};

[0029] Among them, tsfresh(V) 1-12 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 12-dimensional features; tsfresh(I) 1-12 It means that the feature extraction of the tsfresh time series of the charging current is performed to obtain 12-dimensional features.

[0030] Furthermore, in step S60, the health status estimation model is created based on LightGBM or xgboost.

[0031] In a second aspect, the present invention provides a battery health status estimation system supporting any charging interval, comprising the following modules:

[0032] Charging data acquisition module, used to obtain a large amount of charging data of lithium batteries;

[0033] A charging data preprocessing module, used for preprocessing each of the charging data;

[0034] An online charging data set construction module, used to obtain the SOH true value corresponding to each of the charging data, and to construct an online charging data set based on the preprocessed charging data and the SOH true value;

[0035] A time series feature extraction module, used to extract features from the online charging data set using a time series feature extraction tool tsfresh to obtain a 64-dimensional time series feature;

[0036] A time series feature screening module, used to sort and screen the time series features by importance to obtain the 12-dimensional time series features;

[0037] The health state estimation module is used to create a health state estimation model, train the health state estimation model using the 12-dimensional time series features, and use the trained health state estimation model to perform online health state estimation of the lithium battery.

[0038] Further, in the charging data acquisition module, the charging data includes voltage time series data, current time series data and SOC value in different SOC intervals;

[0039] The charging data preprocessing module is specifically used for:

[0040] A deviation ratio threshold is set, and the voltage or current in the charging data whose deviation from the mean exceeds the deviation ratio threshold is eliminated, and the missing voltage or current is filled based on the time adjacent value.

[0041] Further, in the online charging data set construction module, the online charging data set includes an input X and an output Y;

[0042] X = {V, I, SOC};

[0043] Y = {SOH};

[0044] Among them, V represents the charging voltage; I represents the charging current; SOC represents the SOC value under the current charging voltage and charging current; SOH represents the actual value of the current health status of the lithium battery;

[0045] In the time series feature extraction module, the feature extraction is specifically:

[0046] Convert the input X = {V, I, SOC} to X' = {tsfresh(V) 1-64 ,tsfresh(I) 1-64 ,SOC min ,SOC max};

[0047] Where X' represents the time series feature set; tsfresh(V) 1-64 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 64-dimensional features; tsfresh(I) 1-64Indicates the feature extraction of the charging current tsfresh time series to obtain 64-dimensional features; SOC min Indicates the starting value of SOC during charging; SOC max Indicates the end value of SOC during charging.

[0048] Furthermore, the timing feature screening module specifically includes:

[0049] A feature reorganization unit, used for arranging the 64-dimensional time series features in random order and merging them with the initial time series features;

[0050] An importance calculation unit, used to calculate the importance of each of the time series features through a random forest algorithm, and determine whether the number of repetitions is greater than a preset threshold value. If so, enter the feature screening unit; if not, enter the feature recombination unit;

[0051] The feature screening unit is used to select the 12-dimensional time series features with changed importance:

[0052] X'={tsfresh(V) 1-12 ,tsfresh(I) 1-12 ,SOC min ,SOC max};

[0053] Among them, tsfresh(V) 1-12 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 12-dimensional features; tsfresh(I) 1-12 It means that the feature extraction of the tsfresh time series of the charging current is performed to obtain 12-dimensional features.

[0054] Furthermore, in the health status estimation module, the health status estimation model is created based on LightGBM or xgboost.

[0055] The advantages of the present invention are:

[0056] By acquiring charging data including voltage timing data, current timing data and SOC value in different SOC intervals, the true SOH value corresponding to the charging data is obtained, and an online charging data set is constructed based on the preprocessed charging data and the true SOH value. The timing feature extraction tool tsfresh is used to automatically extract features from the online charging data set to replace the traditional manual extraction. The extracted timing features are sorted and screened by importance, and the created health state estimation model is trained using the screened timing features to ensure the accuracy and generalization of the health state estimation model. Ultimately, the health state estimation model can estimate the health state of lithium batteries in any SOC interval (charging interval), thereby greatly expanding the scope of application of online health state estimation of lithium batteries. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The present invention will be further described below in conjunction with embodiments with reference to the accompanying drawings.

[0058] Figure 1 It is a flow chart of a battery health status estimation method supporting any charging interval of the present invention.

[0059] Figure 2 It is a structural schematic diagram of a battery health status estimation system supporting any charging interval of the present invention. DETAILED DESCRIPTION

[0060] The technical solution in the embodiments of the present application has the following overall idea: by obtaining the charging data under different SOC intervals and the real SOH value corresponding to the electrical data, an online charging data set is constructed based on each charging data and the real SOH value, and the time series feature extraction tool tsfresh is used to automatically extract features from the online charging data set to replace the traditional manual extraction, the extracted time series features are sorted and screened by importance, and then the created health state estimation model is trained using the screened time series features, so that the health state estimation model can estimate the health state of lithium batteries in any SOC interval, thereby expanding the scope of application of online health state estimation of lithium batteries.

[0061] Please refer to Figure 1 to Figure 2 As shown, a preferred embodiment of a method for estimating a battery health state supporting any charging interval of the present invention includes the following steps:

[0062] Step S10, obtaining a large amount of charging data of the lithium battery;

[0063] Step S20, pre-processing each of the charging data;

[0064] Step S30, obtaining the SOH true value corresponding to each of the charging data, and constructing an online charging data set based on the preprocessed charging data and the SOH true value;

[0065] The SOH true value is the discharge capacity of a lithium battery after it has been left at room temperature for a period of time, charged to a full charge state using constant current and constant voltage, and then discharged to a cut-off voltage using constant current. In actual measurement, the SOH true value is generally calibrated after a fixed number of cycles. Therefore, it is necessary to perform curve fitting based on the discrete SOH true value to obtain the SOH true value under each cycle. The double exponential empirical capacity decay model can be used for calculation:

[0066] y=a·exp(b·k)+c·exp(d·k);

[0067] Where a, b, c, and d represent fitting parameters; k represents the number of cycles; and y represents the true value of SOH.

[0068] The network loss function uses the square root error between the true SOH value and the predicted SOH value, and the formula is as follows:

[0069]

[0070] Among them, y prodict represents the predicted value of SOH; y truth represents the true value of SOH; w i represents the weight of different charging cycle numbers; m represents the total number of charging cycles; n represents the fitting parameter;

[0071] Step S40: automatically extracting features from the online charging data set using a time series feature extraction tool tsfresh to obtain 64-dimensional time series features;

[0072] tsfresh is a Python time series data feature extraction module that can automatically calculate a large number of time series features, including a variety of statistical and transformation-based time series extraction methods, such as absolute energy value, absolute sum of first-order differences, aggregated statistical features of autocorrelation coefficients of various orders, approximate entropy, autoregressive coefficient, ADF test, lag-order autocorrelation, etc.; tsfresh can be used to convert the indefinite-length input voltage and current data of any charging SOC interval into fixed-length voltage and current feature data, which is conducive to subsequent modeling;

[0073] Step S50: sorting and screening the time series features by importance to obtain the 12-dimensional time series features;

[0074] Step S60: create a health state estimation model, use the 12-dimensional time series features to train the health state estimation model, and use the trained health state estimation model to perform online health state estimation of the lithium battery.

[0075] In step S10, the charging data includes voltage timing data, current timing data and SOC values ​​in different SOC intervals; since the discharge conditions of lithium batteries in actual applications are often more complicated than the charging conditions, and the charging data of the charging conditions are also easier to obtain (for example, obtaining charging condition data when charging new energy vehicles), the present invention performs health status detection based on the charging conditions of lithium batteries.

[0076] The step S20 is specifically as follows:

[0077] A deviation ratio threshold is set, and the voltage or current in the charging data whose deviation from the mean exceeds the deviation ratio threshold is eliminated, and the missing voltage or current is filled based on the time adjacent value.

[0078] In step S30, the online charging data set includes input X and output Y;

[0079] X = {V, I, SOC};

[0080] Y = {SOH};

[0081] Among them, V represents the charging voltage; I represents the charging current; SOC represents the SOC value under the current charging voltage and charging current; SOH represents the actual value of the current health status of the lithium battery;

[0082] In step S40, the feature extraction is specifically as follows:

[0083] Convert the input X = {V, I, SOC} to X' = {tsfresh(V) 1-64 ,tsfresh(I) 1-64 ,SOC min ,SOC max};

[0084] Where X' represents the time series feature set; tsfresh(V) 1-64 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 64-dimensional features; tsfresh(I) 1-64 Indicates the feature extraction of the charging current tsfresh time series to obtain 64-dimensional features; SOC min Indicates the starting value of SOC during charging; SOC max Indicates the end value of SOC during charging.

[0085] The step S50 specifically includes:

[0086] Step S51, after arranging the 64-dimensional time series features in random order, merge them with the initial time series features;

[0087] Step S52, calculating the importance of each of the time series features by using a random forest algorithm, and determining whether the number of repetitions is greater than a preset threshold value, if so, proceeding to step S53; if not, proceeding to step S51;

[0088] Step S53: Select the 12-dimensional time series features with the greatest change in importance:

[0089] X'={tsfresh(V) 1-12 ,tsfresh(I) 1-12 ,SOC min ,SOC max};

[0090] Among them, tsfresh(V) 1-12 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 12-dimensional features; tsfresh(I) 1-12 It means that the feature extraction of the tsfresh time series of the charging current is performed to obtain 12-dimensional features.

[0091] That is, the extracted time series features are sorted and screened by importance through the Boruta method; since the number of output tsfresh time series features is large and the amount of calculation is large, in order to improve the algorithm operation efficiency and avoid extracting irrelevant features, it is necessary to sort the calculated time series features by importance. The general method uses the Pearson correlation coefficient to sort the feature importance. The closer the obtained correlation coefficient is to 1, the higher the importance; however, the feature importance sorting based on the Pearson correlation coefficient can only calculate the correlation of a single feature, and cannot mine the correlation between multiple features, and each feature needs to be calculated separately, which has calculation redundancy. Therefore, the present invention adopts a feature importance selection strategy based on the Boruta method.

[0092] The goal of Boruta is to select all feature sets related to the dependent variable, rather than selecting a feature set that can minimize the model cost function for a specific model. The significance of the Boruta algorithm is that it can more comprehensively understand the influencing factors of the dependent variable, so as to perform feature selection better and more efficiently.

[0093] In step S60, the health status estimation model is created based on LightGBM or xgboost.

[0094] A preferred embodiment of a battery health status estimation system supporting any charging interval of the present invention includes the following modules:

[0095] Charging data acquisition module, used to obtain a large amount of charging data of lithium batteries;

[0096] A charging data preprocessing module, used for preprocessing each of the charging data;

[0097] An online charging data set construction module, used to obtain the SOH true value corresponding to each of the charging data, and to construct an online charging data set based on the preprocessed charging data and the SOH true value;

[0098] The SOH true value is the discharge capacity of a lithium battery after it has been left at room temperature for a period of time, charged to a full charge state using constant current and constant voltage, and then discharged to a cut-off voltage using constant current. In actual measurement, the SOH true value is generally calibrated after a fixed number of cycles. Therefore, it is necessary to perform curve fitting based on the discrete SOH true value to obtain the SOH true value under each cycle. The double exponential empirical capacity decay model can be used for calculation:

[0099] y=a·exp(b·k)+c·exp(d·k);

[0100] Where a, b, c, and d represent fitting parameters; k represents the number of cycles; and y represents the true value of SOH.

[0101] The network loss function uses the square root error between the true SOH value and the predicted SOH value, and the formula is as follows:

[0102]

[0103] Among them, y prodict represents the predicted value of SOH; y truth represents the true value of SOH; w i represents the weight of different charging cycle numbers; m represents the total number of charging cycles; n represents the fitting parameter;

[0104] A time series feature extraction module, used to automatically extract features from the online charging data set using a time series feature extraction tool tsfresh to obtain a 64-dimensional time series feature;

[0105] tsfresh is a Python time series data feature extraction module that can automatically calculate a large number of time series features, including a variety of statistical and transformation-based time series extraction methods, such as absolute energy value, absolute sum of first-order differences, aggregated statistical features of autocorrelation coefficients of various orders, approximate entropy, autoregressive coefficient, ADF test, lag-order autocorrelation, etc.; tsfresh can be used to convert the indefinite-length input voltage and current data of any charging SOC interval into fixed-length voltage and current feature data, which is conducive to subsequent modeling;

[0106] A time series feature screening module, used to sort and screen the time series features by importance to obtain the 12-dimensional time series features;

[0107] The health state estimation module is used to create a health state estimation model, train the health state estimation model using the 12-dimensional time series features, and use the trained health state estimation model to perform online health state estimation of the lithium battery.

[0108] In the charging data acquisition module, the charging data includes voltage timing data, current timing data and SOC values ​​in different SOC intervals; since the discharge conditions of lithium batteries in actual applications are often more complicated than the charging conditions, and the charging data of the charging conditions are also easier to obtain (for example, obtaining charging condition data when charging new energy vehicles), the present invention performs health status detection based on the charging conditions of lithium batteries.

[0109] The charging data preprocessing module is specifically used for:

[0110] A deviation ratio threshold is set, and the voltage or current in the charging data whose deviation from the mean exceeds the deviation ratio threshold is eliminated, and the missing voltage or current is filled based on the time adjacent value.

[0111] In the online charging data set construction module, the online charging data set includes an input X and an output Y;

[0112] X = {V, I, SOC};

[0113] Y = {SOH};

[0114] Among them, V represents the charging voltage; I represents the charging current; SOC represents the SOC value under the current charging voltage and charging current; SOH represents the actual value of the current health status of the lithium battery;

[0115] In the time series feature extraction module, the feature extraction is specifically:

[0116] Convert the input X = {V, I, SOC} to X' = {tsfresh(V) 1-64 ,tsfresh(I) 1-64 ,SOC min ,SOC max};

[0117] Where X' represents the time series feature set; tsfresh(V) 1-64 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 64-dimensional features; tsfresh(I) 1-64 Indicates the feature extraction of the charging current tsfresh time series to obtain 64-dimensional features; SOCmin Indicates the starting value of SOC during charging; SOC max Indicates the end value of SOC during charging.

[0118] The timing feature screening module specifically includes:

[0119] A feature reorganization unit, used for arranging the 64-dimensional time series features in random order and merging them with the initial time series features;

[0120] An importance calculation unit, used to calculate the importance of each of the time series features through a random forest algorithm, and determine whether the number of repetitions is greater than a preset threshold value. If so, enter the feature screening unit; if not, enter the feature recombination unit;

[0121] The feature screening unit is used to select the 12-dimensional time series features with the greatest change in importance:

[0122] X'={tsfresh(V) 1-12 ,tsfresh(I) 1-12 ,SOC min ,SOC max};

[0123] Among them, tsfresh(V) 1-12 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 12-dimensional features; tsfresh(I) 1-12 It means that the feature extraction of the tsfresh time series of the charging current is performed to obtain 12-dimensional features.

[0124] That is, the extracted time series features are sorted and screened by importance through the Boruta method; since the number of output tsfresh time series features is large and the amount of calculation is large, in order to improve the algorithm operation efficiency and avoid extracting irrelevant features, it is necessary to sort the calculated time series features by importance. The general method uses the Pearson correlation coefficient to sort the feature importance. The closer the obtained correlation coefficient is to 1, the higher the importance; however, the feature importance sorting based on the Pearson correlation coefficient can only calculate the correlation of a single feature, and cannot mine the correlation between multiple features, and each feature needs to be calculated separately, which has calculation redundancy. Therefore, the present invention adopts a feature importance selection strategy based on the Boruta method.

[0125] The goal of Boruta is to select all feature sets related to the dependent variable, rather than selecting a feature set that can minimize the model cost function for a specific model. The significance of the Boruta algorithm is that it can more comprehensively understand the influencing factors of the dependent variable, so as to perform feature selection better and more efficiently.

[0126] In the health status estimation module, the health status estimation model is created based on LightGBM or xgboost.

[0127] In summary, the advantages of the present invention are:

[0128] By acquiring charging data including voltage timing data, current timing data and SOC value in different SOC intervals, the true SOH value corresponding to the charging data is obtained, and an online charging data set is constructed based on the preprocessed charging data and the true SOH value. The timing feature extraction tool tsfresh is used to automatically extract features from the online charging data set to replace the traditional manual extraction. The extracted timing features are sorted and screened by importance, and the created health state estimation model is trained using the screened timing features to ensure the accuracy and generalization of the health state estimation model. Ultimately, the health state estimation model can estimate the health state of lithium batteries in any SOC interval (charging interval), thereby greatly expanding the scope of application of online health state estimation of lithium batteries.

[0129] Although the specific implementation modes of the present invention are described above, those skilled in the art should understand that the specific implementation modes described are only illustrative and are not intended to limit the scope of the present invention. Equivalent modifications and changes made by those skilled in the art in accordance with the spirit of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A battery health status estimation method that supports any charging interval, Features: The steps include: Step S10, obtaining a large amount of charging data of the lithium battery; Step S20, pre-processing each of the charging data; Step S30, obtaining the SOH true value corresponding to each of the charging data, and constructing an online charging data set based on the preprocessed charging data and the SOH true value; Step S40, using the time series feature extraction tool tsfresh to extract features from the online charging data set to obtain 64-dimensional time series features; Step S50: sorting and screening the time series features by importance to obtain the 12-dimensional time series features; Step S60: creating a health state estimation model, using the 12-dimensional time series features to train the health state estimation model, and using the trained health state estimation model to perform online health state estimation of the lithium battery; The step S50 specifically includes: Step S51, after arranging the 64-dimensional time series features in random order, merge them with the initial time series features; Step S52, calculating the importance of each of the time series features by using a random forest algorithm, and determining whether the number of repetitions is greater than a preset threshold value, if so, proceeding to step S53; if not, proceeding to step S51; Step S53: Select the 12-dimensional time series features with changed importance: X'={tsfresh(V) 1-12 ,tsfresh(I) 1-12 ,SOC min ,SOC max }; Among them, tsfresh(V) 1-12 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 12-dimensional features; tsfresh(I) 1-12 It means that the feature extraction of the tsfresh time series of the charging current is performed to obtain 12-dimensional features.

2. A method for estimating a battery health state supporting any charging interval as claimed in claim 1, Features: In the step S10, the charging data includes voltage time series data, current time series data and SOC values ​​in different SOC intervals; The step S20 is specifically as follows: A deviation ratio threshold is set, and the voltage or current in the charging data whose deviation from the mean exceeds the deviation ratio threshold is eliminated, and the missing voltage or current is filled based on the time adjacent value.

3. A method for estimating a battery health state supporting any charging interval as claimed in claim 1, Features: In step S30, the online charging data set includes input X and output Y; X = {V, I, SOC}; Y = {SOH}; Among them, V represents the charging voltage; I represents the charging current; SOC represents the SOC value under the current charging voltage and charging current; SOH represents the actual value of the current health status of the lithium battery; In step S40, the feature extraction is specifically as follows: Convert the input X = {V, I, SOC} to X' = {tsfresh(V) 1-64 ,tsfresh(I) 1-64 ,SOC min ,SOC max }; Where X' represents the time series feature set; tsfresh(V) 1-64 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 64-dimensional features; tsfresh(I) 1-64 Indicates the feature extraction of the charging current tsfresh time series to obtain 64-dimensional features; SOC min Indicates the starting value of SOC during charging; SOC max Indicates the end value of SOC during charging.

4. A method for estimating a battery health state supporting any charging interval as claimed in claim 1, Features: In step S60, the health status estimation model is created based on LightGBM or xgboost.

5. A battery health status estimation system that supports any charging interval, Features: Includes the following modules: Charging data acquisition module, used to obtain a large amount of charging data of lithium batteries; A charging data preprocessing module, used for preprocessing each of the charging data; An online charging data set construction module, used to obtain the SOH true value corresponding to each of the charging data, and to construct an online charging data set based on the preprocessed charging data and the SOH true value; A time series feature extraction module, used to extract features from the online charging data set using a time series feature extraction tool tsfresh to obtain a 64-dimensional time series feature; A time series feature screening module, used to sort and screen the time series features by importance to obtain the 12-dimensional time series features; A health state estimation module is used to create a health state estimation model, train the health state estimation model using the 12-dimensional time series features, and perform online health state estimation of the lithium battery using the trained health state estimation model; The timing feature screening module specifically includes: A feature reorganization unit, used for arranging the 64-dimensional time series features in random order and merging them with the initial time series features; An importance calculation unit, used to calculate the importance of each of the time series features through a random forest algorithm, and determine whether the number of repetitions is greater than a preset threshold value. If so, enter the feature screening unit; if not, enter the feature recombination unit; The feature screening unit is used to select the 12-dimensional time series features with changed importance: X'={tsfresh(V) 1-12 ,tsfresh(I) 1-12 ,SOC min ,SOC max }; Among them, tsfresh(V) 1-12 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 12-dimensional features; tsfresh(I) 1-12 It means that the feature extraction of the tsfresh time series of the charging current is performed to obtain 12-dimensional features.

6. A battery health status estimation system supporting any charging interval as claimed in claim 5, Features: In the charging data acquisition module, the charging data includes voltage time series data, current time series data and SOC value in different SOC intervals; The charging data preprocessing module is specifically used for: A deviation ratio threshold is set, and the voltage or current in the charging data whose deviation from the mean exceeds the deviation ratio threshold is eliminated, and the missing voltage or current is filled based on the time adjacent value.

7. A battery health status estimation system supporting any charging interval as claimed in claim 5, Features: In the online charging data set construction module, the online charging data set includes an input X and an output Y; X = {V, I, SOC}; Y = {SOH}; Among them, V represents the charging voltage; I represents the charging current; SOC represents the SOC value under the current charging voltage and charging current; SOH represents the actual value of the current health status of the lithium battery; In the time series feature extraction module, the feature extraction is specifically: Convert the input X = {V, I, SOC} to X' = {tsfresh(V) 1-64 ,tsfresh(I) 1-64 ,SOC min ,SOC max }; Where X' represents the time series feature set; tsfresh(V) 1-64 Indicates the feature extraction of the charging voltage tsfresh time series to obtain 64-dimensional features; tsfresh(I) 1-64 Indicates the feature extraction of the charging current tsfresh time series to obtain 64-dimensional features; SOC min Indicates the starting value of SOC during charging; SOC max Indicates the end value of SOC during charging.

8. A battery health status estimation system supporting any charging interval as claimed in claim 5, Features: In the health status estimation module, the health status estimation model is created based on LightGBM or xgboost.

Images