Storage battery life prediction method and system and electronic equipment

Abstract

The embodiment of the invention provides a storage battery service life prediction method and system and electronic equipment, and relates to the technical field of lead-acid storage batteries, and the method can prolong the service life of a storage battery. The storage battery comprises a plurality of battery units. The method comprises the following steps: receiving battery pack data and user behavior data; battery pack data including charging data and discharging data of each of the plurality of battery cells; the user behavior data comprises behavior data of charging a storage battery by a user; according to the battery pack data and the user behavior data, obtaining features of multiple dimensions corresponding to each battery unit in the multiple battery units; determining at least one feature influencing the service life of the battery from the features of the multiple dimensions; dynamically adjusting the weight corresponding to the at least one feature; according to the dynamically adjusted weight and the features of the multiple dimensions, performing weighted fusion calculation to obtain a weighted feature vector; and inputting the weighted feature vectors into a trained prediction model to obtain a life prediction result corresponding to each battery unit.

Description

Technical Field

[0001] The present application relates to the technical field of lead-acid batteries, and particularly to a method, a system and an electronic device for predicting the service life of a battery. Background Art

[0002] As one of the core components of an automotive power system, an in-vehicle battery not only plays a key role in starting the engine, but its service life and health status directly affect the normal operation of the vehicle.

[0003] Taking a lead-acid battery as an example, as a commonly used in-vehicle battery, due to its low energy density, lead-acid batteries are usually used in groups, and there are differences in parameters such as the chemical composition, temperature influence, self-discharge rate, capacity and internal resistance of each single battery (or battery cell). At present, a unified charging management strategy is adopted for different single batteries, and batteries with lower capacity are more likely to experience overcharging and over-discharging. When single lead-acid batteries are connected to the system in series, the repeated charging and discharging processes exacerbate the imbalance phenomenon, resulting in serious sulfation of the plates of individual batteries, increased internal resistance and reduced effective active substances. The damage of a single battery often leads to the failure of the entire battery pack. Summary of the Invention

[0004] In view of this, the embodiments of the present application provide a method, a system and an electronic device for predicting the service life of a battery, which can predict the remaining service life of each battery cell, so as to formulate a targeted charging strategy and avoid the failure of the entire battery pack caused by unreasonable local charging management strategies of single batteries.

[0005] In a first aspect, the embodiments of the present application provide a method for predicting the service life of a battery. The battery includes a plurality of battery cells, and the method includes: receiving battery pack data and user behavior data; the battery pack data includes charging data and discharging data of each battery cell in the plurality of battery cells; the user behavior data includes behavior data of the user charging the battery; obtaining, according to the battery pack data and the user behavior data, a plurality of dimensions of features corresponding to each battery cell in the plurality of battery cells; determining at least one feature affecting the battery life from the features of the plurality of dimensions; dynamically adjusting the weights corresponding to the at least one feature; performing weighted fusion calculation according to the dynamically adjusted weights and the features of the plurality of dimensions to obtain a weighted feature vector; and inputting the weighted feature vector into a trained prediction model to obtain a life prediction result corresponding to each battery cell.

[0006] In one implementable manner, determining at least one feature that affects the battery life from the features in multiple dimensions includes: determining the initial weights corresponding to the features in multiple dimensions respectively, and obtaining a first weighted feature vector according to the initial weights and the values of the features in multiple dimensions; using the first weighted feature vector as training data to train a prediction model to obtain a first prediction model; obtaining a first life prediction result corresponding to each battery cell through the first prediction model; calculating a first prediction error of the first prediction model according to the first life prediction result and the actual life corresponding to each battery cell; adjusting the first initial weight corresponding to the first feature in the features in multiple dimensions to obtain a first weight; keeping the weights corresponding to the features in other dimensions unchanged; calculating an adjusted second prediction error; calculating the change amount between the first prediction error and the second prediction error; and determining the first feature as a feature that affects the battery life when the change amount meets a preset condition.

[0007] In one implementable manner, adjusting the first initial weight corresponding to the first feature in the features in multiple dimensions includes: adjusting the first initial weight corresponding to the first feature in the features in multiple dimensions according to the following calculation formula: ; where represents the adjusted first weight, represents the first initial weight corresponding to the first feature; γ represents a control factor, and the control factor is a preset constant.

[0008] In one implementable manner, calculating the adjusted second prediction error includes: performing weighted fusion calculation according to the adjusted first weight and the initial weights corresponding to the features in other dimensions to obtain a second weighted feature vector; using the second weighted feature vector to continue training the first prediction model to obtain a second prediction model; obtaining a second life prediction result corresponding to each battery cell through the second prediction model; and calculating a second prediction error of the second prediction model according to the second life prediction result and the actual life.

[0009] In one implementable manner, dynamically adjusting the weights corresponding to at least one feature includes: determining the gain factors corresponding to at least one feature respectively; multiplying the at least one initial weight corresponding to at least one feature by the corresponding gain factor to obtain the adjusted weights.

[0010] In an implementable manner, battery pack data and user charging behavior data uploaded by a battery management device are received, including: receiving battery pack data and user charging behavior data in different cycles; according to the battery pack data and the user charging behavior data, obtaining features of multiple dimensions corresponding to each battery cell among a plurality of battery cells, including: according to the battery pack data and the user charging behavior data, obtaining features of multiple dimensions corresponding to each battery cell among the plurality of battery cells in different cycles respectively; after obtaining the features of multiple dimensions corresponding to different cycles respectively, the method further includes: obtaining a life prediction result corresponding to each battery cell in different cycles respectively according to the features of multiple dimensions corresponding to different cycles respectively; determining gain factors corresponding to at least one feature, including: calculating a difference between a prediction error corresponding to the latest cycle and a prediction error corresponding to the previous cycle according to the life prediction results corresponding to different cycles respectively; determining gain factors corresponding to at least one feature respectively according to a ratio of the difference to the prediction error corresponding to the previous cycle and a multiplication factor; wherein, the multiplication factor is a preset constant.

[0011] In an implementable manner, determining gain factors corresponding to at least one feature respectively according to a ratio of the difference to the prediction error corresponding to the previous cycle and a multiplication factor includes: calculating a ratio of the difference to the prediction error corresponding to the previous cycle; calculating a product of the ratio and the multiplication factor, and taking the calculation result as the value of the gain factor corresponding to a corresponding one of at least one feature.

[0012] In an implementable manner, the prediction error includes: mean square error MSE and / or mean absolute error MAE.

[0013] In one implementable manner, the features of the multiple dimensions include at least one of the following features: charging feature, discharging feature, temperature feature, voltage feature, usage behavior feature, efficiency and life feature; wherein, the charging feature includes at least one of charging frequency, average charging time, and average charging power; wherein, the charging frequency is determined according to the number of charging times of the user within a predetermined time period; the average charging time is obtained by averaging the charging times of at least one charging of the user within a predetermined time period; the average charging power is obtained by averaging the charging powers of at least one charging of the user within a predetermined time period; the discharging feature includes at least one of average discharge depth, discharge duration, and average discharge power; wherein, the average discharge depth is obtained by averaging the discharge depths corresponding to at least one charging of the user within a predetermined time period; the discharge duration is obtained by summing the discharge durations of at least one discharge within a predetermined time period; the average discharge power is obtained by averaging the discharge powers of at least one discharge within a predetermined time period; the temperature feature includes ambient temperature change feature and / or battery temperature change feature; wherein, the ambient temperature change feature is calculated according to the difference between the maximum value and the minimum value of the ambient temperature collected within a predetermined time period; the battery temperature change feature is calculated according to the difference between the maximum value and the minimum value of the battery temperature collected within a predetermined time period; the voltage feature includes voltage fluctuation feature and / or voltage stability feature; wherein, the voltage fluctuation feature is obtained by calculating the standard deviation of the voltage values collected multiple times within a predetermined time period and the average voltage value; the voltage stability feature is obtained by calculating the ratio of the difference between the maximum voltage value and the minimum voltage value collected within a predetermined time period to the average voltage; the usage behavior feature includes charging periodicity feature and / or discharging periodicity feature; the charging periodicity feature is determined according to the charging frequency of the user within a predetermined time period; the discharge cycle feature is determined according to the discharging frequency of the user within a predetermined time period; the efficiency and life feature includes charge-discharge efficiency and / or total number of cycles; wherein, the charge-discharge efficiency is obtained by calculating the ratio of the charging energy to the discharging energy; the total number of cycles represents the sum of the number of charging times and / or discharging times experienced by the battery.

[0014] In a second aspect, an embodiment of the present application further provides a battery management system, including a battery management device and a cloud data platform; the battery management device includes a main control module, a plurality of single-cell monitoring modules, and a user data monitoring module; wherein, the single-cell monitoring module is configured to collect charging data and discharging data of each battery cell in a plurality of battery cells of the battery to obtain battery pack data; the user data monitoring module is configured to monitor the behavior data of the user charging the battery to obtain user behavior data; the main control module is configured to upload the battery pack data and the user behavior data to the cloud data platform; the cloud data platform includes a server, and the server is configured to implement the method according to any one of the first aspects described above.

[0015] In a third aspect, an embodiment of the present application further provides an electronic device, which includes a processor for executing a computer program or instruction in a memory to implement the method described in any one of the above first aspects.

[0016] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which includes a stored program. When the program is executed by a processor, the method described in any one of the above first aspects is implemented.

[0017] In a fifth aspect, an embodiment of the present application further provides a computer program product, which includes a program. When the program is run on an electronic device, the electronic device is enabled to implement the method described in any one of the above first aspects.

[0018] In a sixth aspect, an embodiment of the present application further provides a chip system, which includes a communication interface for inputting and / or outputting data, and a processor for executing a computer-executable program, so that a device installed with the chip system executes the method described in any one of the above first aspects.

[0019] The method proposed in the embodiment of the present application extracts features in multiple dimensions based on battery pack data and user behavior data, and can dynamically adjust the weight of each feature in real time according to the latest uploaded battery pack data and user behavior data. A weighted feature vector is calculated according to the adjusted weight, and the remaining life of each battery cell is predicted based on the weighted feature vector. In this way, it can dynamically respond to feature changes. When the system monitors that the importance of a certain feature (such as charging frequency) has increased significantly, it is because the user's charging behavior has changed, and the weight of the corresponding feature can be increased, so as to predict a more accurate remaining life and improve the accuracy and adaptability of the prediction. After predicting the remaining life of each single battery, a targeted charging management strategy can be formulated according to the remaining life, which can not only improve the life of a single battery cell, but also avoid the failure of the entire battery pack to a certain extent due to unreasonable local charging management strategies of single batteries, that is, reduce battery failures and improve the life of the battery pack. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required to be used in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 It is a schematic structural diagram of the electronic device (server) provided by the embodiment of the present application; Figure 2 An example diagram of the system architecture adopted in some embodiments of the battery life prediction method provided by the embodiments of the present application; Figure 3 A schematic structural diagram of a lead-acid battery management system in some embodiments of the battery life prediction method provided by the embodiments of the present application; Figure 4 A schematic flow diagram of the battery life prediction method provided by the embodiments of the present application. Detailed implementation manners

[0022] To better understand the technical solution of the present application, the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0023] It should be clear that the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present application.

[0024] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms of "a", "the" and "said" used in the embodiments of the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0025] It should be understood that the term " / and / " used herein is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after.

[0026] The main functions of on-vehicle batteries include providing high-power starting current to ensure reliable starting of the engine, assisting the generator to provide stable power supply for various electrical equipment, and maintaining the operation of the vehicle's electronic system when the engine is turned off. Therefore, the good condition of on-vehicle batteries is not only a key factor in optimizing vehicle performance, but also directly related to driving safety and the control of vehicle operation costs.

[0027] Currently, on-vehicle batteries are generally lead-acid batteries. Due to their unique economy and low production cost, and relatively high safety under normal use, lead-acid battery packs are commonly used on-vehicle batteries in the automotive industry. However, due to their low energy density, lead-acid batteries are usually used in groups, and the failure of a single battery often leads to the failure of the entire battery pack.

[0028] Specifically analyzed, a lead-acid battery pack includes multiple single cells. When the single cells are connected to the system in series, there are differences in parameters such as the chemical composition, temperature influence, self-discharge rate, capacity, and internal resistance of each single cell. The single cell with a lower capacity is more likely to experience overcharging and over-discharging phenomena, thus causing the lead-acid battery pack to fall into a vicious cycle of increasing plate sulfation and gradually expanding capacity gap. The repeated charge and discharge processes exacerbate the imbalance phenomenon, resulting in serious plate sulfation, increased internal resistance, and reduced effective active material in individual batteries, leading to a significant shortening of the life of single cells and a significant difference in the lives of different single cells. Currently, the charging strategy does not consider adopting a targeted charging strategy for single cells with different lives, and affected by the life of individual single cells, the life of the entire battery pack is shortened.

[0029] In addition, the performance of lead-acid batteries is also affected at extreme temperatures. Both too low or too high temperatures will reduce their performance.

[0030] It can be seen that a targeted charging strategy needs to be set for the single cells in the lead-acid battery pack, and the remaining life of each single cell is an important reference factor for determining the charging strategy. How to accurately predict the life of each single cell in a plurality of single cells in a lead-acid battery pack has become a technical problem that needs to be solved urgently by those skilled in the art.

[0031] In view of this, the embodiments of the present application propose a method, system, and electronic device for predicting the life of a battery, which can predict the remaining life of each single cell (or battery unit) in a lead-acid battery pack, and then adopt a targeted charging strategy for different single cells, so as to solve the problem of unbalanced charging of individual batteries with different remaining lives in a series-connected lead-acid battery pack, aiming to improve the service life of the series-connected lead-acid battery pack and solve the problem of lack of specificity of the current mainstream equalization charging mode for individual batteries in the battery pack.

[0032] The method, system, and electronic device for predicting the life of a battery proposed in the embodiments of the present application can be applied to various application scenarios that require loading lead-acid batteries. For example, it is used to predict the life of each single cell in a lead-acid battery pack loaded in vehicles such as cars, buses, trucks, and motorcycles, and adjust the charging strategy according to the predicted life. Or, it can also be used to predict the life of lead-acid batteries in the energy storage systems of electric vehicles and hybrid electric vehicles. Moreover, it can also be applied in the communication industry to predict the life of lead-acid batteries in the power supply systems of devices such as telephone exchanges, mobile communication base stations, and optical fiber communication systems, providing a reference for intelligent charging strategies.

[0033] In addition, in the embodiments of the present application, lead-acid batteries are mostly used as examples for illustration. It should not be understood that the method proposed in the embodiments of the present application can only be applied to the charge and discharge management application scenarios of lead-acid batteries, and can also be applied to other types of batteries.

[0034] The electronic device proposed in the embodiments of the present application can be a server, specifically an independent server or a server cluster. Exemplarily, Figure 1 It is a schematic structural diagram of a server in an embodiment of the present application. As Figure 1 shown, the server 100 may include: one or more processors 110, a communication interface 120, a memory 130, and a communication bus 140 connecting different components (including the memory 130, the communication interface 120, and the processor 110).

[0035] The communication bus 140 represents one or more of several types of bus structures, including a memory bus or a memory controller, a peripheral bus, a graphics acceleration port, or a local bus using any of the multiple bus structures. For example, the communication bus 140 may include, but is not limited to, an Industry Standard Architecture (ISA) bus, a Micro Channel Architecture (MCA) bus, an Enhanced ISA bus, a Video Electronics Standards Association (VESA) local bus, and a Peripheral Component Interconnection (PCI) bus.

[0036] An electronic device typically includes a variety of computer system-readable media. These media can be any available media accessible by the electronic device, including volatile and non-volatile media, removable and non-removable media.

[0037] The memory 130 may include computer system-readable media in the form of volatile memory, such as: Random Access Memory (RAM) and / or cache memory. The memory 130 may include at least one program product having a set (for example, at least one) of program modules configured to execute the computing power resource scheduling method provided in the embodiments of the present application.

[0038] A program / util utility having a set (at least one) of program modules can be stored in the memory 130. Such program modules include, but are not limited to, an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment. The program modules generally execute the functions and / or methods in the embodiments of the present application.

[0039] The processor 110 executes various functional applications and data processing by running the programs stored in the memory 130, such as implementing the computing power resource scheduling method provided in the embodiments of the present application.

[0040] It should be understood that Figure 1 The processor 110 in the shown server 100 may be a system-on-chip (SOC). The processor 110 may include a central processing unit (CPU) and may further include other types of processors, such as a graphics processing unit (GPU), etc.

[0041] As Figure 2 shown, the method proposed in the embodiments of the present application can be implemented based on Figure 2 the shown system architecture. The system architecture includes a lead-acid battery pack 201, a battery management device 202, and a cloud data platform 203. The battery management device is used to collect the charging data and discharging data of each single battery in a plurality of single batteries in the lead-acid battery, and record the behavior data of the user U1 charging the lead-acid battery, and upload the collected data to the cloud data platform 203. A prediction model is deployed in the cloud data platform 203, which can perform online learning, extract multi-dimensional features according to the collected data, predict the life of the single battery according to the multi-dimensional features through the prediction model, and calculate the prediction error. According to the prediction error, the weight of at least one feature in the multi-dimensional features is dynamically adjusted. Based on the adjusted weight, the prediction model is trained, and the life of the single battery is predicted through the trained prediction model. And, in some embodiments, the cloud data platform can also configure corresponding charging strategies according to the predicted life, and adopt different charging strategies for different single batteries with different lives to improve the life of a single single battery, thereby avoiding the shortening of the life of the entire battery pack due to the life of a single single battery.

[0042] Specific embodiments are listed below.

[0043] An embodiment of the present application provides a lead-acid battery management system, which is a series-connected on-vehicle battery detection, maintenance, and charging system based on multi-source data fusion. It includes a battery management device and a cloud data platform. The cloud data platform uses a multi-source data fusion algorithm to predict the lifespan of each individual battery, and solves the problem of unbalanced charging of individual batteries with different remaining lifespans in a series-connected lead-acid battery pack.

[0044] Specifically, this system is a battery detection, maintenance, and charging system that uses a battery cell detection unit to detect the voltage, internal resistance, and temperature of the battery in real time. As Figure 3 shown, the battery management device in this system can include a main control module, N single-cell monitoring modules, a user data monitoring module, a power supply, and an interaction module. The main control module in the battery management device is communicatively connected to the cloud data platform. Among them, the N single-cell monitoring modules are used to monitor N single batteries, and one single-cell monitoring module is used to monitor one single battery (or battery cell).

[0045] Among them, the main control module is the core of the entire battery management system. By working in coordination with the single-cell monitoring module and the user data monitoring module, it can comprehensively obtain the operation data of the battery pack and the usage behavior data of the user.

[0046] For example, the single-cell monitoring module monitors the voltage, temperature, and status of each battery cell in real time, while the user data monitoring module collects the usage pattern data of the user, such as charging habits, depth of discharge, etc. The main control module aggregates and processes this information, and analyzes the health status and operation status of the battery pack in real time to make the optimal charging control decision.

[0047] The main control module can be implemented based on an advanced microcontroller architecture, and can maintain high computing performance in complex battery management tasks. Specifically, the main control module can adopt a real-time operating system (RTOS) to enable it to quickly respond to the state changes of the battery pack with millisecond-level precision, ensuring that the system remains stable and safe under dynamic conditions.

[0048] In terms of data communication, the main control module integrates a wireless communication function and supports multiple transmission protocols such as Wi-Fi and Bluetooth. Through these wireless communication methods, the module can quickly and accurately transmit battery data to the cloud data platform.

[0049] In addition, to ensure the security of data transmission, the main control module adopts an encrypted communication protocol to ensure that sensitive information will not be externally attacked or stolen during the transmission process. The cloud data platform can then receive and store this information in real time.

[0050] The interaction module can specifically be an interface for implementing the interaction between the user and the system. For example, the interaction between the user and the system is mainly achieved through an intuitive touch screen interface. This touch screen adopts advanced multi-touch technology, supports the recognition of operations of multiple fingers simultaneously, and brings a smoother user experience. The screen can display various key states of the battery pack in real time, including detailed information such as the current battery power, charge and discharge progress, and remaining usage time.

[0051] In some embodiments, the touch screen can be equipped with an intelligent fault alarm system. When an abnormal situation occurs in the battery pack, the system will immediately notify the user by means of popping up a warning window, color change, or sound prompt. The user can directly view the alarm details, view the recommended solutions, and even remotely contact technical support through touch operations. There are also customizable function modules on the interface, and users can adjust the display content according to their needs to keep track of the information most relevant to their needs in real time.

[0052] The single-cell monitoring module, after collecting real-time data of a single battery (such as voltage, temperature, and internal resistance), uploads the data to the cloud data platform through the main control module. The collected data will be classified according to different application requirements and divided into training samples, validation samples, and test samples for subsequent machine learning and data analysis.

[0053] The user data monitoring module can monitor the user's usage data in real time, including key indicators such as charging frequency, charging duration, and discharge depth. These data are recorded through an efficient acquisition system and uploaded to the cloud data platform in a timely manner using advanced communication technology. On the cloud data platform, this information will be centrally stored.

[0054] The cloud data platform is responsible for receiving the data transmitted from the single-cell monitoring module and the user data monitoring module and centrally summarizing these data. In this process, the platform applies a multi-source data fusion algorithm to fuse and process data from different sources such as the operating state of the battery unit and the user's usage habits (or behavior habits). Through this multi-source data fusion algorithm, the platform can integrate the redundant information of multiple data sources, complement their respective deficiencies, and improve the accuracy and consistency of the data.

[0055] The fused data can not only more accurately reflect the overall operating state of the battery pack but also deeply analyze potential problems with battery performance and user behavior patterns, providing a reliable basis for intelligent fault prediction, health status assessment, and charging strategy optimization. The application of multi-source data fusion ensures that the platform can effectively improve the accuracy and response speed of decision-making when processing large-scale, heterogeneous data.

[0056] The following is based on Figure 2 the system architecture shown and Figure 3The battery management system shown exemplarily elaborates on the method for predicting the service life of a storage battery proposed in the embodiments of the present application.

[0057] The cloud data platform may include a server, and the multi-source data fusion algorithm proposed in the embodiments of the present application can be applied to the server in the cloud data platform. As Figure 4 shown, from the perspective of the server, the method may include the following processes: S10: Receive battery pack data and user behavior data.

[0058] The battery pack data includes the charging data and discharging data of each battery cell in a plurality of battery cells. The user behavior data is the behavior data of the user charging the lead-acid storage battery. Specifically, the battery pack data and user behavior data can be obtained through the single-cell monitoring module and the user data monitoring module, and then these data are uploaded to the cloud data platform in an encrypted manner through the main control module.

[0059] S11: According to the battery pack data and user behavior data, obtain multiple dimensions of features corresponding to each battery cell in the plurality of battery cells.

[0060] After receiving the data from the main control module, the cloud data platform first performs a series of preprocessing operations on the collected data to ensure the quality and consistency of the data. For example, the cloud data platform can clean and denoise the data, remove invalid, duplicate or abnormal data points, and avoid these noises affecting subsequent analysis.

[0061] Next, the cloud data platform normalizes or standardizes the data from different sensors and data sources, unifies its measurement units and ranges, so as to better fuse and compare. Among them, the sensor can be the sensor in the single-cell monitoring module. For the missing parts in the data, the cloud data platform can use the interpolation algorithm to estimate and fill in the missing values to maintain the integrity of the data. In terms of denoising, the cloud data platform can use filtering techniques, such as using low-pass filtering, Kalman filtering, etc., to smooth the noise in the data.

[0062] After the cloud data platform completes the data preprocessing, key features that can reflect the battery performance and user behavior patterns are extracted from the original data, so as to reduce the data dimension and retain the most representative and valuable information for analysis.

[0063] In this embodiment, exemplarily, the original data (including battery pack data and user behavior data) uploaded by the battery management device is refined and characterized into the following multiple dimensions of features: charging feature, discharging feature, temperature feature, voltage feature, usage pattern feature, efficiency and life feature.

[0064] Among them, the feature extraction process of each dimension is as follows: Charging characteristics: including charging frequency, average charging time and average charging power.

[0065] Among them, charging frequency (R_ch): indicates the number of times the user charges the battery in a unit time, for example, the unit time can be one day (24 hours). The calculation method is to record the total number of charging times within the preset time, divide it by the number of days, and get the average daily charging frequency. For example, the preset time can be one week, and the total number of charging times within one week divided by 7 days can be used to get the charging frequency.

[0066] Average charging time (T_ch): indicates the average time taken by the user to charge each time. It can be calculated by summing up all charging times within the preset time and dividing by the number of charging times. For example, if the preset time is one week, record the duration of each charging, sum up the duration of multiple charging times and divide it by the number of charging times to get the average charging time.

[0067] Average charging power (P_ch): Indicates the average power received by the battery during the charging process. It is calculated by summing up all charging powers within the preset time and dividing by the number of charging times. Similarly, the preset time can be 7 days, recording the charging power of each charge within 7 days, summing up the charging power and dividing it by the number of charging times within 7 days to get the average charging power.

[0068] Discharge characteristics: The discharge characteristics can be the average discharge depth (D_d), which indicates the average depth of discharge of the battery during multiple discharges within a preset time. It is calculated by summing all the discharge depths within the preset time and dividing by the number of discharges. The discharge depth is the ratio of the battery's discharged power to its total capacity. For example, the discharge depth of a lead-acid battery is about 80%, that is, it is charged when 20% of the power is left.

[0069] Discharge duration (T_d): indicates the total duration of a battery discharge cycle, which is calculated by summing up all the discharge times between two adjacent charges. For example, the discharge time of at least one discharge of the battery between the end of one charge and the start of the next charge is recorded, and the total duration of one discharge cycle is obtained as the discharge duration.

[0070] Average discharge power (P_d): represents the average power output by the battery during one round of discharge. It is calculated by summing up all discharge powers between two adjacent charges and dividing by the number of discharges. For example, the discharge power of each discharge in at least one discharge of the battery during the period between the end of one charge and the start of the next charge is recorded, summed up, and divided by the number of discharges.

[0071] Temperature characteristics: Ambient temperature change (ΔT_env): Represents the amplitude of the ambient temperature change during one charging process or one discharging process, calculated by the difference between the maximum and minimum ambient temperatures. One charging process can be the time period from the start time of charging to the end time of this charging; one discharging process can be the time period between two adjacent charging processes. Or it can be the time period between the start time and the end time of a single discharging. For example, during the start time to the end time of a single discharging, the maximum and minimum ambient temperatures within this time period can be counted, and the difference can be calculated to obtain the temperature change characteristics.

[0072] Battery temperature change (ΔT_bat): Represents the amplitude of the battery's own temperature change during one charging process or one discharging process, calculated by the difference between the maximum and minimum battery temperatures.

[0073] Voltage characteristics: Voltage fluctuation (V_v): Represents the degree of battery voltage fluctuation during one charging process or one discharging process. It is obtained by calculating the sum of the squares of the differences between each voltage value and the average voltage, then taking the average value and taking the square root. For example, during one charging process or one discharging process, voltage is sampled every ten minutes, and then the average value of the voltage values at all sampling points during one charging or discharging process is calculated.

[0074] Voltage stability (S_v): Represents the degree of voltage stability. The calculation method can be, during one charging process or one discharging process, the standard deviation of the difference between the maximum and minimum battery voltages and the average voltage, reflecting the relative stability of voltage changes. For example, the degree of voltage stability can be obtained during one discharging process of the battery or during one charging process, obtaining the maximum and minimum battery voltages, through which the magnitude of voltage change can be reflected. Then calculate the average value of the voltage values at all sampling points during one charging or discharging process: ; According to the difference between each sampling point's voltage value and the average voltage, calculate the mean of the sum of squares and take the square root: .

[0075] Usage pattern characteristics: Charge-discharge periodicity (C): Represents the frequency of charge and discharge within a preset time. The calculation method is to divide the total number of charge and discharge times within the preset duration by the total number of usage days, reflecting the regularity of the user's usage pattern. For example, the preset duration is one week.

[0076] Efficiency and life characteristics: Include charge-discharge efficiency and total number of cycles.

[0077] Charge-discharge efficiency (E): It represents the energy conversion efficiency during the charging process. The calculation method is to multiply the ratio of charging energy to discharging energy by 100% to obtain a percentage representation.

[0078] Total number of cycles (N_c): It represents the total number of charge-discharge cycles that the battery has experienced and reflects the service life of the battery.

[0079] According to the above exemplary description, features in multiple dimensions can be obtained.

[0080] S12: Determine at least one feature that affects the battery life from the features in multiple dimensions.

[0081] Specifically, it can be to determine one or more features that can significantly affect the battery life from the features in multiple dimensions. It can be understood as determining the importance of the influence degree of each feature on the battery life. In other words, it can be to judge whether the importance of the influence degree of each feature on the battery life is significantly improved.

[0082] In this embodiment, the judgment of whether the importance is significantly improved can be achieved through the following steps: S120: Determine the initial weights of each feature in the multi-dimensional features.

[0083] For the initial weights of charging features, discharging features, temperature features, voltage features, usage pattern features, and efficiency and life features, they can be obtained through preliminary experiments.

[0084] Specifically, for each feature, a separate independent experiment is conducted, that is, when only considering the change of this feature, other features are kept unchanged. By conducting multiple standard charge-discharge cycle experiments on the change of this feature and comparing the remaining life data of each group, the influence weight of this feature can be preliminarily evaluated. For example, for the temperature feature, multiple standard charge-discharge cycle experiments are conducted under different temperature conditions while keeping other features unchanged. Finally, by comparing the remaining life of the experimental groups, the initial weights of each feature are preliminarily determined.

[0085] S121: At the initial stage of training, use the initial weights of all features to train a prediction model to obtain a preliminary prediction model (the first prediction model).

[0086] S122: Use the above-trained preliminary prediction model to predict the remaining life of each battery cell and calculate the prediction error of the model. The prediction error needs to be determined by the mean squared error (MSE) and the mean absolute error (MAE). The formulas are as follows: , ; where is the actual remaining life value, is the predicted remaining life value, n is the number of samples. In this step, based on the currently calculated prediction error and is used as the benchmark error.

[0087] S123: Select a feature and, based on the initial weight, adjust the weight of this feature individually. During the adjustment process, keep the weights of other features (i.e., keep the initial weights) unchanged.

[0088] For example, the weight of this feature can be increased by multiplying it by a control factor γ, and the calculation formula is as follows: ; where represents the adjusted weight, represents the initial weight corresponding to this feature; γ represents the control factor. Generally, the control factor is a preset constant.

[0089] Train the preliminary prediction model obtained above using the adjusted weight to obtain a new battery remaining life prediction model (the second prediction model).

[0090] S124: Use the second prediction model for prediction, calculate the adjusted prediction error to obtain a new error value, and the calculation formula is the same as S122. Compare the changes in the errors before and after adjustment (including MSE and MAE), and calculate the error change amount: , ; To determine whether the influence of the feature on the battery life is significant, set a threshold ϵ for the error change. For example, ϵ = 0.05. Only when the error change amount is greater than this threshold can it be determined that the weight adjustment of this feature is effective, and it shows that the influence degree or importance of this feature on the battery life has increased significantly.

[0091] If and , it is considered that the contribution of this feature to the model prediction result has increased significantly. Otherwise, keep the weight of this feature unchanged, and refer to the above steps S123 and S124 to adjust other features separately to determine whether there are other features that have a significant influence on the life or the importance has increased.

[0092] S13: Dynamically adjust the weights corresponding to at least one feature.

[0093] After determining the features whose weights need to be adjusted through S12, update their weights.

[0094] The cloud data platform synchronously and real-time receives data from the single monitoring module and the user data monitoring module uploaded by the main control module and repeats step S11, aiming to track the behavior changes of the user and the battery status. Specifically, the main control module can upload data to the cloud data platform periodically. For example, one cycle can be one week.

[0095] After uploading the data of the latest cycle and extracting the features of the latest multiple dimensions, the weighted gradient descent algorithm is adopted to automatically adjust the weights of the features during the model training process.

[0096] Specifically, after it is determined through the above S12 that the importance of a certain feature (such as the charging frequency) has been significantly improved (for example, the error has been significantly reduced), the system will, according to the preset rules, for example, the preset rules can be multiplying by a gain factor, to increase the weight of this feature. The purpose of weight adjustment is to dynamically respond to feature changes. When the system monitors that the importance of a certain feature (such as the charging frequency) has been significantly improved, it may be because the charging behavior of the user has changed. For example, if the user has been charging frequently recently, this indicates that the charging frequency may have a stronger correlation with the battery life. By multiplying by a gain factor, the weight of the charging frequency can be quickly increased, making the model pay more attention to this feature during training.

[0097] The gain factor can be determined using the following formula: ; Wherein, the current error is the prediction error corresponding to the life value obtained by inputting the weighted feature vector obtained by weighted fusion calculation of the features of multiple dimensions extracted from the data uploaded in the previous cycle into the trained prediction model (such as the first prediction model). The new error is the prediction error obtained by inputting the weighted feature vector obtained by weighted fusion calculation of the features of multiple dimensions extracted from the data uploaded in the latest cycle into the trained prediction model (such as the first prediction model).

[0098] The multiplication factor can be a preset constant. For example, the multiplication factor is set to 1.5.

[0099] After determining the gain factor, the contribution of the feature to the model prediction can be evaluated in one cycle, and the weight can be dynamically adjusted.

[0100] For example, the weight of this feature is increased by multiplying by a gain factor. After the feature weight is adjusted, continue to use the updated feature weight to retrain the model to obtain the trained third prediction model. The training process will further optimize the parameters of the model, making the model more accurately capture the influence of important features on the battery life prediction.

[0101] S14: Perform weighted fusion calculation based on the dynamically adjusted weights and features of multiple dimensions to obtain a weighted feature vector.

[0102] Here, the weighted fusion algorithm proposed in the embodiments of this application is elaborated in detail: The raw data uploaded by the main control module in the battery management device, including information such as the health status of the battery, historical operation data, and user usage habits, after extracting features of multiple dimensions from it, perform weighted fusion calculation according to the weights corresponding to each feature to obtain a weighted feature vector. Among them, the weights corresponding to some features may be the weights after dynamic adjustment according to the above steps, and the weights corresponding to the remaining features may be initial weights.

[0103] After the processing of multi-source data fusion, the weighted feature vector will more comprehensively and accurately reflect the overall condition and future trend of the battery.

[0104] For example, the system will first extract all relevant feature data from the user behavior and battery status monitoring module. For example, the charging frequency , temperature , depth of discharge , etc. Assume there are n features in the system (such as charging frequency, temperature, depth of discharge, etc.). Then, according to the foregoing steps, dynamically adjust the weight of each feature.

[0105] For each feature , multiply it by the corresponding weight to obtain a weighted feature × . Among them, may be the weight after dynamic adjustment or the initial weight. The final weighted feature vector is: ; where is the weight of each feature (which can be dynamically adjusted), and is the value of the corresponding feature. Through this weighted fusion, the system can generate a comprehensive feature vector (i.e., weighted feature vector) containing features of multiple dimensions.

[0106] Then, the weighted feature vector will be input into the constructed prediction model. For example, the prediction model can be a model obtained based on a deep learning framework. The prediction model will be trained and optimized based on these multi-dimensional features.

[0107] The initialization parameters in the deep learning model are set according to the characteristics of different data sources, and can gradually learn and identify the complex relationships between battery life and various characteristics. Finally, the set output variable is the remaining life of each single battery, and the deep learning model is trained, verified and tested. When the deviation between the remaining life value of a single battery output under more than 99% of the parameters and the actual remaining life value of a single battery is less than or equal to 2%, it proves that the model training is completed; if the deviation value is too large or the compliance rate is less than 99%, the model training parameters are adjusted and retrained.

[0108] Exemplarily, the specific training process is as follows: After obtaining the weighted feature vector, forward propagation is performed. The purpose of forward propagation is to input the data into the neural network and calculate the output of each layer, and finally obtain the prediction result of the model. This step involves calculations through each layer of the network until the last output layer.

[0109] Specifically, the input layer receives the dynamically weighted feature vector , as input, after each feature is multiplied by its corresponding weight , it is passed to the next layer (hidden layer).

[0110] Each neuron in each layer of the hidden layer calculates the weighted sum and performs a non-linear transformation through the activation function. For the th layer, ; where is the activation function, and the ReLU function is selected here, that is, . The ReLU activation function can effectively avoid the problem of gradient disappearance. is the output of the previous layer, is the bias term, is the weight matrix of the current layer. Transmitted to the output layer, in the regression task, the output layer usually contains one neuron, and the predicted value of the remaining battery life is output. Finally, the output of the model is the prediction of the remaining battery life. Then the loss function is calculated, and the role of the loss function is to measure the difference between the prediction result of the model and the actual result.

[0111] Here, the loss function can use the mentioned in the above steps, where is the predicted value of the th sample, is the true value of the th sample, is the number of samples. Subsequently, backpropagation is performed. Backpropagation is a core part of training a neural network. By using the backpropagation algorithm, the gradients of the loss function with respect to each parameter weight are calculated, and these gradients are used to update the model parameters.

[0112] The backpropagation process can include the following steps: Calculate the gradient of the loss function with respect to the activation of the output layer. For each output node , the gradient calculation formula is: ; where is the loss function, is the predicted value, is the true value.

[0113] Then, calculate the gradients of each hidden layer in turn. Through the chain rule, the backpropagation algorithm passes the error from the output layer to each layer: ; where is the gradient of the loss function with respect to the output of the th layer, is the gradient of the hidden layer output with respect to the weight.

[0114] Use the gradients calculated by backpropagation with the gradient descent optimization algorithm to update the weights of the model. The update rule of gradient descent is: ; is the learning rate.

[0115] The last step of the training process is to update the model parameters through gradient descent to complete one round of training. Then continue with the next iteration until the loss function converges or reaches the preset maximum number of iterations. At the end of each epoch, the model dynamically adjusts the weights of the features according to the changes in the loss function and each feature to ensure that the model can adapt to the changes in the data and user behavior. When the training loss of the model reaches the predetermined threshold or after multiple epochs, the model will stop training. At this time, all the weights (including the dynamically adjusted feature weights) have been optimized, and the model can accurately predict the remaining life of the battery or other target variables.

[0116] S15: Input the weighted feature vector into the trained prediction model to obtain the life prediction results corresponding to each battery cell.

[0117] After obtaining the life prediction results (i.e., remaining life values) of each single battery, the cloud data platform outputs the remaining life values of each battery and then transmits this key information to the main control module. The main control module intelligently adjusts the charging strategy according to the remaining life differences of different batteries. For example, when the remaining life of a certain battery in the battery pack is short, the system will set a dedicated charging plan for it, perhaps reducing the charging current or controlling the charging duration to reduce the load on the battery and delay the aging process.

[0118] Meanwhile, the main control module also dynamically adjusts the charging balance of the entire battery pack to ensure that, without affecting the overall performance, the usage time of the battery with a shorter life is extended. This precise charging strategy based on the health status of single batteries can effectively improve the overall life and efficiency of the battery pack, reduce maintenance costs, and enhance the safety and stability of the system.

[0119] The existing charging strategy technologies for series-connected lead-acid vehicle-mounted batteries focus on extracting single or limited features, with relatively single feature combinations and unable to fully extract features. The method proposed in the embodiments of this application uses a method of multi-source data fusion to extract features, which can extract and make full use of complex features such as charging frequency, depth of discharge, and battery voltage fluctuations and temperature change rates from more dimensions, incorporating the important module of user behavior data.

[0120] In addition, the existing charging strategy technologies for series-connected lead-acid vehicle-mounted batteries largely ignore the impact of user behavior and usage patterns on battery life. However, the multi-source data fusion method proposed in the embodiments of this application can dynamically analyze the recent feature data of users, clearly consider user behavior data, combine it with battery performance data, and propose a user-based personalized mode strategy, thereby more accurately predicting battery life and optimizing the charging strategy.

[0121] Further, after feature extraction is completed, the platform uses a multi-source data fusion algorithm to combine and process the data features from different monitoring modules. The multi-source data fusion algorithm can achieve comprehensive analysis of data at multiple levels, including the data level, feature level, and decision level.

[0122] The embodiments of this application also provide an electronic device, which includes: a processor, and the processor is used to execute computer programs or instructions in a memory to implement the method described in any of the above embodiments.

[0123] It should be noted that the processor can be a chip with computing capabilities, and is not limited to a Central Processing Unit (CPU). For example, the processor can be a chip that includes one or more transistors, resistors, capacitors, and other circuit elements for implementing certain functions; or it can be an integrated circuit in various packages that can implement the above methods.

[0124] The embodiment of the present application also provides a chip system, including: a communication interface for inputting and / or outputting data; a processor for executing computer-executable programs, so that a device installed with the chip system executes the method described in any of the above embodiments.

[0125] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center in a wired (such as coaxial cable, optical fiber, digital subscriber line) or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that the computer can access or a data storage device such as a server or data center that includes one or more integrated available media. The available medium can be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a Solid State Disk).

[0126] The above are only the preferred embodiments of the present application and are not used to limit the present application. For those skilled in the art, the present application can have various changes and modifications. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for predicting battery life, characterized in that: The battery comprises a plurality of battery cells, and the method comprises: Receiving battery pack data and user behavior data; the battery pack data includes charging data and discharging data of each battery cell in the plurality of battery cells; the user behavior data includes behavior data of the user charging the battery; Obtaining features of multiple dimensions corresponding to each of the multiple battery cells according to the battery pack data and the user behavior data; Determining at least one feature that affects battery life from the features in the multiple dimensions; Dynamically adjusting a weight corresponding to the at least one feature; According to the dynamically adjusted weights and features of multiple dimensions, weighted fusion calculation is performed to obtain a weighted feature vector; The weighted feature vector is input into a trained prediction model to obtain a life prediction result corresponding to each battery cell.

2. The method according to claim 1, characterized in that Determining at least one feature that affects battery life from the features in the multiple dimensions includes: Determine initial weights corresponding to the features of the multiple dimensions respectively, and obtain a first weighted feature vector according to the initial weights and the values ​​of the features of the multiple dimensions; use the first weighted feature vector as training data to train a prediction model to obtain a first prediction model; Obtaining a first life prediction result corresponding to each battery cell through the first prediction model; calculating a first prediction error of the first prediction model according to the first life prediction result and the actual life corresponding to each battery cell; Adjusting a first initial weight corresponding to a first feature among the features of the multiple dimensions to obtain a first weight; the weights corresponding to the features of other dimensions remain unchanged; Calculate the adjusted second forecast error; The variation between the first prediction error and the second prediction error is calculated; and when the variation satisfies a preset condition, the first feature is determined to be a feature that affects the battery life.

3. The method according to claim 2, characterized in that Adjusting a first initial weight corresponding to a first feature among the features of the multiple dimensions includes: According to the following calculation formula, the first initial weight corresponding to the first feature in the features of the multiple dimensions is adjusted: ； in, represents the adjusted first weight, represents the first initial weight corresponding to the first feature; γ represents a control factor, and the control factor is a preset constant.

4. The method according to claim 2, characterized in that: Calculate the adjusted second forecast error, including: Perform weighted fusion calculation based on the adjusted first weight and the initial weights corresponding to the features of other dimensions to obtain a second weighted feature vector; Using the second weighted feature vector, continue to train the first prediction model to obtain a second prediction model; Obtaining a second life prediction result corresponding to each battery cell through the second prediction model; A second prediction error of the second prediction model is calculated based on the second life prediction result and the actual life.

5. The method according to claim 2, characterized in that: Dynamically adjusting a weight corresponding to the at least one feature includes: Determine a gain factor corresponding to each of the at least one feature; At least one weight corresponding to the at least one feature is multiplied by a corresponding gain factor to obtain an adjusted weight.

6. The method according to claim 5, characterized in that Receive battery pack data and user charging behavior data uploaded by the battery management device, including: Receive battery pack data of different cycles and user charging behavior data uploaded by a battery management device; According to the battery pack data and the user charging behavior data, features of multiple dimensions corresponding to each of the multiple battery cells are obtained, including: According to the battery pack data and the user charging behavior data, characteristics of multiple dimensions corresponding to each battery cell in different cycles are obtained; After obtaining the features of multiple dimensions corresponding to different periods, the method further includes: According to the characteristics of multiple dimensions corresponding to different cycles, the life prediction results of each battery unit corresponding to different cycles are obtained; Determining a gain factor corresponding to each of the at least one feature includes: According to the life prediction results corresponding to the different cycles, the difference between the prediction error corresponding to the latest cycle and the prediction error corresponding to the previous cycle is calculated; The gain factors corresponding to the at least one feature are determined according to the ratio of the difference to the prediction error corresponding to the previous cycle, and a multiplication factor; wherein the multiplication factor is a preset constant.

7. The method according to claim 6, characterized in that Determining the gain factors corresponding to the at least one feature respectively according to the ratio of the difference to the prediction error corresponding to the previous cycle and the multiplication factor, including: Calculate the ratio of the difference to the prediction error corresponding to the previous cycle; calculate the product of the ratio and the multiplication factor, and use the calculation result as the value of the gain factor corresponding to a corresponding one of the at least one feature.

8. The method according to any one of claims 1 to 7, characterized in that The features of the multiple dimensions include at least one of the following features: Charging characteristics, discharging characteristics, temperature characteristics, voltage characteristics, usage behavior characteristics, efficiency and life characteristics; The charging characteristics include at least one of charging frequency, average charging time, and average charging power; wherein the charging frequency is determined according to the number of times the user charges within a predetermined time period; the average charging time is obtained by averaging the charging time of at least one charging of the user within the predetermined time period; and the average charging power is obtained by averaging the charging power of at least one charging of the user within the predetermined time period; The discharge characteristics include at least one of an average discharge depth, a discharge duration, and an average discharge power; wherein the average discharge depth is obtained by averaging the discharge depths corresponding to at least one charge of the user within a predetermined time; the discharge duration is obtained by summing the discharge durations of at least one discharge within a predetermined time; and the average discharge power is obtained by averaging the discharge power of at least one discharge within a predetermined time. Temperature characteristics, including ambient temperature variation characteristics and / or battery temperature variation characteristics; wherein the ambient temperature variation characteristics are calculated based on the difference between the maximum and minimum values ​​of the ambient temperature collected within a predetermined time period; the battery temperature variation characteristics are calculated based on the difference between the maximum and minimum values ​​of the battery temperature collected within a predetermined time period; Voltage characteristics, including voltage fluctuation characteristics and / or voltage stability characteristics; wherein the voltage fluctuation characteristics are obtained by calculating the standard deviation of voltage values ​​collected multiple times within a predetermined time period and the average voltage value; the voltage stability characteristics are obtained by calculating the ratio of the difference between the maximum and minimum voltage values ​​collected within a predetermined time period to the average voltage; Usage behavior characteristics include charging periodicity characteristics and / or discharging periodicity characteristics; the charging periodicity characteristics are determined according to the frequency of charging performed by the user within a predetermined time period; the discharging periodicity characteristics are determined according to the frequency of discharging performed by the user within a predetermined time period; Efficiency and life characteristics, including charge and discharge efficiency and / or total cycle number; wherein, the charge and discharge efficiency is obtained by calculating the ratio of the charging energy to the discharging energy; the total cycle number refers to the sum of the number of charges and / or discharges experienced by the battery.

9. A battery management system, characterized in that: Includes battery management device and cloud data platform; The battery management device includes a main control module, multiple single-cell monitoring modules, and a user data monitoring module; wherein the single-cell monitoring module is used to collect charging data and discharging data of each of the multiple battery cells of the battery to obtain battery pack data; the user data monitoring module is used to monitor the user's behavior data of charging the battery to obtain user behavior data; the main control module is used to upload the battery pack data and the user behavior data to a cloud data platform; The cloud data platform includes a server, and the server is used to implement the method as described in any one of claims 1-8.

10. An electronic device, characterized in that: The electronic device comprises: a processor, wherein the processor is configured to execute a computer program or instruction in a memory to implement the method according to any one of claims 1 to 8.

11. A computer program product, characterized in that The computer program product comprises a program, and when the program is executed by an electronic device, the electronic device implements the method according to any one of claims 1 to 8.

Images