Lithium battery aging state calculation method and device, electronic equipment and storage medium

Abstract

The application discloses a lithium battery aging state calculation method and device, electronic equipment and storage medium. First, based on the offline running data of the reference battery, an offline database and a corresponding offline model are established. Then, the offline model is used to process the historical running data of the target battery. The obtained historical health factor sequence and corresponding historical aging state label are fused with the offline database. Based on the regression analysis algorithm, an online model is established according to the fused database. Finally, the online aging state label corresponding to the online running data is calculated by using the online model. By selecting a battery with the same material system and the same capacity as the target battery as the reference battery, the offline model can reflect the aging characteristics of the target battery to a certain extent. By using the fused database to establish the online model, the applicability of the online model to the target battery is improved, and the accuracy of the calculated online aging state label is improved.

Description

Technical Field

[0001] This application relates to the field of lithium battery technology, and more specifically, to a method, apparatus, electronic device, and storage medium for calculating the aging state of a lithium battery. Background Technology

[0002] To ensure the safe, efficient, and stable operation of pure electric vehicles and hybrid vehicles, it is often necessary to predict the aging state of the lithium batteries that serve as the power source and calculate the corresponding aging state labels.

[0003] Currently, most methods utilize offline regression models built upon offline health factor sequences and offline aging state labels of batteries other than the target battery. These models aim to calculate the current aging state label of the target battery using online data to predict its aging state. However, in calculating the aging state label of the target battery using offline regression models based on offline data from other batteries, the differences between batteries—including differences in material systems, factory capacity, usage scenarios, and operating conditions—lead to variations in aging paths. This results in low applicability of the offline regression models to the target battery, and consequently, low accuracy in calculating the aging state label. Summary of the Invention

[0004] In view of the above problems, this application proposes a method, apparatus, electronic device, and storage medium for calculating the aging state of lithium batteries, so as to solve the problem of low accuracy in calculating the online aging state label of the target lithium battery when the battery aging path is diverse. The specific solution is as follows:

[0005] Firstly, a method for calculating the aging state of lithium batteries is provided, including:

[0006] An offline database is established based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery. The offline database includes multiple sets of offline health factor sequences and corresponding offline aging state labels.

[0007] An offline model is established based on the offline database and a regression analysis algorithm.

[0008] Based on the offline database, the offline model, and the historical operating data of the target battery, an online database for the target battery is established. The online database contains a sequence of historical health factors and corresponding historical aging status labels determined based on the historical operating data.

[0009] An online model is established based on the online database and a regression analysis algorithm.

[0010] Based on the online operating data of the target battery, extract the online health factor sequence;

[0011] Based on the online model, calculate the online aging status label corresponding to the online health factor sequence.

[0012] Secondly, a lithium battery aging state calculation device is provided, comprising:

[0013] An offline database establishment unit is used to establish an offline database based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery. The offline database includes multiple sets of offline health factor sequences and corresponding offline aging state labels.

[0014] An offline model building unit is used to build an offline model based on the offline database using a regression analysis algorithm.

[0015] An online model building unit is used to build an online database for the target battery based on the offline database, the offline model, and the historical operating data of the target battery. The online database contains historical health factor sequences and corresponding historical aging state labels determined based on the historical operating data. An online model is built based on the online database using a regression analysis algorithm.

[0016] An online data extraction unit is used to extract online health factor sequences based on the online operating data of the target battery;

[0017] An online aging status label determination unit is used to calculate an online aging status label corresponding to the online health factor sequence based on the online model.

[0018] Thirdly, an electronic device is provided, comprising: a memory and a processor;

[0019] The memory is used to store programs;

[0020] The processor is used to execute the program to implement the steps of the lithium battery aging state calculation method as described above.

[0021] Fourthly, a storage medium is provided on which a computer program is stored, which, when executed by a processor, implements the various steps of the lithium battery aging state calculation method as described above.

[0022] Using the above technical solution, this application constructs an offline database by extracting offline operating data from reference batteries with the same material system and capacity as the target battery. An offline model is then established based on this database, connecting offline health factor sequences and offline aging state labels. Based on this offline model, historical operating data preceding the online operating data of the target battery are processed, and the processing results are fused with the offline database to form an online database. An online model is then established based on this online database. The online aging state labels corresponding to the online operating data are calculated using the online model. Since the offline model utilizes offline operating data from reference batteries with the same material system and capacity as the target battery, it can reflect the commonalities in aging among lithium batteries with the same material system and capacity. Calculating the historical operating data of the target battery based on the offline model yields more accurate historical aging state labels. Therefore, the online model, built upon an online database containing historical health factor sequences and corresponding historical aging state labels, better reflects the aging characteristics of the target battery than the offline model, making it more suitable for the target battery and improving the accuracy of the calculated online aging state labels corresponding to the online operating data. Attached Figure Description

[0023] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this application. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0024] Figure 1 This is a flowchart illustrating a method for calculating the aging state of a lithium battery according to an exemplary embodiment;

[0025] Figure 2 This is a schematic diagram illustrating an offline database and a corresponding offline regression model according to an exemplary embodiment;

[0026] Figure 3 This is a flowchart illustrating a method for calculating the aging state of a lithium battery according to another exemplary embodiment;

[0027] Figure 4 This is a flowchart illustrating a method for calculating the aging state of a lithium battery according to yet another exemplary embodiment;

[0028] Figure 5 This is a schematic diagram illustrating the process of building an online model based on an exemplary embodiment;

[0029] Figure 6 This is a schematic diagram illustrating the extraction process of a health factor sequence according to an exemplary embodiment;

[0030] Figure 7 This is a schematic diagram of the structure of a lithium battery aging status tag calculation device according to an exemplary embodiment;

[0031] Figure 8 This is a hardware structure block diagram of an electronic device according to an exemplary embodiment. Detailed Implementation

[0032] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0033] Given the poor applicability of offline models to target batteries, the inventors discovered that online models based on the online operating data of the target battery can reflect its aging characteristics. However, data from the early stages of aging is unavailable, making it impossible to calculate the aging status label for the entire battery lifecycle. To address this issue, this application provides a scheme for calculating the aging status label of lithium batteries, applicable to online aging status label calculation tasks, including calculations for various stages of the lithium battery lifecycle, including the early stages of aging.

[0034] The proposed solution can be implemented based on a terminal with data processing capabilities, such as a mobile phone, computer, server, or cloud platform.

[0035] The following explains the basic concepts involved in the embodiments of this application, including battery operating data, health factor sequences, aging state labels, and regression models:

[0036] Battery operating data includes offline operating data of a reference battery, historical operating data of the target battery, and online operating data. For any given data point, the data can be represented as a window curve, which is collected during a single charge cycle. Depending on the data type, window curves can be categorized as voltage, current, and temperature window curves, etc. The following explanation will use a voltage window curve as an example.

[0037] Health factor sequences include offline health factor sequences, historical health factor sequences, and online health factor sequences. The extraction method for any health factor sequence can be to extract aging characteristic parameters based on a single operating data point of a lithium battery. A health factor sequence is composed of multiple aging characteristic parameters. Aging characteristic parameters, i.e. health factors, can include statistical parameters such as window time, window average voltage, window curve skewness, and window curve kurtosis.

[0038] The aging status label can be the State of Health (SOH) value, which characterizes the health status of a lithium battery. It can be the ratio of the battery's usable capacity to its rated capacity, expressed as a percentage.

[0039] A regression model is a regression analysis algorithm that can be used to construct a regression model between a sequence of health factors and an aging state label using a multiple linear regression algorithm.

[0040] Next, combined Figure 1 The illustrations detail the steps of the lithium battery aging state calculation method provided in this application embodiment. The lithium battery aging state calculation method provided in this application embodiment specifically includes the following steps:

[0041] S101. Establish an offline database based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery.

[0042] The target battery is the lithium battery whose aging status label is to be calculated. The offline operating data is the operating data of the reference battery extracted from the car manufacturer's database. The offline database includes multiple sets of offline health factor sequences and corresponding offline aging status labels.

[0043] The automotive manufacturer's database contains a large amount of battery operation data, allowing for the extraction of offline operation data from reference batteries that share the same material system, capacity, and even production batch as the target battery. Furthermore, it allows for the selection of voltage window curves that meet specific requirements, such as those corresponding to specific voltage ranges. In addition, the database includes aging status labels for reference batteries derived from big data analysis, which correspond to the voltage window curves. Therefore, it is possible to extract the required offline aging status labels and corresponding operation data, such as offline aging status labels with specific values. From this operation data, various aging characteristic parameters can be extracted to construct an offline health factor sequence. The number and type of aging characteristic parameters can be defined according to actual needs.

[0044] S102. Establish an offline model based on the offline database using a regression analysis algorithm.

[0045] Figure 2A schematic diagram illustrating a possible method for building an offline model based on an offline database is shown. For example... Figure 2 As shown, using regression analysis algorithm f, the offline health factor matrix h can be used for analysis. f，off and offline aging state label vector H label，off In this case, output the offline model F between the two. off =f(h) f，off H label，off ).

[0046] S103. Based on the offline database, the offline model, and the historical operating data of the target battery, establish an online database for the target battery.

[0047] The historical operating data of the target battery can be multiple operating data points that are distinct from the online operating data. These data points were collected before the online operating data was collected and have different collection times. The online operating data is a single operating data point that corresponds to the online aging state label to be calculated. The online database contains a sequence of historical health factors and corresponding historical aging state labels determined based on the historical operating data.

[0048] It should be noted that during the establishment of the online database, historical operational data was processed, including extracting historical health factor sequences and calculating historical aging state labels corresponding to these sequences based on an offline model. This yielded a portion of the data information that makes up the online database, while the offline database provided another portion. Since each historical operational data point can be a voltage window curve collected during a single charge of the target battery, and the extracted health factors are statistical parameters of the window curve, the corresponding historical health factor sequences can be obtained from the historical operational data.

[0049] S104. Establish an online model based on the regression analysis algorithm using the online database.

[0050] S105. Extract the online health factor sequence based on the online operating data of the target battery.

[0051] S106. Based on the online model, calculate the online aging status label corresponding to the online health factor sequence.

[0052] The lithium battery aging state calculation method provided in this application involves extracting offline operating data from reference batteries that belong to the same material system and have the same capacity as the target battery, constructing an offline database, and establishing an offline model between offline health factor sequences and offline aging state labels based on the offline database. Based on the offline model, historical operating data preceding the online operating data of the target battery are processed, and the processing results are fused with the offline database to form an online database. An online model is then established based on the online database. Finally, the online aging state labels corresponding to the online operating data are calculated using the online model.

[0053] Since the offline model is built using offline operating data from a reference battery with the same material system and capacity as the target battery, it can reflect the common aging characteristics of lithium batteries with the same material system and capacity. Calculating the historical operating data of the target battery based on the offline model can yield more accurate historical aging state labels. Therefore, the online model built based on an online database containing historical health factor sequences and corresponding historical aging state labels is more effective than the offline model in reflecting the aging characteristics of the target battery, making it more suitable for the target battery and improving the accuracy of the calculated online aging state labels corresponding to the online operating data.

[0054] The following is an illustrative explanation of the steps involved in calculating the aging state of a lithium battery. First, the steps for establishing an offline database will be explained.

[0055] In some embodiments provided in this application, the implementation process of step S101 above, establishing an offline database based on offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery, may include:

[0056] S201. Extract the offline aging status label and corresponding operation data of the reference battery from the offline operation data.

[0057] The reference battery can be a battery of the same model and batch as the target battery. Lithium batteries of the same model and batch have smaller differences, including differences in material systems, factory capacity, etc. Therefore, the aging path of such reference batteries is closer to that of the target battery. An offline model built based on the offline operating data of such reference batteries is more suitable for the target battery than models based on other reference batteries.

[0058] S202. Extract offline health factor sequences based on the operational data.

[0059] S203. Construct an offline database based on offline health factor sequences and corresponding offline aging status labels.

[0060] The offline database may include M sets of offline aging status labels and corresponding offline health factor sequences, and each offline health factor sequence may include N offline health factors. Figure 2 A schematic diagram of the possible composition of an offline database is shown, such as... Figure 2 As shown, the offline database may include: an M×N offline health factor matrix and an M×1 offline aging state label vector, wherein each row of the offline health factor matrix corresponds to an offline health factor sequence, and each offline health factor sequence corresponds to an offline aging state label located in the same row of the offline aging state label vector.

[0061] In one possible implementation, step S201 above, the process of extracting the offline aging status tag of the reference battery and the corresponding operating data from the offline operating data, may include:

[0062] In the offline operating data, starting from the initial offline aging state label that represents the reference battery not yet aging, for every increase of X% / M in the change of the offline aging state label, an offline aging state label and the corresponding operating data are extracted.

[0063] The change amount is used to characterize the degree of aging of the reference battery compared to when it had not yet started aging; the initial offline aging status label can be set to 100% to characterize that the battery has not aged; X% is the length of the range of aging status labels available for the target battery, and M is the number of offline aging status labels extracted from the offline operation data. For every increase of X% / M in the change amount of the offline aging status labels compared to the initial value, it is equivalent to every decrease of X% / M in the offline aging status labels, which means that the degree of aging of the reference battery has increased by X% / M.

[0064] In current industry standards, the usable aging status label range for automotive power lithium batteries is generally 100%–70%, meaning the usable aging status label range is 30%. When the aging status label of a lithium battery decays to below 70%, the battery is no longer suitable for automotive power supply. In this case, the step size for extracting the aging status label is 30% / M, meaning that the change in aging status label every 30% / M is used as the condition for extracting operational data.

[0065] The above steps enable the uniform extraction of aging status tags and corresponding operational data within the available aging status tags of the battery. The above steps only provide one possible method for extracting aging status tags. In application, other methods can be adopted according to the actual situation. For example, the number of extracted aging status tags can be appropriately reduced during periods of slow change, while the number can be appropriately increased during periods of rapid change.

[0066] In the specific application scenario of vehicle-mounted power supply, focusing on aging status tags within a certain range, especially within the available range, is more practically meaningful. However, this does not mean that this calculation method is not applicable to calculating aging status tags outside the available aging tag range. For the calculation of aging status tags outside the available range, this method can still obtain relatively accurate calculation results under certain conditions. These conditions may include adding a certain amount of relevant data, such as offline operation data outside the available range.

[0067] Secondly, the process of establishing an online database and online model based on offline databases, offline models, and historical operational data is illustrated by example.

[0068] In cases where the historical operating data consists of multiple operating data points distinct from the online operating data, the process of establishing an online database for the target battery based on the offline database, the offline model, and the historical operating data of the target battery in step S103 will be explained.

[0069] In some embodiments provided in this application, the process may include:

[0070] Based on each piece of historical operational data, the corresponding historical health factor sequence is extracted, and the historical aging status label corresponding to the historical health factor sequence is calculated based on the offline model.

[0071] Add the historical health factor sequence and the corresponding historical aging status label to the offline database to obtain the updated offline database;

[0072] In the updated offline database, the historical aging status tags and offline aging status tags are matched, and the matched offline aging status tags and corresponding offline health factor sequences are deleted from the updated offline database to obtain the online database.

[0073] The above steps, through offline model calculations, obtained the historical aging state labels of the target battery. Therefore, in establishing the online database, the historical health factor sequences and corresponding historical aging state labels of the target battery at each stage, as well as a portion of the offline database, were utilized. The online model established using the online database obtained in the above manner not only reflects the commonalities of aging among batteries belonging to the same material system and with the same capacity, but also reflects the individual aging characteristics of the target battery to a certain extent. It also avoids the problem of missing data in the early stages of aging. Therefore, calculations based on this online model can obtain relatively accurate online aging state labels for each stage of the target battery.

[0074] Figure 3This is a flowchart illustrating a method for calculating the aging state of a lithium battery according to another embodiment. It should be noted that the historical operating data used in this embodiment can be multiple operating data points collected before the online operating data, arranged in chronological order of collection time. For example... Figure 3 As shown, a method for calculating the aging state of a lithium battery may include the following steps:

[0075] S301-S302 correspond one-to-one with the steps S101-S102 mentioned above, and will not be repeated here.

[0076] S303. Obtain the nth historical running data.

[0077] Where n is a positive integer representing the time sequence in which historical operation data was collected, with the initial time n = 1.

[0078] S304. Extract the historical health factor sequence corresponding to the nth historical operational data.

[0079] S305. Calculate the corresponding historical aging status labels based on the online model.

[0080] The initial online model was actually an offline model.

[0081] S306. Calculate the change in the historical aging status label compared to other historical aging status labels that have been obtained.

[0082] Other historical aging status labels can be calculated based on the historical aging status labels of each running data before the nth historical running data.

[0083] S307. Determine whether the change amount meets the preset conditions. If the preset conditions are not met, proceed to step S310. If the preset conditions are met, proceed to step S308.

[0084] S308. Update the online database.

[0085] The process of updating the online database includes adding the historical health factor sequence and the corresponding historical aging status label to the online database, matching the historical aging status label and the offline aging status label in the online database, and deleting the matched offline aging status label and the corresponding offline health factor sequence from the online database to obtain the updated online database. The initial online database can be an offline database.

[0086] S309. Update the online model based on the updated online database using a regression analysis algorithm.

[0087] S310. Determine whether all historical running data has been processed. If yes, proceed to step S312; otherwise, proceed to step S311.

[0088] S311. Let n = n + 1.

[0089] S312-S313 correspond one-to-one with the steps S105-S106 mentioned above, and will not be repeated here.

[0090] The above steps utilize an online model to calculate historical aging state labels corresponding to each historical health factor. Under preset conditions, historical data of the target battery and offline data of the reference battery are fused, and the online model is continuously updated using the fused database. In the early stages of the target battery's aging process, since batteries with the same material system and capacity exhibit very similar aging characteristics, the offline model can be used as the initial online model to calculate a more accurate aging state label for the target battery, solving the problem of missing aging state labels during the online model establishment process. Furthermore, as the target battery's performance deteriorates and its aging degree deepens, the aging characteristics of the target battery are fully considered. Specifically, during the processing of historical operating data, offline data from the reference battery and historical data from the target battery are fused in stages. The online regression model, built based on more operating data from the target battery, is continuously updated when preset conditions are met. The updated online model is then used to process subsequent historical operating data, further improving the online model's adaptability to the target battery. Therefore, the model can more accurately reflect the aging path of the target battery, and a more accurate aging state label corresponding to the online operating data can be obtained based on this model. In addition, since updating the model requires a lot of computation, a preset condition was set, namely, the interval between online database and online model updates, which to some extent solved the problem of the surge in computation caused by frequent online model updates.

[0091] Figure 4 This is a flowchart illustrating a method for calculating the aging state of a lithium battery according to another embodiment. It should be noted that the historical operating data used in this embodiment can be multiple operating data points collected before the online operating data, arranged in chronological order of collection time. For example... Figure 4 As shown, a method for calculating the aging state of a lithium battery may include the following steps:

[0092] S401-S405 correspond one-to-one with the steps S301-S305 mentioned above, and will not be repeated here.

[0093] S406. Calculate the change in the historical aging status label compared to the previous calculation, and include it in the cumulative change.

[0094] S407. Determine whether the change satisfies the first condition.

[0095] Wherein, the first condition is that the change amount reaches the change amount threshold. If the first condition is not met, then step S409 is executed; otherwise, step S408 is executed.

[0096] S408. Add historical health factor sequences and corresponding historical aging status labels to the online database.

[0097] The initial online database was an offline database.

[0098] S409. Determine whether the cumulative change satisfies the second condition.

[0099] The second condition is that the cumulative change reaches the cumulative change threshold. If the second condition is not met, step S413 is executed; if the second condition is met, steps S410 and S412 are executed.

[0100] The cumulative change can be the sum of the changes between two adjacent aging state labels, starting from zero after the previous online model update.

[0101] S410, Update the online database.

[0102] The process of updating the online database includes matching historical aging status labels in the online database with offline aging status labels, deleting the matched offline aging status labels and corresponding offline health factor sequences from the online database to obtain an updated online database, and using the updated online database to update the online model based on a regression analysis algorithm.

[0103] S411. Update the online model based on the updated online database using a regression analysis algorithm.

[0104] S412. The cumulative change is cleared to zero.

[0105] S412 can be a step executed before step S413 is executed if the result of S409 is yes.

[0106] S413. Determine whether all historical running data has been processed. If yes, proceed to step S415; otherwise, proceed to step S414.

[0107] S414. Let n = n + 1.

[0108] Steps S415-S416 correspond one-to-one with steps S312-S313 mentioned above, and will not be repeated here.

[0109] In the above steps, two preset conditions—that the change amount reaches a change threshold and the cumulative change amount reaches a cumulative change threshold—separate the updating of the online database by adding entries and deleting entries. That is, under certain conditions, a record-by-record add-and-delete update operation is performed on the online database. The update operation on the online database can also be implemented as follows: when adding a historical aging status tag, check if the online database contains an offline aging status tag that is the same as the historical aging status tag. If the search result is yes, then delete the found offline aging status tag and its corresponding offline health factor sequence. That is, under another condition, a record-by-record add-and-delete update operation is performed on the online database.

[0110] In one possible implementation, the change threshold is X% / M, where X% is the length of the range of aging state tags available for the target battery, and M is the number of offline aging state tags extracted from the offline operating data.

[0111] In one possible implementation, the cumulative change threshold is N*X% / M, where N is a preset positive integer, X% is the length of the range of aging state tags available for the target battery, and M is the number of offline aging state tags extracted from the offline operating data.

[0112] The above only provides one possible value for the change threshold and the cumulative change threshold. In application, these values ​​can be set according to actual conditions. For example, if the range of usable aging state tags for the target battery is set to 30%, and the number of offline aging state tags extracted from offline operating data is 60, then the change threshold is 30% / 60 = 0.5%. Based on the above, if N = 10, then the cumulative change threshold is 10 * 30% / 60 = 5%. Different thresholds can also be set for different aging stages of the target battery. For example, if aging is slow, the thresholds can be increased appropriately, and vice versa. The change threshold represents the interval for updating the online database, and the cumulative change threshold represents the interval for updating the online model. The two intervals can have a certain correlation, as described above, or can be any value set by the user.

[0113] For example, Figure 5 This illustrates a possible online modeling process for the early stages of target battery aging, where the online database consists of an M×N fusion health factor matrix h. f，on +h f，off and an M×1 aging state label vector H label，on +H label，off , where H label，on It is based on the offline model Foff The calculated health factor matrix has each row corresponding to a health factor sequence, and each health factor sequence corresponds to an aging state label in the same row of the aging state label vector. Based on this, a regression analysis algorithm f can be used to input the fused health factor matrix h. f，on +h f，off and fusion aging state label vector H label，on +H label，off In this case, output the online model F between the two. on =f(h) f，on +h f，off H label，on +H label，off ).

[0114] In the above steps, if the first condition is met, the online database is updated by adding, where the first condition is to set a threshold for the change between two adjacent historical aging state labels; if the second condition is met, preparation is made to update the online model, that is, the online database is updated by deleting, and the online model is updated, where the second condition is to set a threshold for the cumulative amount of the change under certain conditions, wherein the cumulative value of the change includes the change.

[0115] By setting the first condition, the online database is prevented from being arbitrarily updated in a short period of time. By setting the second condition, the threshold for the cumulative change should not be less than the threshold for the change under reasonable settings, allowing for a larger interval to be set between updating the online model and updating the online database. Therefore, the above steps utilize the historical operating data of the target battery to a certain extent, and also solve the problem of increased computational load caused by frequent model updates. Furthermore, since the aging state label will not change abruptly in a short period of time within the available range, updating the online model at certain intervals also avoids the problems of decreased model accuracy and decreased accuracy of aging state label calculation caused by irregular or no model updates. Moreover, the above steps are equivalent to replacing the corresponding content in the offline database with the processing results of the historical operating data, so that the online database no longer contains inaccurate offline health factor sequences and corresponding offline aging state labels relative to the target battery, and contains more processing results of the historical operating data of the target battery than the offline database. Therefore, the online model established based on the above online database can better reflect the aging characteristics of the target battery, and more accurate online aging state labels can be calculated based on this online model.

[0116] Based on the above, to take into account the differences in usage habits of different users, which are mainly reflected in the fact that the voltage change range of the battery is not the same when different users charge the battery, multiple voltage ranges are preset so that at least one voltage range can capture the voltage change corresponding to the online operation data. In other words, when calculating based on the online operation data, at least one voltage range can extract the online health factor sequence corresponding to the online operation, and there is an online model corresponding to that voltage range.

[0117] The reason for setting multiple voltage ranges is that if only a large voltage range, or even the entire voltage range, is set, although the voltage changes during the user's charging process can be captured within that range, differences in user habits may mean that the voltage changes only involve a portion of that range. Furthermore, since the health factors extracted from the voltage window are statistical values ​​of the voltage curve shown in that window, the health factors obtained within that range cannot accurately reflect the true voltage changes. Conversely, if only a small voltage range is set, differences in user charging habits may prevent the capture of voltage changes during the user's charging process, meaning that the online health factor sequence cannot be extracted, and therefore the online aging status label cannot be calculated. Therefore, by setting multiple voltage ranges of appropriate length, the inaccuracy or failure to extract health factor sequences can be avoided to some extent.

[0118] It should be noted that the voltage range setting, including the range and number of voltage ranges, can be based on a fixed range and a fixed number obtained from big data analysis of the car manufacturer's database, with appropriate adjustments made to account for differences in battery type. Alternatively, it can be based on the operating data of the target battery to determine the range and number of voltage ranges that are more suitable for the target battery.

[0119] This application provides a method for calculating the aging state of lithium batteries under multiple preset voltage ranges. The following will exemplarily describe the process of establishing an offline database and offline model, establishing an online database and online model, and calculating online aging state labels based on the online model. Some of these steps can be referred to the above description.

[0120] In some embodiments provided in this application, the above-described step S101, the process of establishing an offline database based on offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery, may include:

[0121] Based on the offline operation data, establish offline databases corresponding to different preset voltage ranges.

[0122] Each offline database includes multiple sets of offline health factor sequences and corresponding offline aging status labels obtained within the corresponding voltage range.

[0123] Figure 6 A schematic diagram illustrating a possible process for extracting health factor sequences in a multi-voltage range scenario is shown. Figure 6 As shown, under the preset voltage ranges a, b, and c, and with each health factor sequence containing N health factors, for an offline aging state label H... label，off It is possible to extract three health factor sequences corresponding to different preset voltage ranges, as shown below, a=[a1 a2 … a N ], b = [b1 b2 … b N ] and c = [c1 c2 … c N ].

[0124] In one possible implementation, step S102 above, the process of establishing an offline model based on the offline database using a regression analysis algorithm, may include:

[0125] Based on the offline database corresponding to each voltage range, an offline model corresponding to each voltage range is established using a regression analysis algorithm.

[0126] In one possible implementation, the online database includes multiple online databases corresponding to different preset voltage ranges, and each online database contains multiple sets of historical health factor sequences and corresponding historical aging state tags obtained within the corresponding voltage range.

[0127] Step S103 above, the process of establishing an online database for the target battery based on the offline database, the offline model, and the historical operating data of the target battery, may include:

[0128] In the process of processing each piece of historical operational data, a voltage range is determined based on the historical operational data. Within this voltage range, the historical health factor sequence is extracted, and the historical aging state label is calculated based on the offline model corresponding to the voltage range. If the calculation involves the amount of change, the comparison object corresponding to the historical aging state label, i.e., other historical aging state labels, is also corresponding to the voltage range. If the database is updated, the updated database is also corresponding to the voltage range.

[0129] It should be noted that if a single historical data point can identify multiple voltage ranges, the above processing needs to be performed separately for each voltage range. If subsequent model updates are involved, the online model should be updated according to the updated database corresponding to each voltage range across all preset voltage ranges.

[0130] In the above steps, the processing of each historical operating data is performed within its corresponding voltage range. That is, the online database corresponding to that voltage range is updated using the health factor sequence and corresponding aging state labels obtained based on the voltage range. This ensures that the online model used subsequently, corresponding to a specific voltage range, better matches the aging state of the target battery and the online operating data. The specific voltage range is determined based on the online operating data.

[0131] In one possible implementation, step S104 above, the process of establishing an online model based on the online database using a regression analysis algorithm, may include:

[0132] Based on the online database corresponding to each voltage range, an online model corresponding to each voltage range is established using a regression analysis algorithm.

[0133] In one possible implementation, step S105 above, the process of extracting the online health factor sequence based on the online operating data of the target battery, may include:

[0134] Based on the online operation data, the target voltage range corresponding to it is determined, the online health factor sequence is extracted within the target voltage range, and the online model corresponding to the target voltage range is obtained.

[0135] The lithium battery aging state calculation method provided in this application, under multiple preset voltage ranges, extracts offline and historical health factor sequences and corresponding offline and historical aging state labels for building online models within multiple voltage ranges, and establishes multiple online models corresponding to each voltage range. This method can effectively match the usage habits of different users. The reason is that when multiple voltage ranges are preset, the health factor sequences extracted from different voltage ranges during the construction of the offline database are different, resulting in differences in the initial online models (i.e., offline models). In other words, the regression models corresponding to different voltage ranges are not universal. Therefore, in the subsequent updating of each online model, updates can only be achieved using the health factor sequences and corresponding aging state labels obtained within that voltage range. In other words, during the establishment of the online model, for any voltage range from which health factor sequences can be extracted, aging state label calculations and subsequent related operations are performed, fully utilizing the historical operating data of the target battery to obtain an online model more suitable for the target battery. Therefore, for each user, the matching online model fully utilizes the battery's operating data under that usage habit during the establishment and updating process, making it more suitable for representing the battery's aging path under that usage habit. Thus, a more accurate aging status label can be obtained based on the online model.

[0136] The lithium battery aging state device provided in the embodiments of this application is described below. The lithium battery aging state calculation device described below and the lithium battery aging state calculation method described above can be referred to in correspondence.

[0137] Figure 7 This is a schematic diagram of the structure of a lithium battery aging state calculation device disclosed in an embodiment of this application. Figure 7 As shown, the device may include:

[0138] The offline database establishment unit 701 is used to establish an offline database based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery.

[0139] The offline model building unit 702 is used to build an offline model based on the offline database using a regression analysis algorithm.

[0140] The online model building unit 703 is used to build an online database of the target battery based on the offline database, the offline model and the historical operating data of the target battery, and to build an online model based on the online database using a regression analysis algorithm.

[0141] The online data extraction unit 704 is used to extract online health factor sequences based on the online operating data of the target battery;

[0142] The online aging status label determination unit 705 is used to calculate the online aging status label corresponding to the online health factor sequence based on the online model.

[0143] Optionally, the offline database establishment unit 701 can be used to establish multiple offline databases corresponding to different preset voltage ranges. Each offline database includes multiple sets of offline health factor sequences obtained in the corresponding voltage range and corresponding offline aging status tags.

[0144] The offline model building unit 702 can be used to build an offline model corresponding to each voltage range based on a regression analysis algorithm and an offline database corresponding to each voltage range.

[0145] The online model building unit 703 can be used to build multiple online databases corresponding to different preset voltage ranges. Each online database contains multiple sets of historical health factor sequences and corresponding historical aging state labels obtained in the corresponding voltage range. Based on the online database corresponding to each voltage range, an online model corresponding to each voltage range is built using a regression analysis algorithm.

[0146] Optionally, the online data extraction unit 704 can be used to determine the target voltage range corresponding to the online operating data, extract the online health factor sequence within the target voltage range, and obtain the online model corresponding to the target voltage range.

[0147] Optionally, the process by which the offline database establishment unit 701 establishes an offline database based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery may include:

[0148] Extract the offline aging status label and corresponding operating data of the reference battery from the offline operating data;

[0149] Extract offline health factor sequences based on the operational data;

[0150] An offline database is constructed based on the offline health factor sequences and the offline aging status labels.

[0151] Optionally, the process by which the offline database establishment unit 701 extracts the offline aging status tag of the reference battery and the corresponding operating data from the offline operating data may include:

[0152] In the offline operation data, starting from the initial offline aging state label that represents the reference battery not yet aging, for every increase of X% / M in the change of the offline aging state label, an offline aging state label and the corresponding operation data are extracted. The change is used to characterize the degree of aging of the reference battery compared to when it has not yet started aging.

[0153] Where X% is the length of the range of aging status tags available for the target battery, and M is the number of offline aging status tags extracted from the offline operating data.

[0154] Optionally, the historical operating data includes multiple operating data points collected in chronological order, where any one of the operating data points is collected during a single charge of the target battery.

[0155] Based on this, the online model building unit 703 builds an online database for the target battery according to the offline database, the offline model, and the historical operating data of the target battery. The process of building an online model based on the online database using a regression analysis algorithm may include:

[0156] The current running data is obtained from the historical running data in chronological order;

[0157] Extract historical health factor sequences based on the current running data;

[0158] The historical aging status label corresponding to the historical health factor sequence is calculated based on the online model, and the change of the historical aging status label compared with the historical aging status labels corresponding to each running data before the current running data is calculated.

[0159] Determine whether the change meets a preset condition and perform corresponding operations, including:

[0160] If the change amount does not meet the preset conditions, the first process is executed. The first process includes obtaining the next piece of running data from the historical running data and returning to execute the step of extracting the historical health factor sequence based on the current piece of running data.

[0161] If the change meets a preset condition, a second process is executed, the second process including:

[0162] The historical health factor sequence and its corresponding historical aging status label are added to the online database. The historical aging status label and the offline aging status label in the online database are matched. The matched offline aging status label and its corresponding offline health factor sequence are deleted from the online database to obtain an updated online database. The online model is updated based on the updated online database using a regression analysis algorithm, and the first process is executed.

[0163] Wherein, the online database at the initial moment is the offline database, and the online model at the initial moment is the offline model.

[0164] Optionally, the process by which the online model building unit 703 calculates the change in the historical aging status label compared to the historical aging status labels corresponding to each previous running data point may include:

[0165] Calculate the change in the historical aging status label compared to the historical aging status label corresponding to the previous running data before the current running data, and include the change in the cumulative change.

[0166] Based on this, the process by which the online model building unit 703 determines whether the change meets preset conditions and performs corresponding operations may include:

[0167] Determine whether the first condition is met and perform the corresponding operation, wherein the first condition is that the change amount reaches the change amount threshold;

[0168] If the first condition is met, the historical health factor sequence and the corresponding historical aging status label are added to the online database, and the process proceeds to the step of determining whether the second condition is met; if the first condition is not met, the process proceeds to the step of determining whether the second condition is met.

[0169] Determine whether the second condition is met and perform the corresponding operation, wherein the second condition is that the cumulative change amount reaches the cumulative change amount threshold;

[0170] If the second condition is not met, the first process is executed. The first process includes obtaining the next running data from the historical running data and returning to the step of extracting the historical health factor sequence based on the current running data. If the second condition is met, the historical aging status labels in the online database are matched with the offline aging status labels, and the matched offline aging status labels and the corresponding offline health factor sequences are deleted from the online database to obtain an updated online database. The online model is updated based on the updated online database using a regression analysis algorithm, the cumulative change is cleared to zero, and the first process is executed.

[0171] Optionally, the change threshold is X% / M, where X% is the length of the range of available aging state tags for the target battery, and M is the number of offline aging state tags extracted from the offline operating data.

[0172] Optionally, the cumulative change threshold is N*X% / M, where N is a preset positive integer, X% is the length of the range of available aging state tags for the target battery, and M is the number of offline aging state tags extracted from the offline operating data.

[0173] Optionally, the historical operating data includes multiple operating data, one of which is data collected during a single charge of the target battery;

[0174] Based on this, the process by which the online model building unit 703 builds the online database of the target battery according to the offline database, the offline model, and the historical operating data of the target battery may include:

[0175] Based on each piece of historical operational data, the corresponding historical health factor sequence is extracted, and the historical aging status label corresponding to the historical health factor sequence is calculated based on the offline model.

[0176] Add the historical health factor sequence and the corresponding historical aging status label to the offline database to obtain the updated offline database;

[0177] In the updated offline database, the historical aging status tags and offline aging status tags are matched, and the matched offline aging status tags and corresponding offline health factor sequences are deleted from the updated offline database to obtain the online database.

[0178] The lithium battery aging state calculation device provided in this application calculates the aging state label corresponding to the historical operating data of the target battery through a regression model. The regression model can be an offline model established based on a reference battery, and it integrates the offline database of the reference battery and the historical health factor sequence and corresponding historical aging state label of the target battery. The integrated database is then used to establish an online model, which solves the problem of missing aging state labels in the early stage of target battery aging, improves the applicability of the online model to the target battery, and improves the accuracy of the calculated online aging state label corresponding to the online operating data.

[0179] The lithium battery aging state calculation device provided in this application embodiment can be applied to electronic devices, such as terminals: mobile phones, computers, etc. Optionally, Figure 8 A hardware block diagram of the electronic device is shown, with reference to... Figure 6 The hardware structure of an electronic device may include: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4;

[0180] In this embodiment of the application, the number of processor 1, communication interface 2, memory 3, and communication bus 4 is at least one, and processor 1, communication interface 2, and memory 3 communicate with each other through communication bus 4;

[0181] Processor 1 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0182] Memory 3 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device;

[0183] The memory stores a program, which the processor can call. The program is used for:

[0184] An offline database is established based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery. The offline database includes multiple sets of offline health factor sequences and corresponding offline aging state labels.

[0185] An offline model is established based on the offline database and a regression analysis algorithm.

[0186] Based on the offline database, the offline model, and the historical operating data of the target battery, an online database for the target battery is established, wherein the online database contains a historical health factor sequence determined based on the historical operating data and corresponding historical aging state labels;

[0187] An online model is established based on the online database and a regression analysis algorithm.

[0188] Based on the online operating data of the target battery, extract the online health factor sequence;

[0189] Based on the online model, calculate the online aging status label corresponding to the online health factor sequence.

[0190] Optionally, the refined and extended functions of the program can be found in the description above.

[0191] This application embodiment also provides a storage medium that can store a program suitable for execution by a processor, the program being used for:

[0192] An offline database is established based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery. The offline database includes multiple sets of offline health factor sequences and corresponding offline aging state labels.

[0193] An offline model is established based on the offline database and a regression analysis algorithm.

[0194] Based on the offline database, the offline model, and the historical operating data of the target battery, an online database for the target battery is established, wherein the online database contains a historical health factor sequence determined based on the historical operating data and corresponding historical aging state labels;

[0195] An online model is established based on the online database and a regression analysis algorithm.

[0196] Based on the online operating data of the target battery, extract the online health factor sequence;

[0197] Based on the online model, calculate the online aging status label corresponding to the online health factor sequence.

[0198] Optionally, the refined and extended functions of the program can be found in the description above.

[0199] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0200] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0201] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for calculating the aging state of a lithium battery, characterized in that, include: An offline database is established based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery. The offline database includes multiple sets of offline health factor sequences and corresponding offline aging state labels. An offline model is established based on the offline database and a regression analysis algorithm. Based on the offline database, the offline model, and the historical operating data of the target battery, an online database for the target battery is established. The online database contains a sequence of historical health factors and corresponding historical aging status labels determined based on the historical operating data. An online model is established based on the online database and a regression analysis algorithm. Based on the online operating data of the target battery, extract the online health factor sequence; Based on the online model, calculate the online aging status label corresponding to the online health factor sequence; The historical operating data includes multiple operating data points collected in chronological order, where any one of the operating data points is data collected during a single charge of the target battery. The step of establishing an online database for the target battery based on the offline database, the offline model, and the historical operating data of the target battery, and establishing an online model based on the online database using a regression analysis algorithm, includes: The current running data is obtained from the historical running data in chronological order; Extract historical health factor sequences based on the current running data; The historical aging status label corresponding to the historical health factor sequence is calculated based on the online model, and the change of the historical aging status label compared with the historical aging status labels corresponding to each running data before the current running data is calculated. If the change amount does not meet the preset conditions, the first process is executed. The first process includes obtaining the next piece of running data from the historical running data and returning to execute the step of extracting the historical health factor sequence based on the current piece of running data. If the change meets a preset condition, a second process is executed, the second process including: The historical health factor sequence and the corresponding historical aging status label are added to the online database. The historical aging status label and the offline aging status label in the online database are matched. The matched offline aging status label and the corresponding offline health factor sequence are deleted from the online database to obtain the updated online database. The online model is updated based on the updated online database using a regression analysis algorithm, and the first process is executed. Wherein, the online database at the initial moment is the offline database, and the online model at the initial moment is the offline model.

2. The method according to claim 1, characterized in that, The offline database includes multiple offline databases corresponding to different preset voltage ranges. Each offline database includes multiple sets of offline health factor sequences obtained in the corresponding voltage range and corresponding offline aging status tags. The step of establishing an offline model based on the offline database using a regression analysis algorithm includes: Based on the offline database corresponding to each voltage range, an offline model corresponding to each voltage range is established using a regression analysis algorithm. The online database includes multiple online databases that correspond to different preset voltage ranges. Each online database contains multiple sets of historical health factor sequences and corresponding historical aging state tags obtained in the corresponding voltage range. The step of establishing an online model based on the online database using a regression analysis algorithm includes: Based on the online database corresponding to each voltage range, an online model corresponding to each voltage range is established using a regression analysis algorithm.

3. The method according to claim 2, characterized in that, The step of extracting online health factor sequences based on the online operating data of the target battery includes: Based on the online operating data, a target voltage range is determined, and an online health factor sequence is extracted within the target voltage range. Obtain the online model corresponding to the target voltage range.

4. The method according to claim 1, characterized in that, The establishment of an offline database based on offline operating data of a reference battery with the same capacity and material system as the target battery includes: Extract the offline aging status label and corresponding operating data of the reference battery from the offline operating data; Extract offline health factor sequences based on the operational data; An offline database is composed of the offline health factor sequences and the offline aging status tags.

5. The method according to claim 4, characterized in that, The step of extracting the offline aging status tag and corresponding operating data of the reference battery from the offline operating data includes: In the offline operation data, starting from the initial offline aging state label that represents the reference battery not yet starting to age, for every increase of X% / M in the change of the offline aging state label, an offline aging state label and the corresponding operation data are extracted. The change is used to characterize the degree of aging of the reference battery compared to when it has not started to age. Where X% is the length of the range of aging status tags available for the target battery, and M is the number of offline aging status tags extracted from the offline operating data.

6. The method according to any one of claims 1 to 5, characterized in that, The calculation of the change in the historical aging status label compared to the historical aging status labels corresponding to each running data point before the current running data point includes: Calculate the change in the historical aging status label compared to the historical aging status label corresponding to the previous running data before the current running data, and include the change in the cumulative change. The preset conditions include: a first condition and a second condition, wherein the first condition is that the change amount reaches a change amount threshold, and the second condition is that the cumulative change amount reaches a cumulative change amount threshold; Then, if the change amount does not meet the preset conditions, the first process is executed, including: If neither the first condition nor the second condition is met, the first process is executed; When the change meets a preset condition, the second process is executed, including: Determine whether the first condition is met. If yes, add the historical health factor sequence and the corresponding historical aging status label to the online database and proceed to the step of determining whether the second condition is met; otherwise, proceed to the step of determining whether the second condition is met. Determine whether the second condition is met. If not, execute the first process. If yes, match the historical aging status labels in the online database with the offline aging status labels, and delete the matched offline aging status labels and corresponding offline health factor sequences from the online database to obtain an updated online database. Use the updated online database to update the online model based on a regression analysis algorithm, clear the cumulative change to zero, and execute the first process.

7. The method according to claim 6, characterized in that, The change threshold is X% / M, where X% is the length of the range of available aging state tags for the target battery, and M is the number of offline aging state tags extracted from the offline operating data.

8. The method according to claim 6, characterized in that, The cumulative change threshold is N*X% / M, where N is a preset positive integer, X% is the length of the range of available aging state tags for the target battery, and M is the number of offline aging state tags extracted from the offline operation data.

9. The method according to any one of claims 1 to 5, characterized in that, The historical operating data includes multiple operating data points, one of which is data collected during a single charge of the target battery. The step of establishing an online database for the target battery based on the offline database, the offline model, and the historical operating data of the target battery includes: Based on each piece of historical operational data, the corresponding historical health factor sequence is extracted, and the historical aging status label corresponding to the historical health factor sequence is calculated based on the offline model. Add the historical health factor sequence and the corresponding historical aging status label to the offline database to obtain the updated offline database; In the updated offline database, the historical aging status tags and offline aging status tags are matched, and the matched offline aging status tags and corresponding offline health factor sequences are deleted from the updated offline database to obtain the online database.

10. A lithium battery aging state calculation device, used to execute the lithium battery aging state calculation method according to any one of claims 1 to 9, characterized in that, include: An offline database establishment unit is used to establish an offline database based on the offline operating data of a reference battery with the same capacity and belonging to the same material system as the target battery. The offline database includes multiple sets of offline health factor sequences and corresponding offline aging state labels. An offline model building unit is used to build an offline model based on the offline database using a regression analysis algorithm. An online model building unit is used to build an online database for the target battery based on the offline database, the offline model, and the historical operating data of the target battery. The online database contains historical health factor sequences and corresponding historical aging state labels determined based on the historical operating data. An online model is built based on the online database using a regression analysis algorithm. An online data extraction unit is used to extract online health factor sequences based on the online operating data of the target battery; An online aging status label determination unit is used to calculate an online aging status label corresponding to the online health factor sequence based on the online model.

11. An electronic device, characterized in that, include: Memory and processor; The memory is used to store programs; The processor is used to execute the program to implement each step of the lithium battery aging state calculation method as described in any one of claims 1 to 9.

12. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements each step of the lithium battery aging state calculation method as described in any one of claims 1 to 9.

Images