Battery life prediction method and device, electronic equipment and storage medium
Through the nonlinear drift Wiener process model and extended Kalman filtering technology, the existing battery life prediction methods are solved, and efficient battery life prediction in different usage scenarios is achieved.
Patent Information
- Application Number
- CN202411000317.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-24
- Publication Date
- 2024-12-03
AI Technical Summary
The existing battery life prediction methods have problems such as low accuracy, high cost, long time period and poor universality, especially in different usage scenarios, which are difficult to accurately predict battery life.
Using the battery life prediction method based on the nonlinear drift Wiener process model, the Wiener degradation process model with nonlinear drift is constructed, combined with extended Kalman filtering, the target battery parameters are estimated and the life prediction is predicted, which is suitable for different uses. Scene.
Improves the accuracy of battery life prediction, reduces costs, shortens the prediction time period, and improves the universality of prediction, which can accurately capture dynamically changing battery capacity status.
Smart Images

Figure CN119066836A_ABST
Abstract
Description
Technical Field
[0001] The embodiments of the present application relate to the field of automobile application technology, and in particular to a battery life prediction method, device, electronic device and storage medium. Background Art
[0002] As time goes by, the industry is constantly improving, and battery manufacturing technology is constantly innovating. The life of new batteries is often difficult to conform to the probability distribution of previous battery life. However, batteries often have long life characteristics, and even if accelerated life tests are performed, it is extremely difficult to obtain enough failure data. If the selected battery is more expensive, once its manufacturing process is better or its resistance to loss in all aspects is strong, the cost will be very expensive, and it will be difficult to complete the degradation experiment in a short time, and finally it will be difficult to obtain complete life data. The use scenarios of a battery are not the same. The external environment and usage conditions have a huge impact on the life of the battery, that is, batteries under different working conditions cannot conform to the same life distribution.
[0003] In related technologies, battery life prediction methods generally use historical life data, obtain the distribution pattern of battery life through statistical analysis, and then infer the probability distribution of battery life based on the pattern. Therefore, this method of battery life prediction has the problems of low accuracy, high cost, long time period and poor universality. Summary of the invention
[0004] The embodiments of the present application provide a battery life prediction method, device, system, electronic device and computer-readable storage medium to solve the problems of low accuracy, high cost, long time period and poor universality in battery life prediction in related technologies.
[0005] In a first aspect, an embodiment of the present application provides a battery life prediction method, the method comprising:
[0006] Constructing a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes the battery degradation;
[0007] Estimate target battery parameters and maximum likelihood estimates corresponding to the target battery parameters according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data;
[0008] updating the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters to obtain a target battery life prediction model;
[0009] The life of the second target battery is predicted according to the target battery life prediction model to obtain a predicted remaining life of the second target battery.
[0010] Optionally, estimating a target battery parameter and a maximum likelihood estimation value corresponding to the target battery parameter according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data includes:
[0011] The acquired battery capacity degradation sample data for different batteries include the first battery capacity degradation sample data corresponding to the first target battery;
[0012] The Wiener degradation process model with nonlinear drift is iterated according to the first battery capacity degradation sample data as a priori samples, and when a tolerance termination condition is met, a target battery parameter and a maximum likelihood estimation value corresponding to the target battery parameter are obtained.
[0013] Optionally, the target battery parameters include target battery a priori parameters and target battery common parameters.
[0014] Optionally, the battery capacity degradation sample data of the different batteries include second battery capacity degradation sample data corresponding to the second target battery;
[0015] The performing life prediction on the second target battery according to the target battery life prediction model to obtain the predicted remaining life corresponding to the second target battery comprises:
[0016] Obtaining a second capacity loss corresponding to a preset number of charge and discharge cycles of the second battery;
[0017] The life of the second target battery is predicted according to the target battery parameter, the maximum likelihood estimate corresponding to the target battery parameter, and the second capacity loss to obtain a predicted remaining life corresponding to the second target battery.
[0018] Optionally, after the step of acquiring battery capacity degradation sample data for different batteries, the method includes:
[0019] Calculating a first capacity loss according to the first battery capacity degradation sample data corresponding to the first target battery, wherein the first capacity loss includes an absolute value of a difference between a discharge capacity at a target time and a discharge capacity at an initial time;
[0020] When the first capacity loss reaches a preset loss value, the first target battery is in a failed state.
[0021] Optionally, after the step of predicting the life of the second target battery according to the target battery life prediction model to obtain a predicted remaining life corresponding to the second target battery, the method includes:
[0022] The target battery parameters corresponding to the target battery life prediction model are updated online through extended Kalman filtering, and the life of the second target battery is predicted online according to the updated target battery life prediction model.
[0023] Optionally, the online updating of the target battery parameters corresponding to the target battery life prediction model by using an extended Kalman filter, and the online life prediction of the target battery according to the updated target battery life prediction model includes:
[0024] The target battery parameters corresponding to the first time corresponding to the battery capacity degradation first reaching the failure threshold corresponding to the second target battery and the second time corresponding to the extended Kalman filter tracking termination point are updated online, and the target battery is predicted to have an online life according to the updated target battery life prediction model.
[0025] In a second aspect, an embodiment of the present application provides a battery life prediction device, the device comprising:
[0026] An initial construction module, used to construct a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes the battery degradation;
[0027] A parameter estimation module, used to estimate target battery parameters and maximum likelihood estimation values corresponding to the target battery parameters according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data;
[0028] A model updating module, used for updating the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters, so as to obtain a target battery life prediction model;
[0029] The prediction module is used to predict the life of the second target battery according to the target battery life prediction model to obtain the predicted remaining life corresponding to the second target battery.
[0030] Optionally, the parameter estimation module includes:
[0031] A first parameter estimation submodule, configured to obtain battery capacity degradation sample data for different batteries, wherein the battery capacity degradation sample data for different batteries includes first battery capacity degradation sample data corresponding to a first target battery;
[0032] The second parameter estimation submodule is used to iterate the Wiener degradation process model with nonlinear drift based on the first battery capacity degradation sample data as a priori samples, and obtain the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters when the tolerance termination condition is met.
[0033] Optionally, the second parameter estimation submodule is specifically used for the target battery parameters including target battery a priori parameters and target battery common parameters.
[0034] Optionally, the battery capacity degradation sample data of different batteries include second battery capacity degradation sample data corresponding to the second target battery; and the prediction module includes:
[0035] A first prediction submodule, used to obtain a second capacity loss corresponding to the second battery during a preset number of charge and discharge cycles;
[0036] The second prediction submodule is used to predict the life of the second target battery according to the target battery parameter, the maximum likelihood estimate corresponding to the target battery parameter and the second capacity loss, so as to obtain the predicted remaining life corresponding to the second target battery.
[0037] Optionally, the extraction module includes:
[0038] A third parameter estimation submodule, configured to calculate a first capacity loss according to the first battery capacity degradation sample data corresponding to the first target battery, wherein the first capacity loss includes an absolute value of a difference between a discharge capacity at a target moment and a discharge capacity at an initial moment;
[0039] The fourth parameter estimation submodule is used to determine that when the first capacity loss reaches a preset loss value, the first target battery is in a failed state.
[0040] Optionally, the device further comprises:
[0041] The Kalman filter prediction module is used to perform online update of the target battery parameters corresponding to the target battery life prediction model through extended Kalman filtering, and perform online life prediction for the second target battery according to the updated target battery life prediction model.
[0042] Optionally, the Kalman filter module includes:
[0043] The Kalman filter prediction submodule is used to update the target battery parameters online according to the first time corresponding to the battery capacity degradation first reaching the failure threshold corresponding to the second target battery and the second time corresponding to the extended Kalman filter tracking termination point, and perform online life prediction on the target battery according to the updated target battery life prediction model.
[0044] In a third aspect, an embodiment of the present application further provides an electronic device, comprising: a processor; and a memory for storing instructions executable by the processor, wherein the processor is configured to execute the instructions to implement any of the battery life prediction methods described above.
[0045] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, which, when instructions in the computer-readable storage medium are executed by a processor of an electronic device, enables the electronic device to execute any of the battery life prediction methods described above.
[0046] In a fifth aspect, an embodiment of the present application further provides a vehicle, comprising the battery life prediction device as described above.
[0047] In an embodiment of the present application, a Wiener degradation process model with nonlinear drift is constructed according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes battery degradation; the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters are estimated according to the Wiener degradation process model with nonlinear drift and the pre-acquired battery capacity degradation sample data; the Wiener degradation process model with nonlinear drift is updated according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters to obtain a target battery life prediction model; the life of the second target battery is predicted according to the target battery life prediction model to obtain the predicted remaining life corresponding to the second target battery. That is, the prediction method based on the Wiener process degradation model with nonlinear drift in the embodiment of the present application is more in line with the nonlinear characteristics appearing in actual use conditions, can improve the prediction accuracy, and has higher reliability than other traditional prediction methods. This application does not require a large amount of prior sample battery data. Only a small number of sample batteries are used to estimate model parameters. The prediction method has low cost and short time period, and the prediction method is not affected by the usage scenario. The random process can accurately capture these dynamic changes and is universal for predicting battery capacity in different usage scenarios.
[0048] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0049] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:
[0050] Figure 1 is a flowchart of the steps of a battery life prediction method provided by an embodiment of the present application;
[0051] Figure 2 is a flowchart of another method for predicting battery life provided by an embodiment of the present application;
[0052] Figure 3 is a flowchart of another method for predicting battery life provided by the present application;
[0053] Figure 4 is a device block diagram of a battery life prediction device provided in an embodiment of the present application;
[0054] Figure 5 is a structural diagram of an electronic device provided in an embodiment of the present application;
[0055] Figure 6 is a schematic diagram of a battery life prediction process provided in an embodiment of the present application;
[0056] Figure 7 is a schematic diagram of a battery capacity degradation curve provided in an embodiment of the present application;
[0057] Figure 8 This is a schematic diagram of the end-of-life cycle number of a battery provided in an embodiment of the present application;
[0058] Fig. 9 This is a model parameter optimization iteration effect diagram provided in the embodiment of the present application
[0059] Fig.10 is a probability density of the remaining life distribution of a partial charge and discharge cycle provided in an embodiment of the present application;
[0060] Fig.11 It is a remaining life corresponding to each charge and discharge cycle provided in an embodiment of the present application;
[0061] Fig.12This is a RUL and actual discharge capacity of each charge and discharge cycle provided in an embodiment of the present application. DETAILED DESCRIPTION
[0062] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.
[0063] Figure 1 is a flowchart of a battery life prediction method provided by an embodiment of the present application, such as Figure 1 As shown, the method may include:
[0064] Step 101, constructing a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes the battery degradation;
[0065] It should be noted that in the embodiment of the present application, the Wiener process with nonlinear drift can be constructed with reference to the following formula 1 to characterize battery degradation.
[0066] Formula 1)
[0068] Among them, the degradation process X(t) is driven by the standard Brownian motion B(t), and X(0) is the initial battery degradation at time 0; represents the average cumulative effect of the battery degradation process, which is called the nonlinear drift of the degradation amount X(t); σ is the diffusion coefficient. Specifically, α represents the random parameter of the difference between batteries, and β and σ are the common parameters representing the degradation of lithium-ion batteries. Without loss of generality, assume that X(0) = 0, and α obeys the mean μ α , with standard deviation σ α The normal distribution of And it is assumed that α and B(t) are statistically independent.
[0069] It should be noted that the above is the average cumulative effect value, and the above α and σ are the common parameters of the initial battery.
[0070] Step 102, estimating target battery parameters and maximum likelihood estimation values corresponding to the target battery parameters according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data;
[0071] It should be noted that, in an embodiment of the present application, the pre-acquired battery capacity degradation sample data can be divided into two parts, that is, the pre-acquired battery capacity degradation sample data for different batteries can be divided into first battery capacity degradation sample data corresponding to the first target battery, and second battery capacity degradation sample data corresponding to the second target battery.
[0072] The first battery capacity degradation sample data corresponding to the first target battery can be used as a priori battery capacity degradation data, the first battery capacity degradation sample data is used for a priori samples, and the second battery capacity degradation sample data is used outside the a priori samples.
[0073] The battery capacity degradation sample data may be based on data collected from existing lithium-ion battery ground tests at NASA Ames Center.
[0074] For example, the capacity degradation data of batteries 1 to 4 in the sample can be extracted. In this application, the failure threshold of the lithium-ion battery in the sample is set to capacity degradation w = 0.4Ah. The capacity at the initial moment is the full capacity, and the capacity loss is X i , represents the absolute value of the difference between the discharge capacity of the battery at the i-th cycle and the discharge capacity at the initial moment.
[0075] Reference Figure 7 , Figure 7 The middle curves b1, b2, b3, and b4 represent the capacity loss of batteries 1 to 4 respectively. When the capacity loss reaches 0.4Ah, it means that the battery has failed, that is, its life span has ended.
[0076] Reference Figure 8 , Figure 8 is the end time of each battery life, battery No. 1 fails at the 107th charge and discharge; battery No. 2 fails at the 60th charge and discharge cycle; battery No. 3 fails at the 129th charge and discharge cycle; battery No. 4 fails at the 81st charge and discharge cycle.
[0077] After constructing the Wiener degradation process model with nonlinear drift and obtaining the battery capacity degradation sample data, the target battery parameters can be estimated based on the historical degradation data of similar batteries, where the target battery parameters include the common parameters β and σ, and the parameter μ in the prior distribution of the random parameter α a and
[0078] Specifically, there are degradation data from M batteries of the same type, and the number of degradation data of the i-th battery is N, (1≤i≤M). Let X(t i,j )(abbreviated as X i,j ) indicates that the ith battery is i,jThe (i.e. jth) degradation data at the moment, and 1≤j≤N, based on formula (1), we have
[0079]
[0080] Where: α0 represents the prior value of the random parameter α, and its distribution is Furthermore, it is assumed that the degradation data of different batteries are uncorrelated, while the degradation data of the same battery are correlated, and the correlation is characterized by the covariance matrix. According to formula (1) and the properties of Brownian motion, the following theorem can be obtained.
[0081] Theorem 1 If the degradation data of the i-th battery is X i =(X i,1 X i,1 X i,3 …X i,N )′, then X i It obeys a multivariate Gaussian distribution, and its mean and covariance are:
[0082] μ i =μ α I i (Formula 3)
[0083]
[0084] In formula 3 and 4,
[0085]
[0086] Based on Theorem 1, the log-likelihood function generated by the degradation data from all M batteries is:
[0087]
[0088] In this formula, Θ=(μ α σ a βσ)′ is an unknown parameter; X 1:M =(X1X2…X M ) represents the degradation data of all M batteries.
[0089] The capacity degradation data of the 1st, 2nd and 3rd battery are used as prior samples to iterate the model and estimate the parameters. The maximum likelihood value is obtained by the fminsearch function in MATLAB, and the options structure is created by optimset. The termination tolerance of the function is set to 1*e-10 in the options structure, that is, the control optimizer ends the iteration when the difference between the current function value and the previous value is less than 1*e-10. To ensure that the parameter optimization will not be terminated early, the maximum number of iterations is set to 1000.
[0090] Reference Fig. 9 , Fig. 9 This is the iterative process of optimizing parameters in the simulation process. When the model iterates to the 158th time and meets the tolerance termination condition, the iteration stops and the maximum likelihood estimation value of each parameter is output. Table 1 summarizes the estimated values of each parameter.
[0091] Table 1 Maximum likelihood estimates of an exemplary model parameter
[0092] <![CDATA[μ α ]]> <![CDATA[σ a 2 ]]> β <![CDATA[σ 2 ]]> 5.204373e-03 8.275787e-26 9.308427e-01 2.818561e-04
[0093] Therefore, the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters can be estimated based on the Wiener degradation process model with nonlinear drift and the pre-acquired battery capacity degradation sample data.
[0094] Step 103, updating the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters to obtain a target battery life prediction model;
[0095] It should be noted that in an embodiment of the present application, the target battery parameters and their maximum likelihood estimates iteratively estimated according to step 102 can be used to replace the Wiener degradation process model with nonlinear drift constructed in the initial step 101, that is, the simulated initial values are replaced with the closest target battery parameters estimated based on the prior sample data of the present application, thereby obtaining a target battery life prediction model.
[0096] Furthermore, after step 103, in order to better reflect the battery evaluation process, a quantity, namely, the first arrival time, is introduced. The main principle of the first arrival time is that the battery is constantly degrading and its remaining life is constantly decreasing. When the total degradation reaches the failure threshold X for the first time f When the battery is considered to be faulty, the remaining life is 0.
[0097] Therefore, the end of battery life can be defined as the time when the random degradation process {X(t), t≥0} first crosses the failure threshold X f The lifetime T is defined as:
[0098] T=inf{t:X(t)≥X f ∣X(0) <X f} (Formula 7)
[0099] In the above formula 7, inf (lower limit function) represents the minimum time in the set, and the conditional probability density function of the life span T under the definition is:
[0100]
[0101] Among them, the unconditional probability density function f(t) of life span T can be obtained by the total probability formula based on formula 9:
[0102] f T (t) = ∫ Ω f T∣α (t||α)f(α)dα (Formula 9)
[0103] In Formula 9, f(α) and Ω are the probability density function and parameter space of the random parameter α, respectively.
[0104] Furthermore, the probability density function of the remaining life of the equipment can be obtained using the result of formula 9. The process is as follows:
[0105] Let t k (k∈N + ) represents the current time, l k Indicates t k The remaining life of the battery at time t is a random variable. If the battery reaches the failure threshold at time t, then the remaining life is l k =tt k Using the independent increment characteristics of Brownian motion, we can get from formula 1:
[0106]
[0107] Among them, Y(l k )=X(t k +l k )-X(t k ), and Y(0)=0.
[0108] It can be seen that formula 10 has a similar functional form to formula 1. Therefore, based on the results of formula 8 and formula 9, the remaining life l of the degradation model of formula 1 is obtained k The probability density function of is as follows: Formula 11.
[0109]
[0110] In the above formula 11, μ α and σ 2 are the mean and variance of α respectively; x k For the battery at t k The degenerate observation at time .
[0111] Step 104 , predicting the life of the second target battery according to the target battery life prediction model to obtain a predicted remaining life of the second target battery.
[0112] It should be noted that in the embodiment of the present application, a target battery outside the prior sample is used as a prediction object for life prediction. Therefore, the life prediction of the second target battery is performed according to the target battery life prediction model to obtain the predicted remaining life corresponding to the second target battery.
[0113] Specifically, based on the maximum likelihood estimate MLE of the model parameters calculated above, the No. 4 battery is used as the experimental object, that is, the second target battery, to predict the life span. Substitute the parameter values into the formula (Formula 11), where x k Represents the No. 4 battery at t k The capacity loss during the charge and discharge cycle. The probability density of the remaining life distribution of the No. 4 battery in each charge and discharge cycle can be obtained by calculation. In order to observe the change trend, the 10th, 20th, 30th, 40th, 50th, and 60th cycles are taken to draw a three-dimensional graph for reference. Fig.10 shown.
[0114] Depend on Fig.10 It can be seen that as the number of charge and discharge cycles increases, the value corresponding to the highest point of the PDF image of the battery RUL gradually decreases, which means that the battery life is getting shorter and shorter, which is in line with the actual situation. The expectation of the probability density function of the RUL distribution is calculated using the formula
[0115] E(x)=∫x*f(x)dx (Formula 12)
[0116] The average remaining life corresponding to each charge and discharge cycle can be calculated, and the remaining life corresponding to each charge and discharge cycle can be drawn. Fig.11 shown.
[0117] According to the definition of the first arrival time, it can be known that the proposed nonlinear drift Wiener degradation process model predicts that the remaining number of cycles of the No. 4 battery life is 0 when it reaches the 81st charge and discharge cycle, and the capacity degradation reaches the failure threshold for the first time, and the battery life ends. Adding the actual capacity decay process of the battery, such as Fig.12 shown.
[0118] Depend on Fig.12 It can be seen from the two curves that the predicted RUL change trend follows the actual capacity of the battery itself, which can well reflect the actual health status of the battery.
[0119] To facilitate statistical analysis, the actual remaining life and predicted remaining life are plotted every five charge and discharge cycles as shown in Table 2. Represents the predicted RUL, in cycles.
[0120] Table 2 An exemplary RUL prediction result based on the Wiener process model
[0121]
[0122]
[0123] Furthermore, after step 104, the battery life prediction method also includes: online updating the target battery parameters corresponding to the target battery life prediction model through extended Kalman filtering, and performing online life prediction on the second target battery according to the updated target battery life prediction model.
[0124] Specifically, the online updating of the target battery parameters corresponding to the target battery life prediction model by using the extended Kalman filter, and the online life prediction of the target battery according to the updated target battery life prediction model includes:
[0125] The target battery parameters corresponding to the first time corresponding to the battery capacity degradation first reaching the failure threshold corresponding to the second target battery and the second time corresponding to the extended Kalman filter tracking termination point are updated online, and the target battery is predicted to have an online life according to the updated target battery life prediction model.
[0126] It should be noted that in the embodiments of the present application, nonlinear processes are common in actual engineering, and the Kalman filter's requirement that the state equation and the observation equation must be linear is too strict. Therefore, the Kalman filter can be expanded to handle nonlinear situations.
[0127] Specifically, in the embodiment of the present application, it is applied to lithium batteries, wherein the RUL prediction of lithium-ion batteries is a prediction of a nonlinear process. For nonlinear problems, one solution is the EKF (Extended Kalman Filter) algorithm, which is an extended form of the standard Kalman filter in a nonlinear situation. The nonlinear function is Taylor expanded (high-order terms are omitted, and the first-order terms of the expanded terms are retained) to achieve linearization of the nonlinearity. Finally, the state estimate and variance estimate of the system are approximated by the Kalman filter algorithm to filter the signal. In the field of mathematics, the Taylor formula can approximate complex nonlinear functions into polynomial sums by using the derivation method. The extended Kalman filter is based on this, and the nonlinear function is expanded to obtain the first-order approximation of the filter value. The first-order Taylor formula (Formula 13) happens to be linear, and the Kalman filter can continue to be applied to complete the prediction work. In other words, EKF is a linear treatment of nonlinear problems.
[0128] f(x)=
[0129] f(a)+f′(a)(xa)+f″(a)((xa)^2) / 2!+f″′(a)((xa)^3) / 3! +…+f^(n)(xa)^n / n! +……
[0130] (Formula 13)
[0131] e^x=1+x+(x^2) / 2! +(x^3) / 3! +(x^4) / 4! +...
[0132] (Formula 14)
[0133] The following is the state-space equation for a nonlinear system:
[0134] X(k+1)=f[k,X(k)]+G(k)W(k) (Formula 15)
[0135] Z(k)=h[k,X(k)]+V(k) (Formula 16)
[0136] Among them, f(**) and h(**) are the state transfer matrix and observation matrix respectively, both of which are nonlinear; G(k) is noise, W(k) and V(k) are process noise and observation noise respectively. X(**) is the system state variable, and Z(**) is the observation variable. In order to realize the prediction of nonlinear process, the system is processed as follows:
[0137]
[0138]
[0139] Among them, X(j) is the state vector, β(j), σ(j), μ a (j) is the state parameter of the degraded observation model, dimensionless. β (j)ω σ (j) is the process noise, V(j) is the observation noise, X(j) is the observed capacity value, and j is the number of cycles.
[0140] The battery failure threshold set in this prediction experiment is 80% of the rated capacity of the battery, that is, when the total capacity degradation reaches 0.4Ah, the battery life ends. The RUL calculation method can be expressed by the formula:
[0141] RUL=k eol -k eom (Formula 19)
[0142] Among them, k eol is the time when the battery capacity degradation first reaches the failure threshold, k eomIt is the time of the filter tracking termination point, that is, the prediction starting point, and the units of both are the number of cycles.
[0143] The embodiment of the present application uses a prediction method based on a nonlinear drift Wiener process degradation model that is more in line with the nonlinear characteristics that appear in actual usage conditions, can improve prediction accuracy, and has higher reliability than other traditional prediction methods. In this application, a large amount of prior sample battery data is not required, and only a small number of sample batteries are used for model parameter estimation. The prediction method has low cost and short time period, and the prediction method is not affected by the usage scenario. The random process can accurately capture these dynamic changes and has universal applicability for predicting battery capacity in different usage scenarios.
[0144] In addition, in the embodiment of the present application, the online life of the lithium battery is predicted by updating the model parameters online through the extended Kalman filter. Based on the different Kemo starting points, it is further verified that the nonlinear Wiener process model prediction has a high reliability.
[0145] Figure 2 is a flowchart of another method for predicting battery life provided by an embodiment of the present application. Figure 2 As shown, the method may include:
[0146] Step 201, constructing a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes the battery degradation;
[0147] It should be noted that, in the embodiment of the present application, the above step 201 refers to the previous discussion and will not be repeated here.
[0148] Step 202, acquiring battery capacity degradation sample data for different batteries, wherein the battery capacity degradation sample data for different batteries includes first battery capacity degradation sample data corresponding to a first target battery;
[0149] Furthermore, after step 202, the method further includes:
[0150] Calculating a first capacity loss according to the first battery capacity degradation sample data corresponding to the first target battery, wherein the first capacity loss includes an absolute value of a difference between a discharge capacity at a target time and a discharge capacity at an initial time;
[0151] When the first capacity loss reaches a preset loss value, the first target battery is in a failed state.
[0152] It should be noted that in the present embodiment, the failure threshold of the lithium-ion battery in the sample is set to capacity degradation w = 0.4Ah. The capacity at the initial moment is the full capacity, and the capacity loss is X i , represents the absolute value of the difference between the discharge capacity of the battery at the i-th cycle and the discharge capacity at the initial moment.
[0153] Step 203, iterating the Wiener degradation process model with nonlinear drift according to the first battery capacity degradation sample data as a priori samples, and obtaining target battery parameters and maximum likelihood estimation values corresponding to the target battery parameters when a tolerance termination condition is met;
[0154] Furthermore, the target battery parameters include target battery a priori parameters and target battery common parameters.
[0155] It should be noted that in the embodiment of the present application, in the above steps 201-203, the capacity degradation data of the 1st, 2nd and 3rd batteries are used as prior samples to perform model iteration to estimate the target battery parameters, wherein the target battery parameters include target battery prior parameters and target battery common parameters, i.e., refer to the four types in Table 1 below.
[0156] Specifically, the maximum likelihood value can be obtained through the fminsearch function in MATLAB, and the options structure can be created using optimset. The termination tolerance of the function can be set to 1*e-10 in the options structure. That is, the optimizer is controlled to terminate the iteration when the difference between the current function value and the previous value is less than 1*e-10. To ensure that the parameter optimization will not be terminated prematurely, the maximum number of iterations is set to 1000.
[0157] Reference Fig. 9 , Fig. 9 This is the iterative process of optimizing parameters in the simulation process. When the model iterates to the 158th time and meets the tolerance termination condition, the iteration stops and the maximum likelihood estimation value of each parameter is output. Table 1 summarizes the estimated values of each parameter.
[0158] Therefore, the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters can be estimated based on the Wiener degradation process model with nonlinear drift and the pre-acquired battery capacity degradation sample data.
[0159] Step 204, updating the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters to obtain a target battery life prediction model;
[0160] Step 205 , predicting the life of the second target battery according to the target battery life prediction model to obtain a predicted remaining life of the second target battery.
[0161] It should be noted that, in the embodiment of the present application, the above steps 204-205 refer to the previous discussion and will not be repeated here.
[0162] The embodiment of the present application uses a prediction method based on a nonlinear drift Wiener process degradation model that is more in line with the nonlinear characteristics that appear in actual usage conditions, can improve prediction accuracy, and has higher reliability than other traditional prediction methods. In this application, a large amount of prior sample battery data is not required, and only a small number of sample batteries are used for model parameter estimation. The prediction method has low cost and short time period, and the prediction method is not affected by the usage scenario. The random process can accurately capture these dynamic changes and has universal applicability for predicting battery capacity in different usage scenarios.
[0163] In addition, in another embodiment of the present application, in order to better evaluate the performance of battery RUL prediction based on EKF, the following evaluation indicators may be introduced to comprehensively evaluate the pros and cons of the prediction results.
[0164] Specifically, the first type, Absolute Error (AE): reflects the difference between the experimental prediction value and the actual life value.
[0165] AE=|RUL pre -RUL act | (Formula 20)
[0166] Among them, RUL pre is the predicted value, RUL act This is the value actually used.
[0167] The second type, relative error (RE), is used to measure the degree of deviation between the measurement result and the true value. The smaller the value, the more accurate and reliable the result.
[0168]
[0169] The calculation of the corresponding evaluation indicators can be summarized in Table 3, where RUL act Indicates the actual RUL, RUL pre Represents the predicted RUL, both units are times (cycles).
[0170] Table 3 An exemplary prediction result and related evaluation indexes obtained by using EKF
[0171]
[0172]
[0173] From Table 3, combined with different keom From the capacity degradation prediction curve, it can be seen that the accuracy of the prediction result will be improved as the prediction starting point moves backward when using the EKF algorithm to predict RUL.
[0174] In the embodiment of the present application, the available historical data of the EKF can increase as the prediction starting point moves backward, so that the parameters of the obtained model are more accurate and reliable, and can better reflect the nonlinear process of capacity degradation of the lithium-ion battery.
[0175] In addition, as the prediction starting point moves backward, the range required for prediction becomes smaller and the data evolution cycle becomes shorter, which also improves the prediction accuracy. Combined with the actual data in Table 3, the calculated statistics AE and RE data, RE are all below 10%, indicating that the EKF prediction method is more accurate.
[0176] Figure 3 is a flowchart of another method for predicting battery life provided by an embodiment of the present application. Figure 3 As shown, the method may include:
[0177] Step 301, constructing a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes the battery degradation;
[0178] Step 302, estimating target battery parameters and maximum likelihood estimation values corresponding to the target battery parameters according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data;
[0179] Step 303, updating the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters to obtain a target battery life prediction model;
[0180] Step 304, obtaining a second capacity loss corresponding to a preset number of charge and discharge cycles of the second battery;
[0181] Step 305 , predicting the life of the second target battery according to the target battery parameters, the maximum likelihood estimation values corresponding to the target battery parameters, and the second capacity loss, to obtain a predicted remaining life corresponding to the second target battery.
[0182] The prediction method based on the nonlinear drift Wiener process degradation model in the embodiment of the present application is more in line with the nonlinear characteristics that appear in actual usage, which can improve the prediction accuracy and has higher reliability than other traditional prediction methods. The prediction method is not affected by the usage scenario, and the random process can accurately capture these dynamic changes, which is universal for predicting the battery capacity in different usage scenarios.
[0183] In addition, the present application uses target batteries outside the prior sample as prediction objects for life prediction, without the need for a large amount of prior sample battery data, and only uses a small number of sample batteries for model parameter estimation. The prediction method has low cost and a short time period.
[0184] Corresponding to the method provided in the above-mentioned battery life prediction method embodiment of the present application, see Figure 4 The present application also provides a logic block diagram of a battery life prediction device. In this embodiment, the device 400 includes:
[0185] An initial construction module 401 is used to construct a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift represents the battery degradation;
[0186] A parameter estimation module 402, configured to estimate target battery parameters and maximum likelihood estimation values corresponding to the target battery parameters according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data;
[0187] A model updating module 403, configured to update the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters, to obtain a target battery life prediction model;
[0188] The prediction module 404 is used to predict the life of the second target battery according to the target battery life prediction model to obtain the predicted remaining life of the second target battery.
[0189] Optionally, the parameter estimation module includes:
[0190] A first parameter estimation submodule, configured to obtain battery capacity degradation sample data for different batteries, wherein the battery capacity degradation sample data for different batteries includes first battery capacity degradation sample data corresponding to a first target battery;
[0191] The second parameter estimation submodule is used to iterate the Wiener degradation process model with nonlinear drift based on the first battery capacity degradation sample data as a priori samples, and obtain the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters when the tolerance termination condition is met.
[0192] Optionally, the second parameter estimation submodule is specifically used for the target battery parameters including target battery a priori parameters and target battery common parameters.
[0193] Optionally, the battery capacity degradation sample data of different batteries include second battery capacity degradation sample data corresponding to the second target battery; and the prediction module includes:
[0194] A first prediction submodule, used to obtain a second capacity loss corresponding to the second battery during a preset number of charge and discharge cycles;
[0195] The second prediction submodule is used to predict the life of the second target battery according to the target battery parameter, the maximum likelihood estimate corresponding to the target battery parameter and the second capacity loss, so as to obtain the predicted remaining life corresponding to the second target battery.
[0196] Optionally, the extraction module includes:
[0197] A third parameter estimation submodule, configured to calculate a first capacity loss according to the first battery capacity degradation sample data corresponding to the first target battery, wherein the first capacity loss includes an absolute value of a difference between a discharge capacity at a target moment and a discharge capacity at an initial moment;
[0198] The fourth parameter estimation submodule is used to determine that when the first capacity loss reaches a preset loss value, the first target battery is in a failed state.
[0199] Optionally, the device further comprises:
[0200] The Kalman filter prediction module is used to perform online update of the target battery parameters corresponding to the target battery life prediction model through extended Kalman filtering, and perform online life prediction for the second target battery according to the updated target battery life prediction model.
[0201] Optionally, the Kalman filter module includes:
[0202] The Kalman filter prediction submodule is used to update the target battery parameters online according to the first time corresponding to the battery capacity degradation first reaching the failure threshold corresponding to the second target battery and the second time corresponding to the extended Kalman filter tracking termination point, and perform online life prediction on the target battery according to the updated target battery life prediction model.
[0203] In summary, the embodiment of the present application provides a battery life prediction device. The embodiment of the present application introduces a more accurate color model, intelligent image segmentation technology and machine learning algorithm by cutting based on a preset black edge removal algorithm, so as to more accurately identify and distinguish the main colors of different areas in the video; for dynamically changing video content, an adaptive adjustment analysis strategy can be implemented to ensure that the main color information can still be effectively extracted under different scene switching or lighting conditions. By processing the video stream, it is not affected by factors such as video format, encoding standard or ambient lighting, and can stably and reliably extract the main color of the video partition in a variety of application scenarios.
[0204] Figure 5 This is a structural diagram of an electronic device M00 provided in an embodiment of the present application. In the diagram, the electronic device M00 includes a processor M01 and a memory M02. The memory M02 stores programs or instructions that can be run on the processor M01. When the program or instructions are executed by the processor M01, the various steps of the above-mentioned battery life prediction method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, they will not be repeated here.
[0205] In an embodiment of the present application, the memory M02 can be used to store software programs and various data. The memory M02 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instructions required for at least one function (such as a sound playback function, an image playback function, etc.), etc. In addition, the memory M02 may include a volatile memory or a non-volatile memory, or the memory x09 may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate synchronous dynamic random access memory (DDRSDRAM), an enhanced synchronous dynamic random access memory (ESDRAM), a synchronous link dynamic random access memory (SLDRAM) and a direct memory bus random access memory (DRRAM). The memory M02 in the embodiment of the present application includes but is not limited to these and any other suitable types of memories.
[0206] The processor M01 may include one or more processing units; optionally, the processor M01 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communication signals, such as a baseband processor. It is understandable that the modem processor may not be integrated into the processor M01.
[0207] An embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the various processes of the above-mentioned battery life prediction method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.
[0208] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk.
[0209] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned battery life prediction method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
[0210] It should be understood that the chip involved in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.
[0211] An embodiment of the present application provides a computer program product, which is stored in a storage medium. The program product is executed by at least one processor to implement the various processes of the battery life prediction method embodiment as described above, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.
[0212] It should be noted that, in this article, the terms "comprise", "include" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise one..." do not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.
[0213] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the relevant technology, can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present application.
[0214] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present application, ordinary technicians in this field can also make many forms without departing from the purpose of the present application and the scope of protection of the claims, all of which are within the protection of the present application.
Claims
1. A battery life prediction method, characterized in that: The method comprises: Constructing a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes the battery degradation; Estimate target battery parameters and maximum likelihood estimation values corresponding to the target battery parameters according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data; updating the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters to obtain a target battery life prediction model; The life of the second target battery is predicted according to the target battery life prediction model to obtain a predicted remaining life of the second target battery.
2. The method according to claim 1, characterized in that The method of estimating the target battery parameter and the maximum likelihood estimation value corresponding to the target battery parameter according to the Wiener degradation process model with nonlinear drift and the pre-acquired battery capacity degradation sample data includes: The acquired battery capacity degradation sample data for different batteries include the first battery capacity degradation sample data corresponding to the first target battery; The Wiener degradation process model with nonlinear drift is iterated according to the first battery capacity degradation sample data as a priori samples, and when a tolerance termination condition is met, a target battery parameter and a maximum likelihood estimation value corresponding to the target battery parameter are obtained.
3. The method according to claim 2, characterized in that The target battery parameters include target battery a priori parameters and target battery common parameters.
4. The method according to claim 2, characterized in that: The battery capacity degradation sample data of different batteries include second battery capacity degradation sample data corresponding to the second target battery; The performing life prediction on the second target battery according to the target battery life prediction model to obtain the predicted remaining life corresponding to the second target battery comprises: Obtaining a second capacity loss corresponding to a preset number of charge and discharge cycles of the second battery; The life of the second target battery is predicted according to the target battery parameter, the maximum likelihood estimate corresponding to the target battery parameter, and the second capacity loss to obtain a predicted remaining life corresponding to the second target battery.
5. The method according to claim 2, characterized in that: After the step of acquiring battery capacity degradation sample data for different batteries, the method includes: Calculating a first capacity loss according to the first battery capacity degradation sample data corresponding to the first target battery, wherein the first capacity loss includes an absolute value of a difference between a discharge capacity at a target time and a discharge capacity at an initial time; When the first capacity loss reaches a preset loss value, the first target battery is in a failed state.
6. The method according to claim 1, characterized in that After the step of predicting the life of the second target battery according to the target battery life prediction model to obtain the predicted remaining life corresponding to the second target battery, the method includes: The target battery parameters corresponding to the target battery life prediction model are updated online through extended Kalman filtering, and the life of the second target battery is predicted online according to the updated target battery life prediction model.
7. The method according to claim 6, characterized in that The online updating of the target battery parameters corresponding to the target battery life prediction model by using the extended Kalman filter, and the online life prediction of the target battery according to the updated target battery life prediction model comprises: The target battery parameters corresponding to the first time corresponding to the battery capacity degradation first reaching the failure threshold corresponding to the second target battery and the second time corresponding to the extended Kalman filter tracking termination point are updated online, and the target battery is predicted to have an online life according to the updated target battery life prediction model.
8. A battery life prediction device, characterized in that: The device comprises: An initial construction module, used to construct a Wiener degradation process model with nonlinear drift according to the pre-acquired initial battery degradation amount, the average cumulative effect value in the battery degradation process, the initial battery common parameters and the standard Brownian motion, wherein the Wiener degradation process model with nonlinear drift characterizes the battery degradation; A parameter estimation module, used to estimate target battery parameters and maximum likelihood estimation values corresponding to the target battery parameters according to the Wiener degradation process model with nonlinear drift and pre-acquired battery capacity degradation sample data; A model updating module, used for updating the Wiener degradation process model with nonlinear drift according to the target battery parameters and the maximum likelihood estimation values corresponding to the target battery parameters, so as to obtain a target battery life prediction model; The prediction module is used to predict the life of the second target battery according to the target battery life prediction model to obtain the predicted remaining life corresponding to the second target battery.
9. An electronic device, characterized in that: include: processor; a memory for storing instructions executable by the processor; The processor is configured to execute the instructions to implement the battery life prediction method as described in any one of claims 1 to 7.
10. A storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the battery life prediction method as described in any one of claims 1 to 7.
Citation Information
Cited By
Method, device and equipment for predicting remaining service life of battery and readable storage medium
CN119535235A
Battery remaining service life prediction method, device, equipment and readable storage medium
CN119535235B
