A stress-based method for defining, evaluating and predicting the remaining life of a battery

Abstract

The application belongs to the field of power battery prediction, and provides a stress-based battery residual life definition, evaluation and prediction method. According to the capacity and internal resistance definition method, and the strong correlation between stress and battery life, the application proposes the definition of the stress-based residual service life of the power battery, further combines various definition methods such as battery stress, capacity, internal resistance and the like, performs weighted summation, comprehensively evaluate the residual life of the battery, and finally uses a model to take current, voltage, capacity, stress and the like as input to predict the residual life of the battery online.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of power battery prediction, and relates to a stress-based battery remaining life definition, evaluation and prediction method. BACKGROUND

[0002] The statements in this section merely provide background information related to the present application and do not necessarily constitute prior art.

[0003] Lithium-ion batteries have been widely used in electric vehicles and energy storage devices. A key but challenging problem is to predict the remaining useful life (RUL) of lithium-ion batteries. Under actual application conditions, the battery performance will gradually deteriorate with continuous charging and discharging until the end of life (EOL). After EOL, the capacity and power of the battery will decrease significantly. This will lead to operational difficulties and accidents, such as shutdown, fire, and even serious damage to the power system. Therefore, obtaining accurate RUL is of great significance to battery health management and safe and reliable operation of the system.

[0004] Currently, RUL is defined as the number of charge-discharge cycles or usage time when the available capacity drops to 80% of the initial value or the internal resistance reaches 1.5 times the initial internal resistance. Generally, the capacity or internal resistance is obtained by controlling the battery charge-discharge test, so as to obtain accurate RUL. However, this method cannot be measured in real time online, and is usually most suitable for operation in the laboratory. Therefore, the traditional RUL definition and evaluation method has certain limitations. How to redefine RUL based on new features or parameters and realize comprehensive evaluation and online prediction of RUL has become an important means to break through the key technical bottlenecks of power battery health management and related industry rapid development.

[0005] Correlation studies have shown that stress is a dynamic variable during battery operation, influenced by battery SOC and gradually increasing with the number of charge-discharge cycles. At the same time, as a mechanical parameter, stress can be directly measured in real time during charging and discharging, and is easy to operate. Some scholars have pointed out that during the charging and discharging process of lithium-ion batteries, Li+ will continuously intercalate and deintercalate in the positive and negative electrodes of the battery, causing changes in the volume of the electrode material, and the decomposition of the electrolyte to produce gas, resulting in mechanical deformation of the battery, changes in stress, and a decrease in performance and life. The increase in internal stress of the battery is strongly correlated with the capacity degradation of the battery, and the battery's charge-discharge capacity will gradually deteriorate as the stress increases, leading to life degradation. Many scholars have conducted extensive research on the definition and evaluation method of RUL, but the current definition and calculation method of lithium-ion battery RUL is relatively single, and cannot be predicted online. For example, Chinese invention patent CN 201810826685.3 estimates battery state of health and remaining useful life based on electrochemical impedance spectroscopy. This method only considers the relationship between battery internal resistance and RUL, but the internal resistance does not change significantly before RUL declines to EOL, and the accuracy is low. Chinese invention patent CN 201911289199.3 only considers battery capacity to define battery RUL, and the evaluation results are easily disturbed by external factors, with large errors and decreasing accuracy over time. Chinese invention patent CN 201310533862.6 defines RUL using capacity and internal resistance, respectively, and comprehensively evaluates the remaining useful life of the battery. However, this method does not consider the influence of internal parameters, and the accuracy and calculation efficiency are low. Chinese invention patent CN 202210735862.3 considers the influence of stress, estimates the state of charge (SOC) of the battery, and improves the accuracy of SOC estimation, providing a new approach for accurate RUL evaluation. Chinese invention patent CN 202010860206.7 improves the credibility of SOC estimation by establishing a battery stress model and an equivalent circuit model and using an extended Kalman filter to estimate SOC online, but the model is easily disturbed by external environmental factors, and different battery models have certain differences, with poor robustness. Chinese invention patent CN 202110681576.9 considers the influence of mechanical strain on battery capacity and uses it as a feature to predict RUL, but this method is still based on the capacity definition method of RUL and does not fundamentally explore the relationship between stress and battery life, with certain limitations. Chinese invention patent CN 201910784277.0 extracts multiple characteristic quantities of the battery IC curve to input a prediction model to evaluate battery life. However, this method has a complex calculation process and low efficiency. Chinese invention patent CN 202110669209.7 uses current, voltage, capacity, and electrolyte content during battery charging and discharging to comprehensively evaluate battery RUL, with high accuracy, but the acquisition of electrolyte content is difficult and the operation is complex. Although there have been many studies on RUL definition and evaluation, most of them are based on battery capacity or internal resistance, and there is a lack of analysis of the influence of battery stress and other characteristic parameters on battery life.At the same time, the existing method cannot predict RUL online, and the application is limited. SUMMARY

[0006] The present application proposes a stress-based definition, evaluation and prediction method of battery remaining life to solve the above problems. According to the capacity and internal resistance definition method, and the strong correlation between stress and battery life, the present application proposes a definition of stress-based power battery remaining life. Further, the present application combines various definition methods such as battery stress, capacity, internal resistance, etc. to perform weighted summation, comprehensively evaluates the battery remaining life, and finally uses a model to take current, voltage, capacity, stress, etc. as input to predict the battery remaining life online.

[0007] According to some embodiments, the present application adopts the following technical solutions:

[0008] A stress-based definition, evaluation and prediction method of battery remaining life includes the following steps:

[0009] Taking the battery stress, current and voltage data of the target power battery as input, the pre-trained model is used to obtain the predicted remaining life;

[0010] The training process of the model includes:

[0011] The battery capacity, internal resistance and battery stress of each charge and discharge obtained by the cycle charge and discharge experiment of the target battery are obtained, and the data is processed;

[0012] The battery remaining life is evaluated by comprehensively considering the battery capacity, internal resistance and battery stress;

[0013] The model is trained by taking the battery stress, current and voltage data as input and the battery remaining life as output.

[0014] As an optional implementation, the specific process of evaluating the battery remaining life by comprehensively considering the battery capacity, internal resistance and battery stress includes:

[0015] When the available capacity of the battery is lower than the set value of the initial capacity, the internal resistance increases to k times the initial internal resistance, and the stress increases to m times the initial stress, the cycle charge and discharge experiment is stopped, and the battery stress at this time is recorded, and k and m are positive numbers;

[0016] The stress data is analyzed, the correlation between the battery stress and the battery life is calculated, and the battery remaining life defined based on the stress is obtained;

[0017] Based on the capacity and internal resistance data, the corresponding battery remaining life is calculated, and the battery remaining life is comprehensively evaluated in combination with the battery remaining life defined based on the stress.

[0018] As further, considering the battery capacity, internal resistance and battery stress comprehensively, the battery remaining life is evaluated, the battery remaining life based on the stress, capacity and internal resistance definition method is comprehensively defined, the weight coefficient of each definition method is selected according to the application occasion, the weighted sum of the obtained different battery remaining life is obtained, and the comprehensive evaluation of the battery remaining life is realized.

[0019] As further, the sum of the weight coefficients of each definition method is one.

[0020] As further, when the weight coefficients of each definition method are selected according to the application occasion, TOPSIS, fuzzy evaluation, grey correlation, rank sum ratio or comprehensive index method is used to determine.

[0021] As an optional implementation, the model is a regularized linear regression network.

[0022] As further, a penalty term is added to the model to minimize the error, and a linear regression network based on lasso or elastic net technology is selected.

[0023] A power battery remaining life prediction system, comprising:

[0024] A prediction module configured to take the battery stress, current and voltage data of the target power battery as input, utilize the pre-trained model, and obtain the predicted remaining life.

[0025] A training module configured to obtain the battery capacity, internal resistance and battery stress of each charge and discharge obtained by the cycle charge and discharge experiment of the target battery, process the data, consider the battery capacity, internal resistance and battery stress comprehensively, evaluate the battery remaining life, take the battery stress, current and voltage data as input, and take the battery remaining life as output to train the model.

[0026] A computer readable storage medium, wherein a plurality of instructions are stored, the instructions are suitable for being loaded by a processor of a terminal device and executing the steps in the method.

[0027] A terminal device, comprising a processor and a computer readable storage medium, the processor is used to implement instructions, and the computer readable storage medium is used to store a plurality of instructions, the instructions are suitable for being loaded by the processor and executing the steps in the method.

[0028] Compared with the prior art, the beneficial effects of the present application are:

[0029] (1) The present application defines the battery RUL based on the battery stress, and the stress measurement is simple and efficient, which can accurately reflect the battery RUL in real time. At the same time, the stress 'inflection point' appears earlier, and the linearity of stress and cycle number after the 'inflection point' is stronger, and even the stress curve is considered to have no 'inflection point'. Therefore, it is simpler and more accurate to define RUL and evaluate and predict RUL based on stress;

[0030] (2) The present application realizes comprehensive evaluation of battery RUL based on various RUL definition methods of stress, capacity, internal resistance and the like, and the result is objective and accurate, not easy to be disturbed by external environment, and has strong robustness;

[0031] (3) The present application realizes online prediction of battery RUL based on regularized linear regression network and stress. The linear regression network has small calculation amount, high prediction accuracy and simple realization.

[0032] In order to make the above-mentioned purposes, characteristics and advantages of the present application more obvious and easy to understand, the following preferred embodiments are described in detail below, and the accompanying drawings are described as follows. BRIEF DESCRIPTION OF DRAWINGS

[0033] The drawings accompanying the specification of the present application are used to provide further understanding of the present application, the illustrative embodiments of the present application and the description thereof are used to explain the present application, and do not constitute improper limitation of the present application.

[0034] Figure 1 is a comprehensive evaluation and prediction flowchart of the present application;

[0035] Figure 2 is a capacity degradation curve;

[0036] Figure 3 is a charge-discharge cycle stress growth curve;

[0037] Figure 4 is a stress-capacity curve;

[0038] Figure 5 is a weight determination method;

[0039] Figure 6 is a comparison diagram of actual life and predicted life. DETAILED DESCRIPTION

[0040] The present application will be further described below in combination with the drawings and embodiments.

[0041] It should be pointed out that the following detailed description is exemplary, and is intended to provide further description of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as generally understood by those skilled in the art to which the present application belongs.

[0042] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, components, and / or groups thereof, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0043] As described in the background, the traditional RUL definition method is based on capacity and internal resistance information, which is not easy to obtain in real time and cannot be predicted online. The battery life degradation is affected by multiple factors coupling, and the RUL defined by capacity and internal resistance only macroscopically reflects the life of the battery itself. Stress, on the other hand, originates from the inside of the battery and is closely related to various chemical reactions and property changes inside the battery. Therefore, according to the definition method of capacity and internal resistance, and the strong correlation between stress and battery life, the present application proposes a stress-based dynamic battery RUL definition method, as shown in formula (1). Further combining various RUL definition methods such as battery stress, capacity, internal resistance, etc., the different RULs are weighted and summed, as shown in formula (2), to comprehensively evaluate the remaining life of the battery. Finally, using the regularized linear regression network, the current, voltage, capacity, stress, etc. are taken as input to predict the battery RUL online.

[0044]

[0045] In the formula, are the cycle number (cycle life) or the use time (calendar life) at the end of life, respectively, are the current cycle number (cycle life) or the use time (calendar life), respectively, and use is the current battery stress, ini is the initial stress, and M is a constant, M ∈ [2.9, 3.2].

[0046]

[0047] In the formula, a, b, c are weight coefficients, which can be selected according to different occasions, and RUL c and RUL r are the remaining life of the battery defined by capacity and internal resistance, respectively, and RUL σ is the new definition of battery life proposed.

[0048] As shown in Figure 1 , the specific steps are as follows:

[0049] (1) Perform charge and discharge tests on the battery to obtain the initial stress, initial internal resistance, and rated capacity of the battery;

[0050] (2) Cycle charge-discharge experiment is carried out on the battery, the battery capacity, internal resistance and battery stress of each charge-discharge are obtained, and the data are processed;

[0051] (3) When the available capacity of the battery is lower than 80% of the initial capacity, the internal resistance increases to 1.5 times of the initial internal resistance, and the stress increases to about 3 times of the initial stress, the cycle charge-discharge experiment is stopped, and the battery stress at this time is recorded;

[0052] (4) The stress data are analyzed, the correlation between the battery stress and the battery life is calculated, and the battery remaining useful life RUL based on the stress definition is obtained σ ;

[0053] (5) Based on the capacity and internal resistance data in (2) and (3), the corresponding battery remaining useful life RUL c , RUL r is calculated respectively, and the battery life is comprehensively evaluated in combination with the battery life RUL σ based on the stress obtained in (4);

[0054] (6) The battery stress, current, voltage and other data are taken as input, and the battery life is taken as output, and a regularized regression network is trained to realize online prediction of RUL.

[0055] Battery life degradation is generally considered as capacity attenuation or internal resistance increase, and currently RUL is mainly defined and evaluated by the changes of battery capacity and internal resistance. The capacity definition method is that the charge-discharge times or use time when the available capacity of the battery decreases to 80% of the rated capacity, that is,

[0056]

[0057] In the formula, respectively are the cycle times or use time at the end of life, respectively are the current cycle times or use time, Q use is the current available capacity of the battery, Q nor is the rated capacity.

[0058] The internal resistance definition method is that the charge-discharge times or use time when the internal resistance of the battery reaches 1.5 times of the initial internal resistance, that is,

[0059]

[0060] In the formula, respectively are the cycle times or use time at the end of life, respectively are the current cycle times or use time, R use is the current internal resistance of the battery, R ini is the initial internal resistance.

[0061] Cycle charge-discharge experiments were carried out on a certain soft package lithium ion battery at different temperatures of 25℃, 45℃ and 60℃ respectively until the available capacity of the battery decreased to 80% of the rated capacity, the internal resistance increased to 1.5 times of the initial internal resistance, and the stress growth was 3 times of the initial stress. The obtained data were pretreated, and the capacity fading and stress growth curves were plotted as shown in Figure 2 、 Figure 3 and Figure 4 .

[0062] As shown in Figure 2 , the battery capacity fading and the increase of cycle number were basically in linear relationship. At the same time, with the increase of temperature, the battery life gradually shortened. At 25℃, the life was the longest, about 750 cycles, while at 60℃, the life decreased to 350 cycles. The battery life was greatly affected by temperature.

[0063] As shown in Figure 3 , with the increase of cycle number, the battery stress also gradually increased. Similar to the capacity curve, the stress at the end of life at different temperatures was basically equal, about 2.9-3.2 times of the initial stress, so the stress could be used as a characteristic to define the battery RUL. At the same time, analyzing the stress curves at different temperatures showed that the time of the stress "inflection point" appeared early, about 50-70 cycles, and the linearity of stress and cycle number was stronger after the "inflection point". When the capacity or internal resistance was used to define RUL, the capacity fading and internal resistance increase were not obvious in the early stage of cycle, i.e. the "inflection point" of capacity and internal resistance appeared late. Therefore, it was more simple and convenient to define RUL and evaluate and predict using stress.

[0064] The stress-capacity relationship curve was plotted by combining the stress and capacity data as shown in Figure 4 . It can be seen from Figure 4 that in the early stage of capacity fading, the battery stress growth rate was fast, and after the appearance of the stress "inflection point", i.e. the capacity decreased to about 98%-97% of the rated capacity, the stress growth rate slowed down, but it still showed a strong linear relationship with the capacity fading. This further illustrated the superiority of defining RUL by stress.

[0065] Further, the Pearson correlation coefficient method was used to analyze the correlation between stress and capacity. The calculation formula of Pearson correlation coefficient r is as follows:

[0066]

[0067] In the formula, cov represents the covariance of variables.

[0068] The Pearson coefficient of capacity and stress at different temperatures is calculated, and the results are shown in Table 1. Since the degradation of battery life is manifested as an increase in stress or a decrease in capacity, the correlation coefficient of stress and capacity should be negative. As can be seen from Table 1, the battery capacity and stress have a good linear relationship.

[0069] Table 1 Pearson correlation coefficient of residual capacity and stress

[0070]

[0071] Based on the above analysis, stress and battery life have a strong correlation, and with the increase of cycle number, the time of stress "inflection point" appears earlier, and the change of stress before and after the "inflection point" with the cycle number is linear. When the current stress of the battery reaches 2.9-3.2 times of the initial stress, it is considered that the battery life is terminated. The cycle number or usage time corresponding to the stress of the battery at the end of life minus the cycle number or usage time corresponding to the stress in the current state is the remaining useful life RUL of the battery.

[0072] Based on the various definition methods of RUL in formula (1), formula (3) and formula (4), weights are set for different RUL definitions, and the battery remaining useful life is comprehensively evaluated, as shown in formula (2). The comprehensive evaluation method combines three definitions of stress, capacity and internal resistance, selects a suitable multi-parameter comprehensive evaluation model according to different application occasions, calculates the weight coefficient, and sums the different RULs weighted to realize the comprehensive evaluation of the battery remaining useful life. The evaluation result is objective and accurate, not easy to be disturbed by external environment, and has strong robustness. The commonly used multi-parameter comprehensive evaluation model is shown in formula (5). Figure 5

[0073] After obtaining the evaluation result, the battery life is predicted by using the regularized linear regression. The current, voltage, temperature, stress and other characteristics are taken as input, and the life is taken as output. The sample is represented by formula (6).

[0074]

[0075] In the formula, subscript m is the number of batteries, n is the number of characteristics, t represents the comprehensive evaluation result, and f is different characteristics.

[0076] The linear regression model used is shown in formula (7).

[0077]

[0078] In the formula, y' m is the predicted life of battery m, x m is a p-dimensional characteristic vector, and ω is a p-dimensional coefficient vector. In order to avoid overfitting, a penalty term is added to the equation by applying regularization technique, as follows.

[0079] ​

[0080] where argmin denotes the value of ω that minimizes the error. X is the feature matrix. The first term can be calculated by using the ordinary least square algorithm. The second term P(ω) is divided into two types according to different regularization techniques. One is the lasso-based technique, as shown in equation (9).

[0081] P(ω) = ||ω||1(9)

[0082] The other is the elastic net technique, as shown in equation (10).

[0083]

[0084] where α is a constant between 0 and 1. When the number of selected features is large, the elastic net technique can achieve better prediction results. Therefore, different features can be selected according to different scenarios, and then the linear regression network based on the lasso or elastic net technique is used to achieve accurate RUL prediction. When a single stress feature is used for prediction, the life prediction values at 25°C, 45°C and 60°C are 683, 539 and 315, respectively. The prediction results are evaluated by absolute error (AE) and relative error (RE), and the calculation formulas are as follows.

[0085] AE = |y pre -y act | (11)

[0086]

[0087] where y pre and y ac represent the life prediction value and the actual value, respectively.

[0088] Table 2 Error comparison table

[0089]

[0090] Table 2 gives the prediction error comparison, Figure 6 is the comparison of actual life and predicted life. From Figure 6 and Table 2, the life prediction value and the actual life are different by 26 cycles, 35 cycles and 1 cycle, respectively, and the error is less than 6%, which shows the superiority of the RUL definition and evaluation based on stress.

[0091] Those skilled in the art will appreciate that embodiments of the present application can be readily used as software, hardware, or a combination of software and hardware. In a software embodiment, the methods can be tangibly embodied in a machine-readable storage medium having stored thereon instructions that can be used to program a computer to perform any of the methods. The software implementation can be for example, in a modulated data signal such as a carrier wave or other transport mechanism, or a computer readable storage medium.

[0092] The present application is described in reference to the drawings, which are as follows: Figure 1 Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks.

[0093] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks.

[0094] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 Figure 1 The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks.

[0095] The application described and claimed herein is not to be limited in scope by the embodiments described herein since the scope of these embodiments is deemed to include any modifications within the spirit and scope of the embodiments. Accordingly, what is desired to be secured by Letters Patent is the application as defined by the following claims, including all equivalents.

[0096] ​​​The above describes the specific embodiments of the present application in combination with the drawings, but is not a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications or variations made by those skilled in the art on the basis of the technical solutions of the present application without creative labor are still within the protection scope of the present application.

Claims

1. A stress-based method for defining, evaluating and predicting the remaining life of a battery, characterized in that, The method comprises the following steps: The battery stress, current and voltage data of the target power battery are taken as input, and the pre-trained model is used to obtain the predicted remaining life; The training process of the model comprises: The battery capacity, internal resistance and battery stress of each charge and discharge obtained through the cycle charge and discharge experiment of the target battery are acquired, and the data is processed; The battery remaining life is evaluated by comprehensively considering the battery capacity, internal resistance and battery stress; The model is trained by taking the battery stress, current and voltage data as input and the battery remaining life as output; The battery remaining life defined based on stress is: wherein , are the cycle number or the time to life termination, respectively, , are the current cycle number or the time to life termination, respectively, σuseis the current battery stress, σinis the initial stress, and M is a constant.

2. The method of claim 1, wherein the stress-based battery remaining life definition, assessment and prediction is characterized by, The specific process of evaluating the battery remaining life by comprehensively considering the battery capacity, internal resistance and battery stress comprises: When the available capacity of the battery is lower than the set value of the initial capacity, the internal resistance increases to k times the initial internal resistance, the stress increases to m times the initial stress, the cycle charge and discharge experiment is stopped, and the battery stress at this time is recorded, k and m are positive numbers; The correlation between the battery stress and the battery life is calculated by analyzing the stress data, and the battery remaining life defined based on stress is obtained; Based on the capacity and internal resistance data, the corresponding battery remaining life is calculated respectively, and the battery remaining life is comprehensively evaluated in combination with the battery remaining life defined based on stress.

3. The method for defining, evaluating and predicting the remaining life of a battery based on stress according to claim 1 or 2, characterized in that, The battery remaining life is evaluated by comprehensively considering the battery capacity, internal resistance and battery stress, and the battery remaining life defined based on stress, capacity and internal resistance is comprehensively evaluated. According to the application occasion, the weight coefficients of each definition method are selected, the different battery remaining lives are weighted and summed, and the comprehensive evaluation of the battery remaining life is realized.

4. The method of claim 3, wherein the stress-based battery remaining life definition, assessment and prediction is characterized by, The battery remaining life evaluated by comprehensively considering the battery capacity, internal resistance and battery stress is: Wherein, a, b, c are weight coefficients, which can be selected according to different occasions. RULc and RULr are the battery remaining life defined based on capacity and internal resistance respectively, and RULσ is the new definition of battery life.

5. The method of claim 3, wherein the stress-based battery remaining life definition, assessment and prediction is characterized by, When the weight coefficients of each definition method are selected according to the application occasion, TOPSIS, fuzzy evaluation, grey correlation, rank sum ratio or comprehensive index method is used to determine.

6. The method for defining, evaluating and predicting the remaining life of a battery based on stress according to claim 1, characterized in that, The model is a regularized linear regression network. Or further, a penalty term is added to the model to minimize the error, and a linear regression network based on lasso or elastic net technology is selected.

7. A stress-based battery remaining life definition, evaluation and prediction system, which performs the steps in the method of any one of claims 1-6, comprising: A prediction module configured to take the battery stress, current and voltage data of the target power battery as input, and use the pre-trained model to obtain the predicted remaining life; A training module configured to acquire the battery capacity, internal resistance and battery stress of each charge and discharge obtained through the cycle charge and discharge experiment of the target battery, process the data, evaluate the battery remaining life by comprehensively considering the battery capacity, internal resistance and battery stress, and train the model by taking the battery stress, current and voltage data as input and the battery remaining life as output.

8. A computer readable storage medium characterized by, A terminal device having a plurality of instructions stored therein, wherein the instructions are adapted to be loaded and executed by the processor of the terminal device to perform the steps in the method of any one of claims 1-6.

9. A terminal device, characterized by, The computer readable storage medium stores a plurality of instructions adapted to be loaded by the processor and to perform the steps of the method of any one of claims 1-6. The computer readable storage medium stores a plurality of instructions adapted to be loaded by the processor and to perform the steps of the method of any one of claims 1-6.

Images