A method and system for evaluating the health status of a battery pack

Abstract

This invention discloses a method and system for assessing the health status of a battery pack. The method includes: obtaining charge and discharge operation data within a first SOC interval from historical charge and discharge operation data using a preset SOC interval threshold; obtaining health indicators corresponding to the battery pack to be assessed from the charge and discharge operation data using a preset health indicator extraction formula; performing feature filtering on the health indicators to obtain a health indicator feature set corresponding to the battery pack to be assessed; inputting the health indicator feature set into a preset battery pack health status assessment model; assessing the health status of the battery pack to be assessed; and obtaining the assessment result corresponding to the battery pack to be assessed, thereby improving the efficiency and accuracy of battery pack health status assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage battery evaluation, and in particular to a method and system for evaluating the health status of a battery pack. Background Art

[0002] With the rapid development of my country's lithium-ion battery industry, lithium-ion batteries have been widely used in electric vehicles and energy storage power stations. Accurate assessment of battery health status is an important guarantee for the safe and stable operation of the system. Inaccurate health status assessment will affect the performance of the battery and even cause safety problems such as thermal runaway. In practical applications, a battery pack is usually composed of a large number of battery cells connected in series and parallel to meet the battery system's requirements for large capacity and high power. Therefore, conducting health status assessment of lithium-ion battery packs is more in line with actual needs.

[0003] However, there are initial performance differences between battery cells before they are grouped. During the cycle aging process of the battery pack, the differences in temperature, voltage and mechanical stress of each cell cause different changes in internal performance, further increasing the inconsistency between cells. Therefore, the performance degradation of the battery pack is not only caused by the aging of its internal cells, but also affected by the inconsistency differences between cells.

[0004] In addition, compared with battery cells, the working process of battery packs is more complicated, and the number and types of parameters that can be monitored are reduced. At present, there are still problems in the research on health status estimation of lithium-ion battery packs, such as extracting features based on the wide SOC range of batteries or complete charge and discharge curves that are difficult to obtain in actual scenarios, and modeling through large sample data. Summary of the invention

[0005] The invention discloses a method and system for evaluating the health status of a battery pack, which improve the accuracy and efficiency of evaluating the health status of a battery pack.

[0006] In order to achieve the above object, the present invention discloses a method for evaluating the health status of a battery pack, comprising:

[0007] Acquire historical charge and discharge operation data of the battery pack to be evaluated, and acquire charge and discharge operation data in a first SOC interval from the historical charge and discharge operation data through a preset SOC interval threshold; the charge and discharge operation data includes charging data of the battery pack in the first SOC interval and battery pack operation data at a time point when the charge and discharge mode of the battery pack is switched in the first SOC interval;

[0008] According to the charge and discharge operation data in the first SOC interval, a health index corresponding to the battery pack to be evaluated is obtained by using a preset health index extraction formula;

[0009] Perform feature screening on the health indicators by using a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated;

[0010] The health indicator feature set is input into a preset battery pack health status assessment model, a health status assessment is performed on the battery pack to be assessed, and an assessment result corresponding to the battery pack to be assessed is obtained.

[0011] The present invention discloses a method for evaluating the health status of a battery pack. First, historical charge and discharge operation data of a battery pack to be evaluated is obtained, and the charge and discharge operation data within a partial SOC range is obtained from the historical charge and discharge operation data. On the one hand, the amount of data processing is reduced and the evaluation efficiency is improved. On the other hand, the health status evaluation is performed based on stable charge and discharge operation data, which can ensure the accuracy of the evaluation result. After obtaining the historical operation data after screening, the screened charge and discharge operation data is subjected to health indicator extraction by a preset health indicator extraction formula, so that the evaluation is performed according to the extracted health indicator, further improving the evaluation efficiency. Then, the health indicator is subjected to feature screening to solve the model overfitting caused by the redundancy of the health indicator and improve the accuracy of the evaluation. Finally, the input feature set is subjected to health status evaluation by using a preset evaluation model to obtain the health status evaluation result of the battery pack.

[0012] As a preferred example, the acquiring of historical charge and discharge operation data of the battery pack to be evaluated, and acquiring charge and discharge operation data in a first SOC interval from the historical charge and discharge operation data by using a preset SOC interval threshold, includes:

[0013] Obtaining a discharge cut-off voltage and a charge cut-off voltage of the battery pack to be evaluated, and obtaining historical charge and discharge operation data of the battery pack to be evaluated according to the discharge cut-off voltage and the charge cut-off voltage;

[0014] Extracting charging data of the battery pack in the first SOC interval within the SOC interval threshold from the historical charging and discharging operation data corresponding to the battery pack to be evaluated according to a preset SOC interval threshold;

[0015] According to the SOC interval threshold, the switching time of the current value in the historical charge and discharge operation data is obtained, and the historical charge and discharge operation data corresponding to the switching time is obtained to obtain the battery pack operation data at the time point of the battery pack charge and discharge mode switching in the first SOC interval.

[0016] The present invention uses the charging cut-off voltage and the discharging cut-off voltage of the battery pack as cycle nodes to obtain historical charging and discharging operation data during the complete operation of the battery pack, and extracts part of the operation data from the historical charging and discharging operation data through a preset SOC interval threshold. On the one hand, stable operation data is used for evaluation to improve the accuracy of the evaluation, and on the other hand, the amount of data processing is reduced to improve the efficiency of the evaluation.

[0017] As a preferred example, according to the charge and discharge operation data in the first SOC interval, obtaining the health index corresponding to the battery pack to be evaluated by a preset health index extraction formula includes:

[0018] According to the charge and discharge operation data in the first SOC interval, a first health indicator characterizing the overall degradation trend of the battery pack to be evaluated is obtained by using a preset first health indicator extraction formula;

[0019] According to the charge and discharge operation data corresponding to each single battery in the battery pack to be evaluated in the first SOC range, a second health indicator characterizing the inconsistency difference between the single batteries is obtained by a preset second health indicator extraction formula.

[0020] The present invention utilizes different preset health index extraction formulas to respectively extract health indicators that characterize the overall performance degradation trend of the battery pack and the inconsistency differences between cells in the battery pack, so as to evaluate the battery pack based on the differences between the whole and the cells, thereby improving the comprehensiveness of the evaluation and further improving the accuracy of the evaluation results.

[0021] As a preferred example, obtaining the first health indicator characterizing the overall degradation trend of the battery pack to be evaluated by using a preset first health indicator extraction formula includes:

[0022] The average voltage of the battery pack to be evaluated in the first SOC interval is obtained, and the calculation formula of the average voltage is as follows:

[0023]

[0024] Among them, the U pack is the battery pack voltage, I is the battery pack charging current, Is the battery pack at t 1 -t 2 The average voltage in some SOC intervals during the charging process;

[0025] Obtain the capacity standard deviation stdQ of the battery pack to be evaluated within the first SOC interval pack and capacity difference standard deviation stddQ pack , the acquisition process is as follows:

[0026] Obtain the voltage range [V pack_min ,V pack_max ] and the voltage range is V interval Equally spaced segments:

[0027] V pack_segment =[V pack_min ,V pack_min +V interval ,V pack_min +V interval *2,...,V pack_max ]

[0028] The capacity sequence of the battery pack to be evaluated within the voltage range is calculated by the ampere-hour integration method:

[0029] Q pack_segment(i) =[Q 2 -Q 1 ,Q 3 -Q 2 ,...,Q n -Q n-1 ],(i=1,2,...,N)

[0030] Among them, 1, 2, ..., n corresponds to the corresponding voltage points within the battery pack voltage range; 1, 2, ..., N corresponds to the number of cycles of the battery pack, Q pack_segment(i) is the capacity sequence of the battery pack under the i-th cycle;

[0031] Subtract the capacity sequence of the battery pack to be evaluated under different cycles from the capacity sequence under the first cycle to obtain the capacity difference sequence:

[0032] Q pack_segment_difference(i) =Q pack_segment(i) -Q pack_segment(1) ,(i=1,2,...,N)

[0033] Among them, Q pack_segment_difference(i) is the capacity difference sequence of the battery pack at the i-th cycle;

[0034] Obtain the capacity sequence Q of the battery pack to be evaluated under different cycles respectively pack_segment The standard deviation stdQ pack and the battery capacity difference sequence Q to be evaluated pack_segment_difference The standard deviation of pack ;

[0035] Obtain the battery pack voltage U at the time of charging mode switching during the charging process of the battery pack to be evaluated pack_changepoint ;

[0036] Obtain the battery pack voltage U after the battery pack to be evaluated is charged and left to stand for a certain period of time pack_rest ;

[0037] Get the voltage drop U of the battery pack within a certain time of instantaneous discharge when the battery pack to be evaluated starts discharging packdrop_5s .

[0038] The present invention extracts a health indicator characterizing the overall performance degradation trend of the battery pack based on the charge and discharge operation data within the first SOC interval, so as to provide data for health status assessment from the overall aspect of the battery pack, thereby improving the accuracy of health status assessment.

[0039] As a preferred example, obtaining the second health indicator characterizing the inconsistency difference between the single cells by using a preset second health indicator extraction formula includes:

[0040] Extract the same features as the battery pack to be evaluated from each single cell in the battery pack to be evaluated Q cell(i) 、stddQ cell(i) With point feature U cell(i)_changepoint , U cell(i)_rest , U cell(i)drop_5s ;

[0041] The range, standard deviation and information entropy between the characteristics of each single cell are calculated to characterize the inconsistency differences between the single cell voltage, capacity and internal resistance. The expressions are as follows:

[0042] ran=x max -x min

[0043] Where, i = 1, 2, ..., n, n is the number of single cells in the battery pack; x max is the maximum value of the single cell characteristic quantity in the battery pack; x min It is the minimum value of the characteristic quantity of the single cell in the battery pack;

[0044]

[0045] Where n is the number of single cells connected in series in the battery pack; x i is the characteristic quantity of the i-th monomer; It is the average value of the characteristic quantities of all series-connected cells in the battery pack;

[0046]

[0047] Where n is the number of single cells connected in series in the battery pack; p(i) is the characteristic value corresponding to the i-th cell.

[0048] On the basis of obtaining the first health indicator, the present invention obtains the health indicator characterizing the inconsistency difference between the battery cells in the battery pack by obtaining the corresponding partial charge and discharge data characterizing the battery cells in the battery pack, so as to subsequently comprehensively evaluate the health status of the battery pack from the perspective of the battery pack as a whole and the cells, thereby improving the accuracy of the evaluation.

[0049] As a preferred example, the feature selection of the health indicator is performed by a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated, including:

[0050] Calculating a Pearson correlation coefficient between the health indicator and the battery pack capacity of the battery pack to be evaluated, and filtering the health indicator according to the Pearson correlation coefficient to obtain an initial health indicator feature set;

[0051] The health indicator feature set is obtained by performing a forward sequence search on the initial health indicator feature set using a preset packaging method and adding new features to the initial health indicator feature set.

[0052] The present invention first calculates the Pearson correlation coefficient between each health indicator and the battery pack capacity, and screens the health indicators according to the Pearson correlation coefficient to reduce the amount of data processing and improve the efficiency of evaluation. At the same time, the health indicators are searched using a packaging method to obtain the optimal health indicator feature set and improve the accuracy of the evaluation.

[0053] As a preferred example, the calculating of the Pearson correlation coefficient between the health indicator and the battery pack capacity of the battery pack to be evaluated includes:

[0054] The Pearson correlation coefficient is calculated using a preset calculation formula, which is:

[0055]

[0056] Among them, x i and i is the true value of the health indicator and battery pack capacity, and is the corresponding mean of the two. The closer the absolute value of the result is to 1, the higher the correlation is.

[0057] The present invention provides a Pearson coefficient calculation formula to calculate the correlation between various health indicators and battery pack capacity, so as to extract health indicators with higher correlation for status evaluation and improve the accuracy and efficiency of evaluation.

[0058] On the other hand, the present invention discloses a battery pack health status assessment system, including a data acquisition module, an indicator extraction module, a feature screening module and a status assessment module;

[0059] The data acquisition module is used to acquire historical charge and discharge operation data of the battery pack to be evaluated, and acquire charge and discharge operation data within a first SOC interval from the historical charge and discharge operation data through a preset SOC interval threshold; the charge and discharge operation data includes the charging data of the battery pack within the first SOC interval and the battery pack operation data at the time point of switching the charge and discharge mode of the battery pack within the first SOC interval;

[0060] The index extraction module is used to obtain the health index corresponding to the battery pack to be evaluated through a preset health index extraction formula according to the charging and discharging operation data in the first SOC interval;

[0061] The feature screening module is used to perform feature screening on the health indicators by a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated;

[0062] The state evaluation module is used to input the health indicator feature set into a preset battery pack health state evaluation model, perform health state evaluation on the battery pack to be evaluated, and obtain an evaluation result corresponding to the battery pack to be evaluated.

[0063] The present invention discloses a health status assessment system for a battery pack. The system first obtains historical charge and discharge operation data of a battery pack to be assessed, and screens the charge and discharge operation data within a partial SOC range from the historical charge and discharge operation data. On the one hand, the data processing volume is reduced and the assessment efficiency is improved. On the other hand, the health status assessment is performed based on stable charge and discharge operation data, which can ensure the accuracy of the assessment result. After obtaining the historical operation data after screening, the screened charge and discharge operation data is subjected to health indicator extraction by a preset health indicator extraction formula, so that the assessment is performed according to the extracted health indicator, further improving the assessment efficiency. The health indicator is then subjected to feature screening to solve the model overfitting caused by health indicator redundancy and improve the assessment accuracy. Finally, the input feature set is subjected to health status assessment using a preset assessment model to obtain the health status assessment result of the battery pack.

[0064] As a preferred example, the data acquisition module includes a historical data unit and a data extraction unit;

[0065] The historical data unit is used to obtain the discharge cut-off voltage and the charge cut-off voltage of the battery pack to be evaluated, and obtain the historical charge and discharge operation data of the battery pack to be evaluated according to the discharge cut-off voltage and the charge cut-off voltage;

[0066] The data extraction unit is used to extract the charging data of the battery group in the first SOC interval within the SOC interval threshold from the historical charging and discharging operation data corresponding to the battery group to be evaluated according to a preset SOC interval threshold; according to the SOC interval threshold, obtain the switching time of the current value in the historical charging and discharging operation data, and obtain the historical charging and discharging operation data corresponding to the switching time, and obtain the battery group operation data at the time point of switching the charging and discharging mode of the battery group in the first SOC interval.

[0067] The present invention uses the charging cut-off voltage and the discharging cut-off voltage of the battery pack as cycle nodes to obtain historical charging and discharging operation data during the complete operation of the battery pack, and extracts part of the operation data from the historical charging and discharging operation data through a preset SOC interval threshold. On the one hand, stable operation data is used for evaluation to improve the accuracy of the evaluation, and on the other hand, the amount of data processing is reduced to improve the efficiency of the evaluation.

[0068] As a preferred example, the indicator extraction module includes a first indicator unit and a second indicator unit;

[0069] The first indicator unit is used to obtain a first health indicator characterizing the overall degradation trend of the battery pack to be evaluated through a preset first health indicator extraction formula according to the charge and discharge operation data in the first SOC interval;

[0070] The second indicator unit is used to obtain a second health indicator characterizing the inconsistency difference between the single cells according to the charge and discharge operation data corresponding to each single cell in the battery pack to be evaluated in the first SOC range through a preset second health indicator extraction formula.

[0071] The present invention utilizes different preset health index extraction formulas to respectively extract health indicators that characterize the overall performance degradation trend of the battery pack and the inconsistency differences between cells in the battery pack, so as to evaluate the battery pack based on the differences between the whole and the cells, thereby improving the comprehensiveness of the evaluation and further improving the accuracy of the evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 : A flowchart of a method for evaluating the health status of a battery pack provided by an embodiment of the present invention;

[0073] Figure 2 : A structural diagram of a battery pack health status assessment system provided by an embodiment of the present invention;

[0074] Figure 3 : A schematic diagram of a battery pack capacity decay trend provided by an embodiment of the present invention;

[0075] Figure 4: A schematic diagram of a trend of an average voltage of a battery pack based on the number of cycles provided in an embodiment of the present invention;

[0076] Figure 5 : A schematic diagram of a standard deviation trend of a battery capacity sequence based on cycle number provided in an embodiment of the present invention;

[0077] Figure 6 : A schematic diagram of a standard deviation trend of a battery pack capacity difference sequence based on cycle number provided in an embodiment of the present invention;

[0078] Figure 7 : A schematic diagram of a battery pack voltage trend at a variable current moment based on the number of cycles provided in an embodiment of the present invention;

[0079] Figure 8 : A schematic diagram of a voltage trend of a battery pack after being fully charged and left to rest based on the number of cycles provided in an embodiment of the present invention;

[0080] Fig. 9 : A schematic diagram of a battery pack voltage drop trend within 5s of instantaneous discharge based on the number of cycles provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0081] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0082] Example

[0083] The embodiment of the present invention provides a method for evaluating the health status of a battery pack. For the specific implementation process of the evaluation method, please refer to Figure 1 , mainly including step 101 to step 104, the steps are:

[0084] Step 101: Obtain historical charge and discharge operation data of the battery pack to be evaluated, and obtain charge and discharge operation data within a first SOC interval from the historical charge and discharge operation data through a preset SOC interval threshold; the charge and discharge operation data include the charging data of the battery pack within the first SOC interval and the battery pack operation data at the time point of switching the charge and discharge mode of the battery pack within the first SOC interval.

[0085] In this embodiment, this step mainly includes: obtaining the discharge cut-off voltage and the charge cut-off voltage of the battery pack to be evaluated, and obtaining the historical charge and discharge operation data of the battery pack to be evaluated based on the discharge cut-off voltage and the charge cut-off voltage; extracting the charging data of the battery pack in the first SOC interval within the SOC interval threshold from the historical charge and discharge operation data corresponding to the battery pack to be evaluated according to a preset SOC interval threshold; according to the SOC interval threshold, obtaining the switching moment of the current value in the historical charge and discharge operation data, and obtaining the historical charge and discharge operation data corresponding to the switching moment, and obtaining the battery pack operation data at the time point of the charge and discharge mode switching of the battery pack in the first SOC interval.

[0086] Exemplarily, in this embodiment, the battery pack to be evaluated is set to be charged from the discharge cut-off voltage (the lowest voltage of the battery pack, if it is 28V) to the charge cut-off voltage (the highest voltage of the battery pack, if it is 34V), and then discharged from the charge cut-off voltage (the highest voltage of the battery pack, if it is 28V) to the discharge cut-off voltage (the lowest voltage of the battery pack, if it is 34V) as a complete charge and discharge process, and the charge and discharge operation data of the battery pack to be evaluated in the complete charge and discharge process is obtained. At the same time, according to the certain corresponding relationship between the battery pack SOC and the battery pack voltage, the battery pack voltage is used as a condition here to extract the battery pack charging data within a partial SOC range from the complete battery pack charge and discharge data. (For example, during the complete charging process of the battery pack, the battery pack voltage increases from 28V to 34V, and only the battery pack voltage, current, operating time and other data corresponding to the 30V-32V interval are extracted (the corresponding SOC interval is 30% SOC-70% SOC), which is the battery pack charging data within part of the SOC interval). The charging and discharging operation data at the time of switching the charging and discharging mode can be obtained by using the current as a judgment condition. The current of the battery pack to be evaluated during the charging and discharging process will not remain unchanged, and the current values ​​in different stages are different (for example, 100A charging in the first stage-50A charging in the second stage-25A charging in the third stage-200A discharging in the fourth stage). The battery pack current, voltage, and operating time at the switching moments between different stages (for example, when the current switches from 100A to 50A, and from 25A to -200A) are the battery pack operation data at the time of switching the charging and discharging mode.

[0087] Step 102: According to the charge and discharge operation data in the first SOC interval, a health index corresponding to the battery pack to be evaluated is obtained by using a preset health index extraction formula.

[0088] In this embodiment, this step mainly includes: according to the charging and discharging operation data within the first SOC range, obtaining the first health indicator characterizing the overall degradation trend of the battery group to be evaluated through a preset first health indicator extraction formula; according to the charging and discharging operation data corresponding to each single battery in the battery group to be evaluated within the first SOC range, obtaining the second health indicator characterizing the inconsistency difference between the single batteries through a preset second health indicator extraction formula.

[0089] Furthermore, in this embodiment, the first health indicator characterizing the overall degradation trend of the battery pack to be evaluated is obtained by using a preset first health indicator extraction formula, including:

[0090] The average voltage of the battery pack to be evaluated in the first SOC interval is obtained, and the calculation formula of the average voltage is as follows:

[0091]

[0092] Among them, the U pack is the battery pack voltage, I is the battery pack charging current, Is the battery pack at t 1 -t 2 The average voltage in some SOC intervals during the charging process;

[0093] Obtain the capacity standard deviation stdQ of the battery pack to be evaluated within the first SOC interval pack and capacity difference standard deviation stddQ pack , the acquisition process is as follows:

[0094] Obtain the voltage range [V pack_min ,V pack_max ] and the voltage range is V interval Equally spaced segments:

[0095] V pack_segment =[V pack_min ,V pack_min +V interval ,V pack_min +V interval *2,...,V pack_max ]

[0096] The capacity sequence of the battery pack to be evaluated within the voltage range is calculated by the ampere-hour integration method:

[0097] Q pack_segment(i) =[Q 2 -Q 1 ,Q 3 -Q 2 ,...,Qn -Q n-1 ],(i=1,2,...,N)

[0098] Among them, 1, 2, ..., n corresponds to the corresponding voltage points within the battery pack voltage range; 1, 2, ..., N corresponds to the number of cycles of the battery pack, Q pack_segment(i) is the capacity sequence of the battery pack under the i-th cycle;

[0099] Subtract the capacity sequence of the battery pack to be evaluated under different cycles from the capacity sequence under the first cycle to obtain the capacity difference sequence:

[0100] Q pack_segment_difference(i) =Q pack_segment(i) -Q pack_segment(1) ,(i=1,2,...,N)

[0101] Among them, Q pack_segment_difference(i) is the capacity difference sequence of the battery pack at the i-th cycle;

[0102] Obtain the capacity sequence Q of the battery pack to be evaluated under different cycles respectively pack_segment The standard deviation stdQ pack and the battery capacity difference sequence Q to be evaluated pack_segment_difference The standard deviation of pack ;

[0103] Obtain the battery pack voltage U at the time of charging mode switching during the charging process of the battery pack to be evaluated pack_changepoint ;

[0104] Obtain the battery pack voltage U after the battery pack to be evaluated is charged and left to stand for a certain period of time pack_rest ;

[0105] Get the voltage drop U of the battery pack within a certain time of instantaneous discharge when the battery pack to be evaluated starts discharging packdrop_5s .

[0106] At the same time, the second health indicator characterizing the inconsistency difference between the single cells is obtained by using a preset second health indicator extraction formula, including:

[0107] Extract the same features as the battery pack to be evaluated from each single cell in the battery pack to be evaluated Q cell(i) 、stddQ cell(i) With point feature U cell(i)_changepoint , U cell(i)_rest , U cell(i)drop_5s ;

[0108] The range, standard deviation and information entropy between the characteristics of each single cell are calculated to characterize the inconsistency differences between the single cell voltage, capacity and internal resistance. The expressions are as follows:

[0109] ran=x max -x min

[0110] Where, i = 1, 2, ..., n, n is the number of single cells in the battery pack; x max is the maximum value of the single cell characteristic quantity in the battery pack; x min It is the minimum value of the characteristic quantity of the single cell in the battery pack;

[0111]

[0112] Where n is the number of single cells connected in series in the battery pack; x i is the characteristic quantity of the i-th monomer; It is the average value of the characteristic quantities of all series-connected cells in the battery pack;

[0113]

[0114] Where n is the number of single cells connected in series in the battery pack; p(i) is the characteristic value corresponding to the i-th cell.

[0115] For example, in this embodiment, the health index is extracted based on the battery pack operation data of the 30%-70% SOC interval and the charge and discharge mode switching time point obtained in the above steps. In the example, the decline trend of the battery pack capacity is as follows: Figure 3 ,according to Figure 3 It can be seen that as the number of battery pack cycles increases, the battery pack capacity gradually decreases. Secondly, based on the external characteristics of the partial charge and discharge curves of the battery pack and the preset first health indicator extraction formula, a health indicator representing the overall performance decline trend of the battery pack is extracted. The trend of the health indicator evolving with the aging of the battery pack can be referred to Figures 4 to 9 , including Figure 4 Schematic diagram of the average voltage trend of the battery pack based on the number of cycles, Figure 5 Schematic diagram of the standard deviation trend of battery capacity series based on the number of cycles, Figure 6 Schematic diagram of the standard deviation trend of battery pack capacity difference series based on cycle number, Figure 7 Schematic diagram of battery pack voltage trend at variable current based on cycle number, Figure 8 Schematic diagram of battery pack voltage trend after full charge and rest based on cycle number Fig. 9A schematic diagram of the voltage drop trend of the battery pack within 5s of instantaneous discharge based on the number of cycles, and then obtaining a health indicator characterizing the inconsistency difference between the cells according to the range, standard deviation and information entropy of the corresponding characteristic values ​​of the battery cells in the battery pack and the preset second health indicator extraction formula. The extracted health indicators are shown in the following table;

[0116]

[0117] Step 103: Perform feature screening on the health indicators using a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated.

[0118] In this embodiment, this step mainly includes: calculating the Pearson correlation coefficient between the health indicator and the battery pack capacity of the battery pack to be evaluated, and filtering the health indicator according to the Pearson correlation coefficient to obtain an initial health indicator feature set; using a preset packaging method to perform a forward sequence search on the initial health indicator feature set, adding new features to the initial health indicator feature set, and obtaining the health indicator feature set.

[0119] Furthermore, in this embodiment, the calculating of the Pearson correlation coefficient between the health indicator and the battery pack capacity of the battery pack to be evaluated includes:

[0120] The Pearson correlation coefficient is calculated using a preset calculation formula, which is:

[0121]

[0122] Among them, x i and i is the true value of the health indicator and battery pack capacity, and is the corresponding mean of the two. The closer the absolute value of the result is to 1, the higher the correlation is.

[0123] Exemplarily, in this embodiment, based on the health indicators obtained in the above table, a filtering method is used, that is, the Pearson correlation coefficient between each health indicator in the above table and the battery pack capacity is calculated, and the health indicators are screened according to the Pearson correlation coefficient. Preferably, irrelevant health indicators with a Pearson correlation coefficient less than 0.2 with the battery pack capacity are filtered out, and at the same time, features with a Pearson correlation coefficient greater than 0.9 with the battery pack capacity are selected from the health indicators that characterize the overall performance degradation trend of the battery pack as the initial feature subsets of the packaging method, and then the packaging method is used to adopt a forward sequence search, and by continuously adding new features to the initial feature subset, the optimal feature subset that improves the model effect well is selected. The obtained optimal feature subset is shown in the following table;

[0124]

[0125]

[0126] Preferably, after obtaining the optimal feature subset, the quality of the optimal feature subset can be evaluated using a preset root mean square error expression, and the root mean square error expression is:

[0127]

[0128] Among them, y i is the true value of the battery capacity, y′ i Estimated value of battery pack capacity.

[0129] In this embodiment, when the difference between the root mean square error corresponding to the optimal feature subset in the subsequent round of search and the previous round of search according to the packaging method is less than one third of the minimum root mean square error in the previous round of search, the search is stopped, and the optimal feature subset in the current situation is selected by analyzing and comparing the N feature subsets with different initial features from two aspects: the number of features contained in the feature subsets and the corresponding minimum root mean square error.

[0130] Step 104: input the health indicator feature set into a preset battery pack health status assessment model, perform health status assessment on the battery pack to be assessed, and obtain an assessment result corresponding to the battery pack to be assessed.

[0131] In this embodiment, the battery pack health state estimation model is constructed in combination with the machine learning algorithm of support vector regression, and the battery pack health state estimation model is used to evaluate the health state of the battery pack to be evaluated according to the obtained optimal feature subset. After obtaining the battery pack health state estimation model, the optimal feature subset obtained by screening under the preset training set can be used as the model input to achieve a good evaluation of the health state of the lithium-ion battery pack throughout its life cycle. The estimation results of the test set under different training sets are as follows: Figure 5 As shown, the maximum error and root mean square error are shown in the following table:

[0132] Training set ratio Maximum relative error of the test set Test set root mean square error 20％ 1.60％ 0.77Ah 50％ 1.63％ 0.68Ah 80％ 1.41％ 0.67Ah

[0133] On the other hand, the present invention also discloses a battery pack health status assessment system. The specific structure of the assessment system can be referred to Figure 2 , mainly includes a data acquisition module 201, an indicator extraction module 202, a feature screening module 203 and a status evaluation module 204.

[0134] The data acquisition module 201 is used to obtain historical charge and discharge operation data of the battery pack to be evaluated, and obtain charge and discharge operation data within a first SOC interval from the historical charge and discharge operation data through a preset SOC interval threshold; the charge and discharge operation data includes the charging data of the battery pack within the first SOC interval and the battery pack operation data at the time point when the charge and discharge mode of the battery pack is switched within the first SOC interval.

[0135] The indicator extraction module 202 is used to obtain the health indicator corresponding to the battery pack to be evaluated through a preset health indicator extraction formula according to the charging and discharging operation data in the first SOC range.

[0136] The feature screening module 203 is used to perform feature screening on the health indicators by using a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated.

[0137] The state evaluation module 204 is used to input the health indicator feature set into a preset battery pack health state evaluation model, perform health state evaluation on the battery pack to be evaluated, and obtain an evaluation result corresponding to the battery pack to be evaluated.

[0138] In this embodiment, the data acquisition module 201 includes a historical data unit and a data extraction unit.

[0139] The historical data unit is used to obtain the discharge cut-off voltage and the charge cut-off voltage of the battery group to be evaluated, and obtain the historical charge and discharge operation data of the battery group to be evaluated according to the discharge cut-off voltage and the charge cut-off voltage.

[0140] The data extraction unit is used to extract the charging data of the battery group in the first SOC interval within the SOC interval threshold from the historical charging and discharging operation data corresponding to the battery group to be evaluated according to a preset SOC interval threshold; according to the SOC interval threshold, obtain the switching time of the current value in the historical charging and discharging operation data, and obtain the historical charging and discharging operation data corresponding to the switching time, and obtain the battery group operation data at the time point of switching the charging and discharging mode of the battery group in the first SOC interval.

[0141] In this embodiment, the indicator extraction module 202 includes a first indicator unit and a second indicator unit.

[0142] The first indicator unit is used to obtain a first health indicator characterizing the overall degradation trend of the battery pack to be evaluated through a preset first health indicator extraction formula based on the charging and discharging operation data in the first SOC interval.

[0143] The second indicator unit is used to obtain a second health indicator characterizing the inconsistency difference between the single cells according to the charge and discharge operation data corresponding to each single cell in the battery pack to be evaluated in the first SOC range through a preset second health indicator extraction formula.

[0144] The embodiment of the present invention provides a method and system for evaluating the health status of a battery pack. The method extracts health indicators that characterize the overall performance degradation trend of the battery pack and the inconsistency differences between cells in the battery pack through partial charge and discharge data, and realizes the stable screening of the optimal feature subset through a feature selection method that combines filtering method with packaging method. Thus, a model for estimating the health status of the battery pack is established in combination with a support vector regression algorithm, and an accurate evaluation of the health status of the lithium-ion battery pack over its entire life cycle is realized. This method can solve the problem of model overfitting caused by small amount of training data and redundancy between health indicators to a certain extent, and has strong adaptability to battery packs with different operating conditions and different grouping methods.

[0145] The specific embodiments described above further illustrate the purpose, technical solutions and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. It is particularly pointed out that for those skilled in the art, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating the health status of a battery pack, characterized in that: include: Acquire historical charge and discharge operation data of the battery pack to be evaluated, and acquire charge and discharge operation data in a first SOC interval from the historical charge and discharge operation data through a preset SOC interval threshold; the charge and discharge operation data includes charging data of the battery pack in the first SOC interval and battery pack operation data at a time point when the charge and discharge mode of the battery pack is switched in the first SOC interval; According to the charge and discharge operation data in the first SOC interval, the health index corresponding to the battery pack to be evaluated is obtained by a preset health index extraction formula; wherein, according to the charge and discharge operation data in the first SOC interval, the first health index characterizing the overall decline trend of the battery pack to be evaluated is obtained by a preset first health index extraction formula; according to the charge and discharge operation data corresponding to each single cell in the battery pack to be evaluated in the first SOC interval, the second health index characterizing the inconsistency difference between the single cells is obtained by a preset second health index extraction formula; The first health indicator characterizing the overall decline trend of the battery pack to be evaluated is obtained by using a preset first health indicator extraction formula, including: obtaining the average voltage of the battery pack to be evaluated in the first SOC interval, and the calculation formula of the average voltage is as follows: Among them, the U pack is the battery pack voltage, I pacK is the battery pack charging current, It is the average voltage of the battery pack in the partial SOC interval during the charging process from t1 to t2; Obtain the capacity standard deviation stdQ of the battery pack to be evaluated within the first SOC interval pack and capacity difference standard deviation stddQ pack , the acquisition process is as follows: Obtain the voltage range [V pack_min ,V pack_max ] and the voltage range is V interval Equally spaced segments: V pack_segment =[V pack_min ,V pack_min +V interval ,V pack_min +V interval *2,...,V pack_max ] The capacity sequence of the battery pack to be evaluated within the voltage range is calculated by the ampere-hour integration method: Q pack_segment(i) =[Q2-Q1,Q3-Q2,...,Q n -Q n-1 ],(i=1,2,...,N) Among them, 1, 2, ..., n corresponds to the corresponding voltage points within the battery pack voltage range; 1, 2, ..., N corresponds to the number of cycles of the battery pack, Q pack_segment(i) is the capacity sequence of the battery pack under the i-th cycle; Subtract the capacity sequence of the battery pack to be evaluated under different cycles from the capacity sequence under the first cycle to obtain the capacity difference sequence: Q pack_segment_difference(i) =Q pack_segment(i) -Q pack_segment(1) ,(i=1,2,...,N) Among them, Q pack_segment_difference(i) is the capacity difference sequence of the battery pack at the i-th cycle; Obtain the capacity sequence Q of the battery pack to be evaluated under different cycles respectively pack_segment The standard deviation stdQ pack and the battery capacity difference sequence Q to be evaluated pack_segment_difference The standard deviation of pack ; Obtain the battery pack voltage U at the time of charging mode switching during the charging process of the battery pack to be evaluated pack_changepoint ; Obtain the battery pack voltage U after the battery pack to be evaluated is charged and left to stand for a certain period of time pack_rest ; Get the voltage drop U of the battery pack within a certain time of instantaneous discharge when the battery pack to be evaluated starts discharging packdrop_5s ; The second health indicator extracting formula is used to obtain the second health indicator that characterizes the inconsistency difference between the single cells, including: extracting the same characteristics as the battery pack to be evaluated from each single cell in the battery pack to be evaluated. Q cell(i) 、stddQ cell(i) With point feature U cell(i)_changepoint , U cell(i)_rest , U cell(i)drop_5s ; The range, standard deviation and information entropy between the characteristics of each single cell are calculated to characterize the inconsistency differences between the single cell voltage, capacity and internal resistance. The expressions are as follows: ran=x max -x min Where, i = 1, 2, ..., n, n is the number of single cells in the battery pack; x max is the maximum value of the single cell characteristic quantity in the battery pack; x min It is the minimum value of the characteristic quantity of the single cell in the battery pack; Where n is the number of single cells connected in series in the battery pack; x i is the characteristic quantity of the i-th monomer; It is the average value of the characteristic quantities of all series-connected cells in the battery pack; Where n is the number of cells connected in series in the battery pack; p(i) is the characteristic value corresponding to the i-th cell; Perform feature screening on the health indicators by using a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated; The health indicator feature set is input into a preset battery pack health status assessment model, a health status assessment is performed on the battery pack to be assessed, and an assessment result corresponding to the battery pack to be assessed is obtained.

2. A method for evaluating the health status of a battery pack as claimed in claim 1, characterized in that: The acquiring of historical charge and discharge operation data of the battery pack to be evaluated, and acquiring charge and discharge operation data in a first SOC interval from the historical charge and discharge operation data by using a preset SOC interval threshold, includes: Obtaining a discharge cut-off voltage and a charge cut-off voltage of the battery pack to be evaluated, and obtaining historical charge and discharge operation data of the battery pack to be evaluated according to the discharge cut-off voltage and the charge cut-off voltage; Extracting charging data of the battery pack in the first SOC interval within the SOC interval threshold from the historical charging and discharging operation data corresponding to the battery pack to be evaluated according to a preset SOC interval threshold; According to the SOC interval threshold, the switching time of the current value in the historical charge and discharge operation data is obtained, and the historical charge and discharge operation data corresponding to the switching time is obtained to obtain the battery pack operation data at the time point of the battery pack charge and discharge mode switching in the first SOC interval.

3. A method for evaluating the health status of a battery pack as claimed in claim 1, characterized in that: The method of screening the health indicators by a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated includes: Calculating a Pearson correlation coefficient between the health indicator and the battery pack capacity of the battery pack to be evaluated, and filtering the health indicator according to the Pearson correlation coefficient to obtain an initial health indicator feature set; The health indicator feature set is obtained by performing a forward sequence search on the initial health indicator feature set using a preset packaging method and adding new features to the initial health indicator feature set.

4. A method for evaluating the health status of a battery pack as claimed in claim 3, characterized in that: The calculating the Pearson correlation coefficient between the health indicator and the battery pack capacity of the battery pack to be evaluated includes: The Pearson correlation coefficient is calculated using a preset calculation formula, which is: Among them, x i and i is the true value of the health indicator and battery pack capacity, and is the corresponding mean of the two. The closer the absolute value of the result is to 1, the higher the correlation is.

5. A battery pack health status assessment system, characterized in that: It includes data acquisition module, indicator extraction module, feature screening module and status assessment module; The data acquisition module is used to acquire historical charge and discharge operation data of the battery pack to be evaluated, and acquire charge and discharge operation data within a first SOC interval from the historical charge and discharge operation data through a preset SOC interval threshold; the charge and discharge operation data includes the charging data of the battery pack within the first SOC interval and the battery pack operation data at the time point of switching the charge and discharge mode of the battery pack within the first SOC interval; The indicator extraction module is used to obtain the health indicator corresponding to the battery pack to be evaluated through a preset health indicator extraction formula according to the charge and discharge operation data in the first SOC interval; wherein the indicator extraction module includes a first indicator unit and a second indicator unit; the first indicator unit is used to obtain the first health indicator characterizing the overall decline trend of the battery pack to be evaluated through a preset first health indicator extraction formula according to the charge and discharge operation data in the first SOC interval; the second indicator unit is used to obtain the second health indicator characterizing the inconsistency difference between the single cells according to the charge and discharge operation data corresponding to each single cell in the battery pack to be evaluated in the first SOC interval through a preset second health indicator extraction formula; The first health indicator characterizing the overall decline trend of the battery pack to be evaluated is obtained by using a preset first health indicator extraction formula, including: obtaining the average voltage of the battery pack to be evaluated in the first SOC interval, and the calculation formula of the average voltage is as follows: Among them, the U pack is the battery pack voltage, I pacK is the battery pack charging current, It is the average voltage of the battery pack in the partial SOC interval during the charging process from t1 to t2; Obtain the capacity standard deviation stdQ of the battery pack to be evaluated within the first SOC interval pack and capacity difference standard deviation stddQ pack , the acquisition process is as follows: Obtain the voltage range [V pack_min ,V pack_max ] and the voltage range is V interval Equally spaced segments: V pack_segment =[V pack_min ,V pack_min +V interval ,V pack_min +V interval *2,...,V pack_max ] The capacity sequence of the battery pack to be evaluated within the voltage range is calculated by the ampere-hour integration method: Q pack_segment(i) =[Q2-Q1,Q3-Q2,...,Q n -Q n-1 ],(i=1,2,...,N) Among them, 1, 2, ..., n corresponds to the corresponding voltage points within the battery pack voltage range; 1, 2, ..., N corresponds to the number of cycles of the battery pack, Q pack_segment(i) is the capacity sequence of the battery pack under the i-th cycle; Subtract the capacity sequence of the battery pack to be evaluated under different cycles from the capacity sequence under the first cycle to obtain the capacity difference sequence: Q pack_segment_difference(i) =Q pack_segment(i) -Q pack_segment(1) ,(i=1,2,...,N) Among them, Q pack_segment_difference(i) is the capacity difference sequence of the battery pack at the i-th cycle; Obtain the capacity sequence Q of the battery pack to be evaluated under different cycles respectively pack_segment The standard deviation stdQ pack and the battery capacity difference sequence Q to be evaluated pack_segment_difference The standard deviation of pack ; Obtain the battery pack voltage U at the time of charging mode switching during the charging process of the battery pack to be evaluated pack_changepoint ; Obtain the battery pack voltage U after the battery pack to be evaluated is charged and left to stand for a certain period of time pack_rest ; Get the voltage drop U of the battery pack within a certain time of instantaneous discharge when the battery pack to be evaluated starts discharging packdrop_5s ; The second health indicator extracting formula is used to obtain the second health indicator that characterizes the inconsistency difference between the single cells, including: extracting the same characteristics as the battery pack to be evaluated from each single cell in the battery pack to be evaluated. Q cell(i) 、stddQ cell(i) With point feature U cell(i)_changepoint , U cell(i)_rest , U cell(i)drop_5s ; The range, standard deviation and information entropy between the characteristics of each single cell are calculated to characterize the inconsistency differences between the single cell voltage, capacity and internal resistance. The expressions are as follows: ran=x max -x min Where, i = 1, 2, ..., n, n is the number of single cells in the battery pack; x max is the maximum value of the single cell characteristic quantity in the battery pack; x min It is the minimum value of the characteristic quantity of the single cell in the battery pack; Where n is the number of single cells connected in series in the battery pack; x i is the characteristic quantity of the i-th monomer; It is the average value of the characteristic quantities of all series-connected cells in the battery pack; Where n is the number of cells connected in series in the battery pack; p(i) is the characteristic value corresponding to the i-th cell; The feature screening module is used to perform feature screening on the health indicators by a preset feature selection method to obtain a health indicator feature set corresponding to the battery pack to be evaluated; The state evaluation module is used to input the health indicator feature set into a preset battery pack health state evaluation model, perform health state evaluation on the battery pack to be evaluated, and obtain an evaluation result corresponding to the battery pack to be evaluated.

6. A battery pack health status assessment system as claimed in claim 5, characterized in that: The data acquisition module includes a historical data unit and a data extraction unit; The historical data unit is used to obtain the discharge cut-off voltage and the charge cut-off voltage of the battery pack to be evaluated, and obtain the historical charge and discharge operation data of the battery pack to be evaluated according to the discharge cut-off voltage and the charge cut-off voltage; The data extraction unit is used to extract the charging data of the battery group in the first SOC interval within the SOC interval threshold from the historical charging and discharging operation data corresponding to the battery group to be evaluated according to a preset SOC interval threshold; according to the SOC interval threshold, obtain the switching time of the current value in the historical charging and discharging operation data, and obtain the historical charging and discharging operation data corresponding to the switching time, and obtain the battery group operation data at the time point of switching the charging and discharging mode of the battery group in the first SOC interval.

Images