A battery health analysis and prediction method, system, electronic device and storage medium

Abstract

The present application relates to a battery health analysis and prediction method, system, electronic device and storage medium, which includes identifying battery characteristic influencing parameters and environmental influencing parameters, building a primary analysis and prediction model based on a neural network algorithm; acquiring historical environmental parameters and historical battery characteristic parameters, adjusting the primary analysis and prediction model with the historical environmental parameters and the historical battery characteristic parameters to obtain a basic analysis and prediction model; monitoring the battery based on the basic analysis and prediction model, and detecting whether the number of battery charging times reaches a preset value; when it is detected that the number of battery charging times reaches a preset value, obtaining user habit influencing parameters; adjusting the basic analysis and prediction model based on the user habit influencing parameters to obtain a target analysis and prediction model; monitoring the battery based on the target analysis and prediction model, and outputting battery performance analysis and prediction results. The present application has the effect of improving the accuracy of battery health analysis and prediction.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to a battery health analysis and prediction method, system, electronic device and storage medium. Background Art

[0002] The battery is an important part of the car, used for vehicle starting, lighting and ignition, etc. At present, most cars are equipped with battery sensors and intelligent car networking application battery management system (BMS), so that users can check the remaining battery power information in time.

[0003] However, the performance of the battery changes with time and usage conditions, and the remaining capacity of the battery will gradually decrease with the number of charge and discharge cycles. The existing battery health analysis and prediction methods are generally based on the measurement and calculation of battery characteristic parameters such as the voltage, current, SOC, etc. of the battery cell, and the data is input into the battery health analysis and prediction model, and then the remaining capacity of the battery is estimated, thereby predicting the performance and life of the battery. However, the prediction of battery performance and life by only estimating the remaining capacity of the battery is not accurate. Therefore, it is necessary to develop a battery health analysis method, system and storage medium. Summary of the invention

[0004] The purpose of this application is to provide a battery health analysis and prediction method, system, electronic device and storage medium, which have the characteristics of improving the accuracy of battery health analysis and prediction.

[0005] In a first aspect, the present application provides a battery health analysis and prediction method, which adopts the following technical solution:

[0006] A battery health analysis and prediction method, comprising:

[0007] Obtain battery characteristic influencing parameters and environmental influencing parameters, and build a primary analysis and prediction model based on the neural network algorithm;

[0008] Acquiring historical environmental parameters and historical battery characteristic parameters, wherein the historical environmental parameters and the historical battery characteristic parameters are used to adjust the primary analysis and prediction model to obtain a basic analysis and prediction model;

[0009] Monitor the battery based on the basic analysis and prediction model, and detect whether the battery charging times reaches a preset value;

[0010] When it is detected that the number of battery charging times reaches a preset value, the user habit influence parameter is obtained;

[0011] Adjusting the basic analysis and prediction model based on the user habit influencing parameter to obtain a target analysis and prediction model;

[0012] The battery is monitored based on the target analysis and prediction model, and the battery performance analysis and prediction results are output, wherein the battery performance analysis and prediction results include capacity decay rate and life prediction.

[0013] By adopting the above technical solution, a basic analysis and prediction model is first set up based on historical data to detect the battery. After detecting that the number of battery charging times reaches the preset value, a sufficient number of charging times is collected, and the user habit influencing parameter data can be further used to make adjustments on the basic analysis and prediction model with strong generalization, so that the model is more personalized. At the same time, the accuracy of the analysis and prediction is improved according to the user's usage habits. However, if the amount of charging times data does not reach the preset value, the user habits are used to influence the parameters, which will not only easily cause inaccurate output of the battery performance analysis and prediction results, but also increase the amount of system processing in the early stage.

[0014] Optionally, the step of obtaining battery characteristic influencing parameters and environmental influencing parameters and constructing a primary analysis and prediction model based on a neural network algorithm includes:

[0015] A conditional trigger node is added to the input layer of the primary analysis and prediction model based on the neural network algorithm, and the trigger condition of the conditional trigger node is detecting that the number of battery charging times reaches a preset value.

[0016] By adopting the above technical solution, additional nodes are added to the input layer of the primary analysis and prediction model, so that when the model is subsequently monitored and used, when it is detected that the number of battery charges reaches a preset value, it will be triggered to input data in the input layer. When it is detected that the number of battery charges does not reach the preset value, no data is input, and the input feature is kept blank, which does not affect the result.

[0017] Optionally, the step of adjusting the basic analysis and prediction model based on the user habit influencing parameter to obtain a target analysis and prediction model includes:

[0018] Score the user habit influencing parameters to obtain the user habit score parameters;

[0019] Construct the objective function based on the user habit score parameter;

[0020] The numerical value calculated by the objective function is added as an additional feature to the conditional trigger node of the input layer of the basic analysis and prediction model as an input feature, so as to adjust the basic analysis and prediction model and obtain the target analysis and prediction model.

[0021] By adopting the above technical solution, the parameters affecting user habits are scored, which is helpful to simplify data analysis, extract key information from large amounts of data, reduce data dimensions, and improve data processing efficiency.

[0022] Optionally, the step of constructing an objective function according to the user habit score parameter includes:

[0023] Construct loss function and reward function based on user habit score parameters;

[0024] The objective function is constructed by combining the loss function and the reward function.

[0025] By adopting the above technical solution, the objective function is constructed by combining the loss function and the reward function, which can better measure the difference between the model analysis prediction results and the real labels from the two levels of the shortcomings and / or goodness of user usage habits. Therefore, the model parameters can be adjusted by optimizing the objective function to better fit the data, making the model more adaptable and robust.

[0026] Optionally, the step of performing scoring processing on the user habit influencing parameter to obtain the user habit score parameter includes:

[0027] The user habit influencing parameters include charging frequency, discharge depth, and charging time. The user habit influencing parameters are scored according to preset scoring rules to obtain user habit score parameters;

[0028] The preset scoring rule is that each data item of the user habit influencing parameter is set with multiple levels, the number of multiple levels set for each data item of the user habit influencing parameter is the same, and each level corresponds to a score.

[0029] By adopting the above technical solution, the number of levels of the three types of data, namely, charging frequency, discharge depth, and charging time, is set to be consistent, which can unify the processing of different types of data in different units and improve the subsequent data processing rate.

[0030] Optionally, the step of constructing a loss function and a reward function according to the user habit score parameters includes: performing a preset threshold comparison on each user habit score parameter to screen out loss function construction parameters and reward function construction parameters;

[0031] The formula for constructing the loss function based on the loss function parameters is as follows:

[0032]

[0033] Wherein, L is the output result of the loss function, a1 is the weight coefficient of the charging frequency, x1[i] is the score parameter of the i-th charging frequency higher than the preset threshold, n1 is the number of score parameters of the charging frequency higher than the preset threshold, a2 is the weight coefficient of the discharge depth, x2[i] is the score parameter of the i-th discharge depth higher than the preset threshold, and n2 is the number of score parameters of the discharge depth higher than the preset threshold;

[0034] The reward function formula is as follows:

[0035]

[0036] Among them, R is the output result of the reward function, b1 is the weight coefficient of the charging frequency, x1[j] is the score parameter for the jth charging frequency not higher than the preset threshold, m1 is the number of score parameters for the charging frequency not higher than the preset threshold, b2 is the weight coefficient of the discharge depth, x2[j] is the score parameter for the jth discharge depth not higher than the preset threshold, m2 is the number of score parameters for the discharge depth not higher than the preset threshold, b3 is the weight coefficient of the charging time, x3[j] is the score parameter for the jth charging time not higher than the preset threshold, and m3 is the number of score parameters for the charging time not higher than the preset threshold.

[0037] By adopting the above technical solution, a preset threshold is compared for each user habit score parameter, and the loss function construction parameters and reward function construction parameters are screened out. The preset threshold has different values ​​for different parameters. Parameters higher than the preset threshold are defined as the case of bad user habits, which are used to construct the loss function. Conversely, parameters not higher than the preset threshold are defined as the case of good user habits, which are used to construct the reward function. For the loss function, the discharge depth and charging frequency are selected in the loss function for construction, which is also conducive to improving the processing rate. For the reward function, appropriately reducing the charging time so that the battery is not fully charged can effectively protect the battery. Therefore, the charging time score parameter not higher than the preset threshold can be used to construct the reward function to make the data more accurate.

[0038] Optionally, the step of constructing an objective function by combining the loss function and the reward function includes:

[0039] The objective function is constructed by combining the loss function and the reward function, where the objective function formula is as follows: J = α × L-β × R;

[0040] Among them, J is the output result of the objective function, α is the trade-off coefficient of the loss function, and β is the trade-off coefficient of the reward function.

[0041] By adopting the above technical solution, the loss function and the reward function are combined to construct the objective function. The goal is to minimize the battery life loss while maximizing the impact of good power usage habits. α and β are coefficients that weigh the loss and reward, and are used to adjust the influence of the loss function and the reward function in the objective function.

[0042] In a second aspect, the present application provides a battery health analysis and prediction system, which adopts the following technical solutions:

[0043] A battery health analysis and prediction system, comprising:

[0044] Preliminary model building module: used to obtain battery characteristic influencing parameters and environmental influencing parameters, and build a primary analysis and prediction model based on the neural network algorithm;

[0045] Basic model building module: used to obtain historical environmental parameters and historical battery characteristic parameters, and the historical environmental parameters and the historical battery characteristic parameters are used to adjust the primary analysis and prediction model to obtain a basic analysis and prediction model;

[0046] Basic data monitoring module: used to monitor the battery based on the basic analysis and prediction model, and detect whether the number of battery charging times reaches a preset value;

[0047] User parameter determination module: when it is detected that the battery charging times reaches the preset value, the user habit influencing parameters are obtained;

[0048] Target model building module: used to adjust the basic analysis and prediction model based on the user habit influencing parameters to obtain a target analysis and prediction model;

[0049] Target data monitoring module: used to monitor the battery based on the target analysis and prediction model, and output the battery performance analysis and prediction results, which include capacity decay rate and life prediction.

[0050] In a third aspect, the present application provides an electronic device, which adopts the following technical solution:

[0051] An electronic device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned battery health analysis and prediction method when executing the computer program.

[0052] In a fourth aspect, the present application provides a computer storage medium, which adopts the following technical solution:

[0053] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned battery health analysis and prediction method are implemented.

[0054] In summary, the beneficial technical effects of this application include:

[0055] First, a basic analysis and prediction model is set up based on historical data to test the battery. After the number of battery charging times reaches the preset value, a sufficient number of charging times is collected. The user habit influencing parameter data can be further used to adjust the basic analysis and prediction model with strong generalization, making the model more personalized. At the same time, the accuracy of the analysis and prediction is improved according to the user's usage habits. However, if the amount of charging times data does not reach the preset value, the user habit influencing parameters will be used, which will not only easily cause inaccurate output of battery performance analysis and prediction results, but also increase the amount of system processing in the early stage. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a flowchart of a battery health analysis and prediction method in one embodiment of the present application.

[0057] Figure 2 It is a schematic diagram of the specific step flow of step S310 in another embodiment of the present application.

[0058] Figure 3 It is a schematic diagram of the specific step flow of step S3101 in another embodiment of the present application.

[0059] Figure 4 It is a structural diagram of a battery health analysis and prediction method according to one embodiment of the present application.

[0060] Figure 5 It is a principle block diagram of an electronic device in one embodiment of the present application.

[0061] In the figure, 1. Preliminary model building module; 2. Basic model building module; 3. Basic data monitoring module; 4. User parameter determination module; 5. Target model building module; 6. Target data monitoring module. DETAILED DESCRIPTION

[0062] In order to enable technicians in this field to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0063] In the description of the embodiments of the present application, words such as "for example" or "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "for example" or "for example" in the embodiments of the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "for example" or "for example" is intended to present related concepts in a specific way.

[0064] In the description of the embodiments of the present application, the meaning of the term "multiple" refers to two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "include", "comprise", "have" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0065] The present application embodiment provides a battery health analysis and prediction method. In one embodiment, please refer to Figure 1 , Figure 1 This is a flowchart of a battery health analysis and prediction method provided by an embodiment of the present application. The method can be implemented by a computer program, which can be integrated into an application or run as an independent tool application. The method can also be implemented by a single-chip microcomputer or run in a battery health analysis and prediction system. Specifically, the method can include the following steps:

[0066] S1. Obtain parameters affecting battery characteristics and environmental impact parameters, and build a primary analysis and prediction model based on a neural network algorithm.

[0067] Among them, the parameters affecting battery characteristics include battery voltage and current, initial capacity of the battery, internal resistance, chemical composition, manufacturing date and charge and discharge cycle data, temperature, humidity and other parameters. The remaining capacity of the battery can be estimated by voltage and current. The remaining capacity of the battery can be compared with the capacity of a new battery to reflect the health of the battery. The internal resistance of the battery can be calculated by measuring the open circuit voltage (OCV) and load voltage of the battery. Environmental influencing parameters include temperature, humidity and altitude. The various parameter data of the battery can be monitored in real time through BMS and external sensors. Then, these raw data are preprocessed, such as filtering and denoising, interpolation processing, etc. Corresponding cleaning and filling processing are required to eliminate the interference of these abnormal data on the model. Then, the recurrent neural network (RNN) model is trained with the battery characteristic influencing parameters and the environmental influencing parameter data set to obtain a primary analysis and prediction model.

[0068] S2. Obtain historical environmental parameters and historical battery characteristic parameters, adjust the primary analysis and prediction model with the historical environmental parameters and the historical battery characteristic parameters, and obtain a basic analysis and prediction model.

[0069] Among them, the performance of the trained primary analysis and prediction model is evaluated using the historical environmental parameters and historical battery characteristic parameter sets of other similar batteries. According to the evaluation results, the hyperparameters or structure of the primary analysis and prediction model are adjusted to improve the generalization ability and prediction accuracy of the model, and the basic analysis and prediction model is obtained. The basic analysis and prediction model is adjusted based on the past data of similar batteries and has generalization.

[0070] S3. Monitor the battery based on the basic analysis and prediction model, and detect whether the battery charging times reaches a preset value.

[0071] S30: When it is detected that the battery charging times do not reach the preset value, the battery performance analysis prediction result is output.

[0072] S31. When it is detected that the number of battery charging times reaches a preset value, a user habit influencing parameter is obtained.

[0073] Among them, the battery is monitored based on the basic analysis and prediction model, and the number of battery charging times is detected to see if it has reached the preset value. This step is to collect enough usage times, so as to further adjust the basic analysis and prediction model with strong generalization, so that the model is more personalized and the accuracy of analysis and prediction is improved according to user usage habits. If the amount of data does not reach the preset value, the user's habits are used to influence the parameters, which will not only easily lead to inaccurate output of battery performance analysis and prediction results, but also increase the amount of system processing in the early stage.

[0074] S310: Adjust the basic analysis and prediction model based on the user habit influencing parameters to obtain a target analysis and prediction model.

[0075] Among them, after the user uses the battery for a period of time, the user habit influencing parameters are collected. By using the user habit influencing parameters as the input features of the model to adjust the basic analysis and prediction model, personalized analysis and prediction can be performed for each user, and the accuracy of the analysis and prediction can also be improved.

[0076] S311. Monitor the battery based on the target analysis prediction model and output the battery performance analysis prediction results, which include capacity decay rate and life prediction.

[0077] Among them, the battery performance analysis prediction results mainly include the battery capacity decay rate and life prediction, which is conducive to users to manage battery use, so as to adjust their battery usage habits and thus improve the battery life.

[0078] Based on the above embodiment, as an optional embodiment, step S1: obtaining battery characteristic influencing parameters and environmental influencing parameters, and building a primary analysis and prediction model based on a neural network algorithm, further includes:

[0079] A conditional trigger node is added to the input layer of the primary analysis and prediction model based on the neural network algorithm. The trigger condition of the conditional trigger node is to detect that the number of battery charging times reaches a preset value.

[0080] Specifically, by adding additional nodes to the input layer of the primary analysis and prediction model, when the model is subsequently monitored and used, when it is detected that the number of battery charges has reached a preset value, it will be triggered to input data at the input layer. When it is detected that the number of battery charges has not reached the preset value, no data will be input, and the input feature will be kept blank, which will not affect the results.

[0081] Reference Figure 2 Based on the above embodiment, as an optional embodiment, step S310: adjusting the basic analysis prediction model based on the user habit influencing parameter to obtain the target analysis prediction model includes:

[0082] S3100, score the user habit influencing parameter to obtain the user habit score parameter;

[0083] S3101, constructing an objective function according to the user habit score parameter;

[0084] S3102. Add the numerical value calculated by the objective function as an additional feature to the conditional trigger node of the input layer of the basic analysis and prediction model as an input feature, so as to adjust the basic analysis and prediction model and obtain the target analysis and prediction model.

[0085] Among them, scoring the parameters that affect user habits helps to simplify data analysis, extract key information from large amounts of data, reduce data dimensions, and improve data processing efficiency.

[0086] Reference Figure 3 Based on the above embodiment, as an optional embodiment, step S3101: constructing an objective function according to the user habit score parameter includes:

[0087] S030. Construct a loss function and a reward function according to the user's habit score parameters;

[0088] S031. Construct the objective function by combining the loss function and the reward function.

[0089] Among them, the objective function is constructed by combining the loss function and the reward function, which can better measure the difference between the model analysis prediction results and the real labels from the two levels of the shortcomings and / or goodness of user usage habits. Therefore, the model parameters can be adjusted by optimizing the objective function to better fit the data, making the model more adaptable and robust.

[0090] Based on the above embodiment, as an optional embodiment, step S3100: performing scoring processing on the user habit influencing parameter to obtain the user habit score parameter includes:

[0091] The user habit influencing parameters include charging frequency, discharge depth, and charging time. The user habit influencing parameters are scored according to preset scoring rules to obtain user habit score parameters;

[0092] Among them, the preset scoring rule is that multiple levels are set for each data of the user habit influencing parameter, the number of multiple levels set for each data of the user habit influencing parameter is consistent, and each level corresponds to a score.

[0093] As mentioned above, the charging frequency, the discharge depth and the charging time are all detected within a period of time. Therefore, there are multiple levels of these three types of data. By setting the number of levels of these three types of data to be consistent, different types of data in different units can be processed in a unified manner, thereby improving the subsequent data processing rate.

[0094] Based on the above embodiment, as an optional embodiment, step S030: constructing a loss function and a reward function according to the user habit score parameter includes:

[0095] Compare each user habit score parameter with a preset threshold to filter out the loss function construction parameters and reward function construction parameters;

[0096] The formula for constructing the loss function based on the loss function parameters is as follows:

[0097]

[0098] Wherein, L is the output result of the loss function, a1 is the weight coefficient of the charging frequency, x1[i] is the score parameter of the i-th charging frequency higher than the preset threshold, n1 is the number of score parameters of the charging frequency higher than the preset threshold, a2 is the weight coefficient of the discharge depth, x2[i] is the score parameter of the i-th discharge depth higher than the preset threshold, and n2 is the number of score parameters of the discharge depth higher than the preset threshold;

[0099] The reward function formula is as follows:

[0100]

[0101] Among them, R is the output result of the reward function, b1 is the weight coefficient of the charging frequency, x1[j] is the score parameter for the jth charging frequency not higher than the preset threshold, m1 is the number of score parameters for the charging frequency not higher than the preset threshold, b2 is the weight coefficient of the discharge depth, x2[j] is the score parameter for the jth discharge depth not higher than the preset threshold, m2 is the number of score parameters for the discharge depth not higher than the preset threshold, b3 is the weight coefficient of the charging time, x3[j] is the score parameter for the jth charging time not higher than the preset threshold, and m3 is the number of score parameters for the charging time not higher than the preset threshold.

[0102] In the above, the preset threshold value is compared for each user habit score parameter to screen out the loss function construction parameters and the reward function construction parameters. The preset threshold value has different values ​​for different parameters. Parameters higher than the preset threshold value are defined as the case of bad user habits and are used to construct the loss function. Conversely, parameters not higher than the preset threshold value are defined as the case of good user habits and are used to construct the reward function.

[0103] Among them, for the loss function, since the battery life loss is related to the discharge depth and charging frequency, and has little correlation with the charging time, the existing battery automatically shuts down after being fully charged. Therefore, the long charging time actually has no correlation with the battery life. In order to avoid such situations as much as possible, the discharge depth and charging frequency are selected in the loss function for construction, which is also conducive to improving the processing rate. a1 and a2 are the weight coefficients of each power usage habit data in the loss function, reflecting the degree of influence of different power usage habits on battery life.

[0104] However, for the reward function, appropriately reducing the charging time and not fully charging the battery can effectively protect the battery. Therefore, the charging time score parameter that is not higher than the preset threshold can be used to construct the reward function. b1, b2, and b3 are the weight coefficients of various power usage habit data in the reward function, reflecting the positive impact of good power usage habits on battery life.

[0105] Based on the above embodiment, as an optional embodiment, step S031 of constructing an objective function by combining the loss function and the reward function may further include:

[0106] The objective function is constructed by combining the loss function and the reward function, where the objective function formula is as follows: J = α × L-β × R;

[0107] Among them, J is the output result of the objective function, α is the trade-off coefficient of the loss function, and β is the trade-off coefficient of the reward function.

[0108] In the above, the loss function and the reward function are combined to construct the objective function. The goal is to minimize the loss of battery life while maximizing the impact of good power usage habits. α and β are coefficients that weigh the loss and reward and are used to adjust the influence of the loss function and the reward function in the objective function.

[0109] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0110] The present application also provides a battery health analysis and prediction system, which corresponds one-to-one to the battery health analysis and prediction method in the embodiment. Figure 4 The battery health analysis and prediction system includes: a preliminary model building module 1, a basic model building module 2, a basic data monitoring module 3, a user parameter determination module 4, a target model building module 5 and a target data monitoring module 6. The detailed description of each functional module is as follows:

[0111] Preliminary model building module 1: used to obtain battery characteristic influencing parameters and environmental influencing parameters, and build a primary analysis and prediction model based on the neural network algorithm.

[0112] Basic model building module 2: used to obtain historical environmental parameters and historical battery characteristic parameters, which adjust the primary analysis and prediction model to obtain the basic analysis and prediction model;

[0113] Basic data monitoring module 3: used to monitor the battery based on the basic analysis and prediction model, and detect whether the battery charging times reach the preset value;

[0114] User parameter determination module 4: when it is detected that the battery charging times reaches a preset value, the user habit influencing parameter is obtained;

[0115] Target model building module 5: used to adjust the basic analysis and prediction model based on the user habit influencing parameters to obtain the target analysis and prediction model;

[0116] Target data monitoring module 6: used to monitor the battery based on the target analysis prediction model and output the battery performance analysis prediction results, which include capacity decay rate and life prediction.

[0117] Among them, the system can obtain battery characteristic influencing parameters and environmental influencing parameters through the preliminary model construction module 1, build a primary analysis and prediction model based on the neural network algorithm, and then use the basic model construction module 2 to obtain historical environmental parameters and historical battery characteristic parameters, adjust the primary analysis and prediction model with historical environmental parameters and historical battery characteristic parameters to obtain a basic analysis and prediction model, and then use the basic data monitoring module 3 to monitor the battery based on the basic analysis and prediction model, and detect whether the battery charging times reaches the preset value, and use the user parameter determination module 4 to determine whether the battery charging times reaches the preset value, and then obtain the user habit influencing parameters, and then use the target model construction module 5 to adjust the basic analysis and prediction model based on the user habit influencing parameters to obtain the target analysis and prediction model, and finally use the target data monitoring module 6 to monitor the battery based on the target analysis and prediction model, and output the battery performance analysis and prediction results, which include capacity decay rate and life prediction.

[0118] For the specific limitations of the battery health analysis and prediction system, please refer to the limitations of the battery health analysis and prediction method in the context, which will not be repeated here. Each module in the above-mentioned battery health analysis and prediction system can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the electronic device in the form of hardware, or can be stored in the memory of the electronic device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0119] In this embodiment, an electronic device is provided, and the electronic device is a computer. Figure 5 , the electronic device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the electronic device is used to store a detection data table. The network interface of the electronic device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a battery health analysis and prediction method is implemented.

[0120] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the following steps when executing the computer program:

[0121] S1. Obtain parameters affecting battery characteristics and environmental impact parameters, and build a primary analysis and prediction model based on a neural network algorithm.

[0122] S2. Obtain historical environmental parameters and historical battery characteristic parameters, adjust the primary analysis and prediction model with the historical environmental parameters and the historical battery characteristic parameters, and obtain a basic analysis and prediction model.

[0123] S3. Monitor the battery based on the basic analysis and prediction model, and detect whether the battery charging times reaches a preset value.

[0124] S30: When it is detected that the battery charging times do not reach the preset value, the battery performance analysis prediction result is output.

[0125] S31. When it is detected that the number of battery charging times reaches a preset value, a user habit influencing parameter is obtained.

[0126] S310: Adjust the basic analysis and prediction model based on the user habit influencing parameters to obtain a target analysis and prediction model.

[0127] S311. Monitor the battery based on the target analysis prediction model and output the battery performance analysis prediction results, which include capacity decay rate and life prediction.

[0128] Step S310 specifically includes the following steps:

[0129] S3100: Score the user habit influencing parameters to obtain user habit score parameters.

[0130] S3101. Construct an objective function according to user habit score parameters.

[0131] S3102. Add the numerical value calculated by the objective function as an additional feature to the conditional trigger node of the input layer of the basic analysis and prediction model as an input feature, so as to adjust the basic analysis and prediction model and obtain the target analysis and prediction model.

[0132] Step S3101 specifically includes the following steps:

[0133] S030. Construct a loss function and a reward function according to the user's habit score parameters;

[0134] S031. Construct the objective function by combining the loss function and the reward function.

[0135] An embodiment of the present application also discloses a computer-readable storage medium storing a computer program that can be loaded and executed by a processor. When the computer program is executed by the processor, the steps of any of the above-mentioned battery health analysis and prediction methods are implemented and the same effect can be achieved.

[0136] The computer-readable storage medium includes, for example, a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0137] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. A battery health analysis and prediction method, characterized in that: include: Obtain battery characteristic influencing parameters and environmental influencing parameters, and build a primary analysis and prediction model based on the neural network algorithm; Acquiring historical environmental parameters and historical battery characteristic parameters, wherein the historical environmental parameters and the historical battery characteristic parameters are used to adjust the primary analysis and prediction model to obtain a basic analysis and prediction model; Monitor the battery based on the basic analysis and prediction model, and detect whether the battery charging times reaches a preset value; When it is detected that the number of battery charging times reaches a preset value, the user habit influence parameter is obtained; The basic analysis and prediction model is adjusted based on the user habit influencing parameter to obtain a target analysis and prediction model, including: Score the user habit influencing parameters to obtain the user habit score parameters; Construct the objective function based on the user habit score parameter; The numerical value calculated by the objective function is added as an additional feature to the conditional trigger node of the input layer of the basic analysis and prediction model as an input feature, so as to adjust the basic analysis and prediction model and obtain the target analysis and prediction model; The battery is monitored based on the target analysis and prediction model, and the battery performance analysis and prediction results are output, wherein the battery performance analysis and prediction results include capacity decay rate and life prediction.

2. The method according to claim 1, characterized in that The step of obtaining battery characteristic influencing parameters and environmental influencing parameters and constructing a primary analysis and prediction model based on a neural network algorithm includes: A conditional trigger node is added to the input layer of the primary analysis and prediction model based on the neural network algorithm, and the trigger condition of the conditional trigger node is detecting that the number of battery charging times reaches a preset value.

3. The method according to claim 1, characterized in that: The step of constructing the objective function according to the user habit score parameter includes: Construct loss function and reward function based on user habit score parameters; The objective function is constructed by combining the loss function and the reward function.

4. The method according to claim 3, characterized in that: The step of scoring the user habit influencing parameter to obtain the user habit score parameter includes: The user habit influencing parameters include charging frequency, discharge depth, and charging time. The user habit influencing parameters are scored according to preset scoring rules to obtain user habit score parameters; The preset scoring rule is that each data item of the user habit influencing parameter is set with multiple levels, the number of multiple levels set for each data item of the user habit influencing parameter is the same, and each level corresponds to a score.

5. The method according to claim 4, characterized in that The step of constructing a loss function and a reward function according to the user habit score parameter includes: Compare each user habit score parameter with a preset threshold to filter out the loss function construction parameters and reward function construction parameters; The formula for constructing the loss function based on the loss function parameters is as follows: Wherein, L is the output result of the loss function, a1 is the weight coefficient of the charging frequency, x1[i] is the score parameter of the i-th charging frequency higher than the preset threshold, n1 is the number of score parameters of the charging frequency higher than the preset threshold, a2 is the weight coefficient of the discharge depth, x2[i] is the score parameter of the i-th discharge depth higher than the preset threshold, and n2 is the number of score parameters of the discharge depth higher than the preset threshold; The reward function formula is as follows: Among them, R is the output result of the reward function, b1 is the weight coefficient of the charging frequency, x1[j] is the score parameter for the jth charging frequency not higher than the preset threshold, m1 is the number of score parameters for the charging frequency not higher than the preset threshold, b2 is the weight coefficient of the discharge depth, x2[j] is the score parameter for the jth discharge depth not higher than the preset threshold, m2 is the number of score parameters for the discharge depth not higher than the preset threshold, b3 is the weight coefficient of the charging time, x3[j] is the score parameter for the jth charging time not higher than the preset threshold, and m3 is the number of score parameters for the charging time not higher than the preset threshold.

6. The method according to claim 5, characterized in that The step of constructing the objective function by combining the loss function and the reward function includes: The objective function is constructed by combining the loss function and the reward function. The objective function formula is as follows: J = α × L - β × R; Among them, J is the output result of the objective function, α is the trade-off coefficient of the loss function, and β is the trade-off coefficient of the reward function.

7. A battery health analysis and prediction system, characterized in that: include: Preliminary model building module (1): used to obtain battery characteristic influencing parameters and environmental influencing parameters, and build a primary analysis and prediction model based on a neural network algorithm; Basic model building module (2): used to obtain historical environmental parameters and historical battery characteristic parameters, wherein the historical environmental parameters and the historical battery characteristic parameters are used to adjust the primary analysis and prediction model to obtain a basic analysis and prediction model; Basic data monitoring module (3): used to monitor the battery based on the basic analysis and prediction model, and detect whether the number of battery charging times reaches a preset value; User parameter determination module (4): when detecting that the number of battery charging times reaches a preset value, obtaining the user habit influence parameter; A target model building module (5) is used to adjust the basic analysis and prediction model based on the user habit influencing parameters to obtain a target analysis and prediction model, including: Score the user habit influencing parameters to obtain the user habit score parameters; Construct the objective function based on the user habit score parameter; The numerical value calculated by the objective function is added as an additional feature to the conditional trigger node of the input layer of the basic analysis and prediction model as an input feature, so as to adjust the basic analysis and prediction model and obtain the target analysis and prediction model; Target data monitoring module (6): used to monitor the battery based on the target analysis prediction model and output the battery performance analysis prediction result, wherein the battery performance analysis prediction result includes capacity decay rate and life prediction.

8. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and executes a battery health analysis and prediction method as claimed in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer program is stored which can be loaded by a processor and execute the battery health analysis and prediction method according to any one of claims 1 to 6.

Images