Lithium battery degradation threshold-impact model construction method and device, equipment and storage medium

Abstract

The application provides a lithium battery degradation threshold-impact model construction method and device, equipment and storage medium. It relates to the field of lithium battery life prediction. The method comprises: establishing a lithium battery inherent degradation model using a nonlinear Wiener process to describe the nonlinear characteristics of capacity attenuation; constructing an external impact damage model based on a non-homogeneous Possion process to represent the time-varying characteristics of random impact; coupling the degradation process and the impact damage through a double-threshold failure mechanism to define the capacity attenuation soft failure threshold and the impact cumulative hard failure threshold; dynamically linking the degradation parameters and the impact parameters to correct them using the Bayesian update and optimization algorithm; and realizing the probability distribution prediction of the remaining life based on Monte Carlo simulation. The application can improve the accuracy and engineering applicability of lithium battery remaining life prediction under complex working conditions.

Description

Technical Field

[0001] This application relates to the field of lithium battery life prediction technology, and in particular to a method, apparatus, device and storage medium for constructing a lithium battery degradation threshold-impact model. Background Technology

[0002] Fault prediction and health management (PHM) technologies integrate equipment condition monitoring, fault diagnosis, life prediction, and maintenance decision-making, serving as a core support for predictive equipment maintenance. Among these, remaining useful life prediction, a core task of PHM, acts as a link between real-time equipment condition monitoring and precise maintenance decisions, and is crucial for ensuring the safe and economical operation of high-value equipment.

[0003] When lithium batteries operate under complex conditions, their performance degradation is affected by both inherent aging effects and external random shocks (such as overcurrent and mechanical vibration), leading to significant uncertainty in RUL prediction results. Although degradation models based on nonlinear Wiener processes can characterize capacity decay and derive the RUL probability distribution, they still have the following limitations:

[0004] (1) Traditional lithium battery models only focus on the inherent degradation process and ignore the coupling effect of external impact damage to the battery, resulting in prediction bias in scenarios with frequent impacts (such as rapid acceleration of electric vehicles and overload of energy storage systems);

[0005] (2) The existing nonlinear Wiener model is not adaptable to the accelerated degradation stage, especially under high-rate charge and discharge conditions, the power-law drift term is difficult to accurately fit the capacity change phenomenon.

[0006] (3) The homogeneous Poisson process is commonly used in the modeling of impact events, which cannot characterize the non-homogeneous characteristics of the actual impact frequency increasing over time (such as the battery being more susceptible to impact damage after aging), resulting in the cumulative damage amount S(t) being underestimated.

[0007] Current RUL prediction methods that integrate degradation and shock typically treat the threshold failure mechanism and parameter update process in isolation: a single failure threshold cannot distinguish between inherent degradation failure and shock-induced failure; the parameter update mechanism lacks linkage, such as Bayesian updates only correcting degradation parameters without simultaneously optimizing the hyperparameters of the shock intensity function. This results in a low degree of matching between the constructed degradation-shock model and the actual battery failure mode, severely limiting the reliability of RUL prediction accuracy in engineering applications.

[0008] Therefore, there is an urgent need to develop a joint modeling method that couples nonlinear degradation modeling, nonhomogeneous impact damage analysis, and dual-threshold failure mechanism. By establishing a degradation-impact dynamic feedback mechanism, high-precision quantitative prediction of lithium battery RUL can be achieved. Summary of the Invention

[0009] This application provides a method, apparatus, device, and storage medium for constructing a lithium battery degradation threshold-impact model. The aim is to propose a joint modeling scheme for lithium battery degradation-threshold-impact (DTS) based on nonlinear Wiener processes and Poisson stochastic processes. By integrating inherent degradation modeling, external impact damage modeling, and dual-threshold failure mechanism, it can achieve high-precision prediction of the remaining life of lithium batteries under complex operating conditions.

[0010] In a first aspect, this application provides a method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes, the method comprising:

[0011] Nonlinear degradation modeling includes: establishing an inherent degradation model of lithium batteries based on a nonlinear Wiener process, wherein the nonlinear Wiener process uses drift parameters in the form of power functions to characterize the nonlinear characteristics of lithium battery degradation;

[0012] Impact damage modeling includes: constructing a lithium battery impact damage model based on a non-homogeneous Poisson process, wherein the impact intensity of the non-homogeneous Poisson process is a function of time to characterize the non-homogeneous characteristics of the impact frequency changing with time.

[0013] Impact strength parameter identification includes: using nonlinear least squares fitting method to identify non-homogeneous parameters.

[0014] The intensity function parameters of the Poisson process are optimized and estimated.

[0015] The dual-threshold failure mechanism is defined as follows: setting a soft failure threshold w and a hard failure threshold D, where the soft failure threshold w is the amount of degradation corresponding to the capacity decay to 80% of the initial capacity, and the hard failure threshold D is the critical value of cumulative impact damage. When the total degradation of the lithium battery is ≥ w or the cumulative impact damage is ≥ D, the lithium battery is determined to be in failure.

[0016] The lithium battery degradation-threshold model is constructed by coupling the inherent degradation amount of the nonlinear Wiener process with the impact damage amount of the nonhomogeneous Poisson process to obtain the expression for the total degradation amount and construct the lithium battery degradation-threshold model.

[0017] Dynamic parameter updates include: based on Bayesian theory, using historical degradation data and real-time observation data of lithium batteries, updating the drift parameters of the degradation model. The update process calculates the posterior distribution by using the likelihood function of the degradation data and the prior distribution of the parameters, thereby achieving dynamic correction of the parameters.

[0018] The lithium battery impact-threshold model is constructed, including: based on the non-homogeneous composite Poisson process, combined with the normal distribution characteristics of the number of impacts and single impact damage, a model is constructed to correlate impact damage with the threshold.

[0019] Reliability analysis includes: using Monte Carlo simulation to perform reliability analysis on the impact-threshold model, simulating the number of impact events and impact intensity, statistically obtaining the probability density function, cumulative distribution function and reliability curve of total impact damage, and predicting the remaining life and reliability of lithium batteries.

[0020] In one possible design, an inherent degradation model for lithium batteries is established based on the nonlinear Wiener process, including:

[0021] A nonlinear Wiener degradation model is established using drift parameters in power function form as an intrinsic degradation model; wherein, the intrinsic degradation model is expressed as:

[0022]

[0023] In the formula, X(t) is the amount of lithium battery degradation at time t, λ(t; θ) is the drift coefficient of the degradation process, x0 is the amount of lithium battery degradation at the initial time, θ is the parameter vector of the nonlinear function, which describes the nonlinear characteristics of lithium battery degradation, and B(t) is Brownian motion, which reflects dynamic uncertainty.

[0024] Express λ(t; θ) as a power function abt t-1 The form is obtained as follows:

[0025] X(t) = x0 + at b +δ B B(t)

[0026] In the formula, a is the scaling factor of the nonlinear degradation trend, which determines the overall amplitude and rate of the degradation curve; b is the exponential term of time t, which controls the nonlinear characteristics of the degradation rate; δ B Let be the diffusion coefficient of Brownian motion (B(t)), which quantifies the intensity of random fluctuations;

[0027] The probability density function f of the first arrival time when the inherent degradation model crosses the failure threshold w is... T|θ (t|θ) is represented as:

[0028]

[0029] In the formula, T is a continuous random variable, representing the remaining lifespan of the lithium battery, and exp is an exponential function with the natural constant as its base.

[0030] In one possible design, a lithium-ion battery impact damage model is constructed based on a non-homogeneous Poisson process, including:

[0031] Let the impact intensity v be a function of time t, and let the counting process satisfy the following conditions:

[0032] The increment N(t1,t2) follows a Poisson distribution with parameter v(t2-t1):

[0033]

[0034] In the formula, Let v2(t) be the integral of the impact intensity function of a lithium battery impact event over the interval [a,b]. Let n be the average number of impact events occurring within the interval [a,b], where [a,b] is a continuous time interval, t1 is the starting point of the observation window, t2 is the ending point of the observation window, n is a discrete variable, e is a natural constant, and P{N[t1,t2]=n} is the probability that the total number of events occurring within the interval [t1,t2] is n.

[0035] When n independent normally distributed random variables are added together Follows a normal distribution N(nu,nσ) 2 The probability density function is determined as follows:

[0036]

[0037] In the formula, f represents the probability density function, describing the distribution of the sum of n independent and identically distributed random variables; x represents the independent variable of the formula; and Y... i Let represent the i-th random variable, where i represents the specific position in the sequence of random variables, u represents the expected value of a single random variable, and nσ 2 This represents the variance of the sum of random variables;

[0038] The impact strength function is expressed as:

[0039] v2(t)=ax b (a, b > 0)

[0040] In one possible design, the inherent degradation of the nonlinear Wiener process is coupled with the impact damage of the nonhomogeneous Poisson process to obtain an expression for the total degradation, thus constructing a lithium battery degradation-threshold model, including:

[0041] When the inherent degradation and impact damage are independent, the total degradation energy and impact damage are determined by the following formula:

[0042] Y(t) = X(t) + S(t)

[0043]

[0044] In the formula, Y(t) is the total degradation energy at time t, S(t) is the impact damage at time t, X(t) is the inherent degradation at time t, N(t) is the number of impacts, and W(t) is the total degradation energy at time t. i ) represents the damage value corresponding to the impact, ti This refers to the specific time point when the i-th event occurs;

[0045] The degradation expression for the lithium battery degradation-threshold model is determined as follows:

[0046]

[0047] In the formula, x0 is the initial state value of the system, a is the scaling factor of the nonlinear trend, and σ B (t) is the diffusion coefficient that varies with time t, used to quantify the dynamic range of random fluctuations during degradation.

[0048] In one possible design, based on Bayesian theory, the drift parameters of the degradation model are updated using historical degradation data and real-time observation data of lithium batteries. The update process calculates the posterior distribution by combining the likelihood function of the degradation data with the prior distribution of the parameters, thereby achieving dynamic parameter correction, including:

[0049] Based on historical degradation detection data of lithium batteries X 1:k To estimate the drift parameter μ, based on Bayesian theory, the posterior distribution of the drift parameter μ is as follows:

[0050] p(μ|x 1:k )∝p(x 1:k |μ)π0(μ)

[0051] In the formula, p(μ|x 1:k p(x) represents the posterior distribution of the drift parameter μ; 1:k |μ) is the likelihood function of the degraded data when the drift parameter μ is known; π0(μ) is the prior distribution of the drift parameter μ;

[0052] Based on the distribution of the drift parameter μ and the properties of Brownian motion, the likelihood function of the degenerate data is determined by the following formula:

[0053]

[0054] In the formula, n is the discrete-time index, k is the total number of observation points, and t n Let t be the end time of the nth time interval. n-1 Let x be the starting time of the (n-1)th time interval. n For in t n The observed value at time x n-1 For in t n-1 The observation at time t, λ(t; η) is a time-dependent intensity function that describes the occurrence rate of the event at time t. The change pattern is adjusted by the parameter η, which is the shape parameter of the intensity function, and δ is the diffusion coefficient.

[0055] Based on the likelihood function of degraded data, the updated The mean and variance of the parameters are obtained using the following formulas:

[0056]

[0057] In the formula, This represents the posterior estimate of the drift parameters at the current time. The posterior estimate of the drift parameters at the previous time step. The standard deviation of the estimated value at the current time. Let be the standard deviation of the estimate at the previous time step, b be the nonlinear degradation exponent, and t be the standard deviation of the estimate at the previous time step. k t is the current observation time. k-1 x is the previous observation time. k Let x be the observed value at the current moment. k-1 This is the observation value from the previous moment.

[0058] In one possible design, based on a non-homogeneous composite Poisson process, and combining the normal distribution characteristics of the number of impacts and single impact damage, a model relating impact damage to a threshold is constructed, including:

[0059] The probability that the total damage S2(t) is less than the threshold w is:

[0060]

[0061] In the formula, N(t) represents the number of impacts, R(t) represents the reliability of the system at time t, P is the probability operator, and W(T) represents the system reliability at time t. i ) is at time T i The instantaneous damage caused by the i-th event;

[0062] If no impact event occurs during the process of battery capacity degradation to 80% of its original capacity, the reliability value is 1, which is represented as:

[0063] R(t|N(t)=0)=1

[0064] If a battery experiences a shock event during the process of its capacity degrading to 80% of its original capacity, the reliability is expressed as follows:

[0065]

[0066] In the formula, n represents the number of impact events, and Y... D This is a dynamic failure threshold;

[0067] The cumulative distribution function of the nonhomogeneous composite Poisson process is determined as follows:

[0068]

[0069] In the formula, F S(s,t) is the cumulative distribution function of the composite random process S at time t, where S is the total dynamic damage. i Let denot be the random damage amount of the i-th event, Λ(t) be the cumulative event occurrence rate function, s be the independent variable of the distribution function, nu represent the mean of the sum of damage from n events, and nσ represent the standard deviation amplification term of random damage.

[0070] The corresponding probability density function is determined as follows:

[0071]

[0072] In the formula, f S (s,t) is the probability density function of the total dynamic damage S at time t, and v2(t) is the impact intensity function.

[0073] In one possible design, the Monte Carlo simulation method is used to perform reliability analysis on the impact-threshold model. By simulating the number of impact events and the impact force, the probability density function, cumulative distribution function, and reliability curve of the total impact damage are statistically obtained, including:

[0074] The Monte Carlo simulation method was used to simulate the number of events N(t) based on the impact rate of the NHPP process. For each impact event, random impact forces were generated according to a normal distribution, and the total impact damage was obtained by accumulating them. After multiple simulations, the total impact damage was statistically analyzed to obtain the probability density function and cumulative distribution function of the non-homogeneous composite Poisson distribution. The probability density and cumulative distribution of the non-homogeneous composite Poisson distribution at different time points were obtained.

[0075] Secondly, this application provides a device for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes, the device comprising:

[0076] The nonlinear degradation modeling module is configured to: establish an inherent degradation model of lithium battery based on the nonlinear Wiener process, wherein the nonlinear Wiener process uses drift parameters in the form of power functions to characterize the nonlinear characteristics of lithium battery degradation;

[0077] The impact damage modeling module is configured to: construct a lithium battery impact damage model based on a non-homogeneous Poisson process, wherein the impact intensity of the non-homogeneous Poisson process is a function of time to characterize the non-homogeneous characteristics of the impact frequency changing with time.

[0078] The impact strength parameter identification module is configured to optimize and estimate the strength function parameters of the non-homogeneous Poisson process using a nonlinear least squares fitting method.

[0079] The dual-threshold failure mechanism definition module is configured to set a soft failure threshold w and a hard failure threshold D, where the soft failure threshold w is the amount of degradation corresponding to the capacity decay to 80% of the initial capacity, and the hard failure threshold D is the critical value of cumulative impact damage. When the total degradation of the lithium battery is ≥ w or the cumulative impact damage is ≥ D, the lithium battery is determined to be in failure.

[0080] The lithium battery degradation-threshold model construction module is configured to: couple the inherent degradation amount of the nonlinear Wiener process with the impact damage amount of the nonhomogeneous Poisson process to obtain the expression for the total degradation amount and construct the lithium battery degradation-threshold model.

[0081] The parameter dynamic update module is configured to: based on Bayesian theory, use historical degradation data and real-time observation data of lithium batteries to update the drift parameters of the degradation model. The update process calculates the posterior distribution by using the likelihood function of the degradation data and the prior distribution of the parameters, thereby realizing the dynamic correction of the parameters.

[0082] The lithium battery impact-threshold model construction module is configured to: construct a model relating impact damage to a threshold based on a non-homogeneous composite Poisson process, combined with the normal distribution characteristics of the number of impacts and single impact damage;

[0083] The reliability analysis module is configured to: use Monte Carlo simulation to perform reliability analysis on the impact-threshold model, and obtain the probability density function, cumulative distribution function and reliability curve of total impact damage by simulating the number of impact events and impact intensity, and predict the remaining life and reliability of lithium battery.

[0084] Thirdly, embodiments of this application provide an electronic device, including: at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, causing the at least one processor to execute the lithium battery degradation threshold-impact model construction method based on the Wiener and Possion processes as described in the first aspect and various possible designs of the first aspect.

[0085] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions. When a processor executes the computer-executable instructions, it implements the lithium battery degradation threshold-impact model construction method based on the Wiener and Possion processes as described in the first aspect and various possible designs of the first aspect.

[0086] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the lithium battery degradation threshold-impact model construction method based on the Wiener and Possion processes as described in the first aspect and various possible designs of the first aspect.

[0087] The method, apparatus, device, and storage medium for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes provided in this application have at least the following beneficial effects:

[0088] (1) To address the problem that traditional models ignore the coupling of external impact damage, a degradation-impact fusion modeling framework is proposed: the nonlinear Wiener process and the nonhomogeneous Poisson process are fused to quantify the interaction effect between inherent degradation and random impact; a dual-threshold failure mechanism is constructed to significantly improve the matching degree between failure modes and actual scenarios.

[0089] (2) To address the problem of isolated updates between degradation parameters and impact parameters, a dynamic linkage update mechanism is designed: By using Bayesian theory and optimization algorithms to collaboratively update degradation model parameters and impact intensity parameters, closed-loop parameter feedback is achieved, reducing the uncertainty of prediction results. Attached Figure Description

[0090] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0091] Figure 1 The flowchart of a lithium battery degradation threshold-impact model construction method based on Wiener and Possion processes provided in this application embodiment. Figure 1 ;

[0092] Figure 2 The flowchart of a lithium battery degradation threshold-impact model construction method based on Wiener and Possion processes provided in this application embodiment. Figure 2 ;

[0093] Figure 3 The actual vehicle charging data provided in this application embodiment includes a monthly charging characteristic graph;

[0094] Figure 4 The result diagram of the nonlinear least squares fitting method in parameter identification provided in the embodiments of this application;

[0095] Figure 5 The NHPP prediction results diagram provided for the embodiments of this application;

[0096] Figure 6A schematic diagram of the degradation-threshold failure process and cumulative impact model provided in the embodiments of this application; wherein, (a) is a diagram of the degradation-threshold failure process; and (b) is a cumulative impact model.

[0097] Figure 7 A diagram illustrating the parameter iteration process of Bayesian update provided in this application embodiment;

[0098] Figure 8 The following are prediction results of the nonlinear Wiener degradation model and DT degradation model provided in the embodiments of this application: (a) normal prediction result of actual vehicle No. 1; (b) prediction result of actual vehicle No. 1 considering impact.

[0099] Figure 9 The PDF curves based on the DS degradation model are provided for embodiments of this application; wherein, (a) is the PDF curve of damage within a certain month; (b) and (d) are the PDF curves of damage up to a certain month;

[0100] Figure 10 The CDF curves based on the DS degradation model are provided for embodiments of this application; wherein, (a) the damage PDF curve within a certain month; (b) and (d) the damage PDF curve up to a certain month;

[0101] Figure 11 The lithium battery reliability curve provided in the embodiments of this application;

[0102] Figure 12 This is a structural diagram of a lithium battery degradation threshold-impact model construction device based on the Wiener and Possion processes, provided in an embodiment of this application.

[0103] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0104] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0105] The collection, storage, use, processing, transmission, provision, and disclosure of financial data or user data involved in the technical solution of this application all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0106] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0107] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0108] This application provides a method for constructing a degradation threshold-impact model for lithium batteries based on the Wiener and Poisson processes, including the following steps: constructing an inherent degradation model of the lithium battery based on the nonlinear Wiener process; establishing an external impact damage model by combining the nonhomogeneous Poisson process; defining a dual-threshold failure mechanism to distinguish between capacity decay failure and impact burst failure; designing a dynamic linkage update mechanism for degradation parameters and impact parameters; and predicting the probability distribution of remaining lifetime through Monte Carlo simulation. This method addresses the challenge of insufficient quantification of the coupled failure mechanism of inherent aging and external impact in lithium batteries under complex operating conditions. It proposes a degradation-threshold-impact (DTS) joint modeling framework, realizes the construction of the RUL prediction framework, and completes the output of the probability density function of remaining lifetime and reliability assessment results.

[0109] Specifically, such as Figure 1 As shown, the method for constructing the lithium battery degradation threshold-impact model based on the Wiener and Possion processes includes the following steps S10-S80.

[0110] S10: Nonlinear degradation modeling, including: establishing an inherent degradation model of lithium battery based on the nonlinear Wiener process, wherein the nonlinear Wiener process uses drift parameters in the form of power functions to characterize the nonlinear characteristics of lithium battery degradation.

[0111] In this embodiment, an inherent degradation model of lithium battery is established based on the nonlinear Wiener process, providing inherent degradation basis data for subsequent total degradation modeling of coupled impact damage.

[0112] In some embodiments, the specific method of step S10 is as follows:

[0113] Considering the nonlinear characteristics exhibited during lithium battery degradation, a nonlinear Wiener degradation model is established using drift parameters in the form of a power function. The degradation amount of the lithium battery at time t can then be expressed as:

[0114]

[0115] In the formula, B(t) represents Brownian motion, reflecting dynamic uncertainty. λ(t; θ) is the drift coefficient of the degradation process; x0 is the degradation amount of the lithium battery at the initial moment, which is generally 0; θ is the parameter vector of the nonlinear function, characterizing the nonlinear characteristics of lithium battery degradation; when the drift coefficient λ(t; θ) is λ, it is a linear Wiener model. λ(t; θ) is defined as a power function abt. t-1 The form is X(t) = x0 + at b +δ B B(t).

[0116] Assumption 1: If the lithium battery is working properly at time 1, then no failure event has occurred before time 1.

[0117] Assumption 2: If at time t, the corresponding degradation reaches the failure threshold w, then it is assumed that the probability of reaching the failure threshold w in the random process before time t is negligible.

[0118] Theorem 1: For the lithium battery degradation process, given Assumption 1 and Assumption 2, the probability density function of the first arrival time crossing the failure threshold w is:

[0119]

[0120] In the formula, f T|a (t|a) is the conditional probability density function of lifetime T given parameter a; S(t) is the cumulative degradation function; λ(t; θ) is the time-varying degradation rate function; σ B is the diffusion coefficient; t is the time variable.

[0121] Theorem 2: For the degradation process, given Assumption 1 and Assumption 2, the probability density function of the first arrival time of {X(t), t≥0} crossing the failure threshold w is:

[0122]

[0123] In the formula, θ represents the unknown parameter (a, b).

[0124] S20: Impact damage modeling, including: constructing a lithium battery impact damage model based on a non-homogeneous Poisson process, wherein the impact intensity of the non-homogeneous Poisson process is a function of time to characterize the non-homogeneous characteristics of the impact frequency changing with time.

[0125] In this embodiment, a lithium battery impact damage model is constructed based on the Poisson process. This model, together with the inherent degradation model in step S10, forms a dual-factor model for lithium battery performance degradation, serving as input for the overall degradation analysis.

[0126] In some embodiments, the specific method of step S20 is as follows:

[0127] Modeling of nonhomogeneous composite Poisson distribution:

[0128] The non-homogeneous Poisson process (NHPP) is a generalization of the Poisson distribution. Unlike the HPP distribution, it describes a situation where the occurrence rate of a shock event is not constant but a function of time within a given interval. In the definition of a Poisson process, if the shock intensity *v* is a function of time *t*, and the counting process satisfies the following conditions:

[0129] (1) N(0) = 0.

[0130] (2) N(t) is an independent increment process.

[0131] (3)P{N(h)=1}=v2(t)h+o(h).

[0132] (4) P{N(h)≥2}=o(h).

[0133] For any 0 ≤ t1 ≤ t2, the corresponding increment N(t1, t2) follows a Poisson distribution with parameter v(t2 - t1), i.e.:

[0134]

[0135] In the formula, Let v1 be the integral of the impact intensity function v2(t) of a lithium battery impact event over the interval [a,b]. Let n be the average number of impact events occurring within the interval [a,b]. Let [a,b] be a continuous time interval, t1 be the starting point of the observation window, t2 be the ending point of the observation window, n be a discrete variable, e be a natural constant, and P{N[t1,t2]=n} be the probability that the total number of events occurring within the interval [t1,t2] is n.

[0136] The Poisson process {N(t), t≥0} is called an NHPP process with intensity function v(t). Let v2(t) be the integral of the rate function of lithium battery impact events over the interval [a,b], representing the average number of impact events occurring within the interval [a,b]. In fact, the HPP distribution is a special case of the NHPP distribution. When v2(t) = v1 (a constant), the NHPP distribution degenerates into the HPP distribution. Therefore, the mean and variance functions of the NHPP distribution are as follows:

[0137]

[0138] Therefore, similar to the HPP distribution, the variance of the NHPP distribution is equal to its expectation.

[0139] The main difference between the HPP and NHPP distributions lies in their intensity. Since the damage distribution pattern for each event is consistent, the difference between the homogeneous composite Poisson distribution and the non-homogeneous composite Poisson distribution is the difference in v1 and v2 values. Therefore, the mean and variance of the non-homogeneous composite Poisson distribution are:

[0140]

[0141] Let the probability density function of S2 be f S(x) Then we have:

[0142]

[0143] In the formula The probability density function.

[0144] Because Y i ~N(u,σ 2 When n independent normally distributed random variables are added together, Follows a normal distribution N(nu,nσ) 2 Its probability density function is:

[0145]

[0146] Let P(N([a,b])=n) and Substituting into the above formula, we get:

[0147]

[0148] By analyzing real vehicle data, the impact strength function can be set as follows:

[0149] v2(t)=ax b (a, b > 0)

[0150] S30: Impact strength parameter identification, including: optimizing the estimation of the strength function parameters of the nonhomogeneous Poisson process by using nonlinear least squares fitting and minimizing the negative log-likelihood function estimation method.

[0151] In this embodiment, the parameters of the nonhomogeneous Poisson process strength model are optimized by nonlinear least squares method and maximum likelihood estimation method, respectively. The optimization results are used to improve the accuracy of the impact damage model constructed in step S20.

[0152] In some embodiments, the specific method of step S30 is as follows:

[0153] The incidence function in the NHPP model is v2(t) = ax b (a,b>0), the parameters to be identified are a and b, and the model parameters are estimated by nonlinear least squares fitting method and minimizing negative log-likelihood function estimation method.

[0154] Nonlinear least squares fitting method

[0155] The nonlinear least squares fitting method is essentially a method to find the optimal parameters by minimizing the sum of squared residuals between the predicted and actual values. It is defined as a model error function, which is based on the sum of squared differences between the model's predicted values ​​f1(x) and observed values ​​f2(x), i.e.:

[0156]

[0157] In the formula: y — the actual value of the data;

[0158] a1, b1 — Model recognition parameter values.

[0159] Secondly, the initial parameter values ​​of a1 and b1 are set to 0.5. The "fminsearch" function is used to minimize the error function F(x) to find the optimal model parameters p. The basic principle is to use...

[0160] The Nelder-Mead simplex algorithm is used to find the minimum value. The main principle of the simplex method is to find the optimal solution by moving along the vertices of the feasible region. In two-dimensional space, these vertices form a polygon, while in higher-dimensional space, they form a polyhedron. The algorithm starts from an initial feasible vertex and then moves to adjacent vertices along the direction that the objective function value decreases until the optimal solution of the model is found or the problem is determined to be unbounded.

[0161] S40: Definition of dual-threshold failure mechanism, including: setting soft failure threshold w and hard failure threshold D, where soft failure threshold w is the amount of degradation corresponding to capacity decay to 80% of the initial capacity, and hard failure threshold D is the critical value of cumulative impact damage. When the total degradation of the lithium battery is ≥ w or the cumulative impact damage is ≥ D, the lithium battery is determined to be in failure.

[0162] In this embodiment, a soft failure threshold w (capacity decay threshold) and a hard failure threshold D (cumulative impact damage threshold) are set to provide a clear standard for the model to determine the failure state of the lithium battery in steps S50 and S70.

[0163] In some embodiments, the specific method of step S40 is as follows:

[0164] The dual-threshold failure mechanism is defined as follows: a soft failure threshold w (capacity decay threshold) and a hard failure threshold D (cumulative impact damage threshold) are set. Battery failure is determined when Y(t)≥w or ∑S(t)≥D. Damage to the power battery system is the result of competition between soft and hard failures. On the one hand, soft failure occurs when the degradation performance index exceeds the set threshold level, and external random impacts can further lead to performance degradation. On the other hand, when the magnitude of the cumulative impact exceeds its threshold level, it is considered a hard failure. A cumulative impact model can be used to model this. Both threshold levels are positively correlated with time-dependent degradation performance. This paper uses a nonlinear Wiener process to model the continuous degradation process of the battery, while the NHPP process is used to handle the occurrence of random fast charging times during charging.

[0165] The soft failure threshold is defined as follows: the soft failure threshold level w is a constant. The curve within the time period 0 to t1, ..., t3 to t4 represents the performance loss caused by normal charging and discharging of the battery. The jump at the point t1 to t4 represents the performance loss caused by impact. When the sum of the normal degradation loss X(t) and the loss S(t) caused by impact, the total degradation loss Y(t) exceeds the threshold w, it indicates that the battery performance degradation value has reached 80% of the initial value, and the battery is considered to have experienced a soft failure. The principle is as follows: Figure 5 As shown.

[0166] Hard failure threshold definition: The cumulative impact model refers to a system that, when subjected to external random impacts, does not fail immediately. Instead, the damage gradually increases with each impact until the cumulative damage reaches or exceeds a set failure threshold, at which point the system fails. Each impact causes varying degrees of damage to the battery. Over time, when the cumulative damage from all arriving random impacts exceeds or equals the set threshold level w, hard failure occurs, signifying the end of the battery pack's lifespan. The principle is detailed in the appendix. Figure 4 As shown.

[0167] S50: Construction of lithium battery degradation-threshold model, including: coupling the inherent degradation amount of the nonlinear Wiener process with the impact damage amount of the nonhomogeneous Poisson process to obtain the expression of the total degradation amount, and constructing the lithium battery degradation-threshold model.

[0168] In this embodiment, a lithium battery degradation-threshold (DT) model is constructed, which combines the inherent degradation model of step S10, the impact damage model optimized by step S30 in step S20, and the dual-threshold failure mechanism in step S40 to establish the correlation between degradation amount and failure threshold.

[0169] In some embodiments, the specific method of step S5 is as follows:

[0170] In the degradation threshold model, the capacity degradation process also depends on random impacts. To some extent, each impact suddenly increases the cumulative impact damage S(t), thereby increasing the overall degradation performance. Based on the physical failure analysis of MEMS systems, similar to the reliability models of competing failure processes in many related studies, this study assumes that X(t) and S(t) are independent. The total degradation energy Y(t) and S(t) are described as follows:

[0171] Y(t) = X(t) + S(t)

[0172]

[0173] In the formula, Y(t) is the total degradation energy at time t, S(t) is the impact damage at time t, X(t) is the inherent degradation at time t, N(t) is the number of impacts, and W(t) is the total degradation energy at time t. i ) represents the damage value corresponding to the impact, t i Let be the specific time point when the i-th event occurs.

[0174] As can be seen from the preceding text, the total loss caused by the nonlinear Wiener degradation process and the shock is:

[0175] X(t) = x0 + at b +σ B (t)

[0176]

[0177] Therefore, the degenerate expression of the DT model can be obtained as:

[0178]

[0179] In the formula, x0 is the initial state value of the system, a is the scaling factor of the nonlinear trend, and σ B (t) is the diffusion coefficient that varies with time t, used to quantify the dynamic range of random fluctuations during degradation.

[0180] S60: Parameter dynamic update, including: based on Bayesian theory, using historical degradation data and real-time observation data of lithium batteries, updating the drift parameters of the degradation model. The update process calculates the posterior distribution by using the likelihood function of the degradation data and the prior distribution of the parameters, thereby realizing the dynamic correction of the parameters.

[0181] In this embodiment, the degradation-threshold model parameter update and model result analysis based on Bayesian theory are performed: Bayesian theory is used to dynamically update the degradation parameters in the degradation-threshold model constructed in step S50, so as to improve the real-time performance and accuracy of the model in describing the degradation process of lithium batteries.

[0182] In some embodiments, the specific method of step S60 is as follows:

[0183] Dynamic parameter update mechanism:

[0184] The drift parameter μ can be obtained by using historical degradation detection data X of lithium batteries. 1:k =(X1,X2,…X) k Therefore, during the lifespan of a lithium battery, t... k Time, based on t k Previous degradation data and t k The data observed at each moment can be used to update the drift parameter

[58] . According to Bayesian theory, the posterior distribution of the drift parameter μ is as follows:

[0185] p(μ|x 1:k )∝p(x 1:k |μ)π0(μ)

[0186] In the formula, p(μ|x 1:k p(x) represents the posterior distribution of the drift parameter μ; 1:k |μ) is the likelihood function of the degraded data when the drift parameter μ is known; π0(μ) is the prior distribution of the drift parameter μ.

[0187] Based on the distribution of μ and the properties of Brownian motion, the likelihood function of degenerate data can be obtained as follows:

[0188]

[0189] In the formula, n is the discrete-time index, k is the total number of observation points, and t n Let t be the end time of the nth time interval. n-1 Let x be the starting time of the (n-1)th time interval. n For in t n The observed value at time x n-1 For in t n-1 The observed value at time t, λ(t; η) is a time-dependent intensity function that describes the occurrence rate of the event at time t. The change pattern is adjusted by the parameter η, which is the shape parameter of the intensity function, and δ is the diffusion coefficient.

[0190] Therefore, based on the above formula, the updated The mean and variance of the parameters can be obtained using the following formulas:

[0191]

[0192] In the formula, This represents the posterior estimate of the drift parameters at the current time. The posterior estimate of the drift parameters at the previous time step. The standard deviation of the estimated value at the current time. Let be the standard deviation of the estimate at the previous time step, b be the nonlinear degradation exponent, and t be the standard deviation of the estimate at the previous time step. k t is the current observation time. k-1 x is the previous observation time. k Let x be the observed value at the current moment. k-1 This is the observation value from the previous moment.

[0193] As can be seen from the above formula, once at time t... k By observing the degraded capacity of lithium batteries, the drift parameters can be updated by inputting this data.

[0194] S70: Construction of lithium battery impact-threshold model, including: based on non-homogeneous composite Poisson process, combined with the normal distribution characteristics of impact number and single impact damage, to construct a model relating impact damage to threshold.

[0195] In this embodiment, the lithium battery impact threshold model (DS) is constructed based on the impact damage model in step S20, the optimized parameters in step S30, and the hard failure threshold in step S40, establishing the correlation between impact damage and failure threshold.

[0196] In some embodiments, the specific method of step S70 is as follows:

[0197] The non-homogeneous composite Poisson model was chosen as the ST model to predict the battery structure. Assume N... i (i = 1, 2, 3, ...) represents the number of impacts in a given month, N j (j = 1, 2, 3, ...) represents the number of impacts up to a certain month. Therefore, the expression for the total damage up to a certain month and up to that month is:

[0198]

[0199] For a non-homogeneous composite Poisson distribution, the impact force follows a normal distribution. The probability that the total damage S²(t) is less than a certain threshold w is:

[0200]

[0201] In the formula, N(t) represents the number of impacts, R(t) represents the reliability of the system at time t, P is the probability operator, and W(T) represents the system reliability at time t. i ) is at time T i The instantaneous damage caused by the i-th event.

[0202] It can be seen that if the battery has not experienced any impact events until its capacity degrades to 80% of its original capacity, the reliability value is 1, which is represented as:

[0203] R(t|N(t)=0)=1

[0204] If the number of fast charging events that occur during this period is n, i.e., N(t) = n > 0, then the reliability can be expressed as:

[0205]

[0206] In the formula, n represents the number of impact events, and Y... D This is the dynamic failure threshold.

[0207] From the above formula, the cumulative distribution function of the non-homogeneous composite Poisson process can be obtained as follows:

[0208]

[0209] In the formula, F S (s,t) is the cumulative distribution function of the composite random process S at time t, where S is the total dynamic damage. i Let denot be the random damage amount of the i-th event, Λ(t) be the cumulative event occurrence rate function, s be the independent variable of the distribution function, nu represent the mean of the sum of damage from n events, and nσ represent the amplification term of the standard deviation of random damage.

[0210] The corresponding probability density function is:

[0211]

[0212] In the formula, f S (s,t) is the probability density function of the total dynamic damage S at time t, and v2(t) is the impact intensity function.

[0213] S80: Reliability analysis, including: using Monte Carlo simulation to perform reliability analysis on the impact-threshold model, by simulating the number of impact events and the impact force, to statistically obtain the probability density function, cumulative distribution function and reliability curve of total impact damage, and to predict the remaining life and reliability of lithium batteries.

[0214] In this embodiment, the reliability analysis of the lithium battery impact-threshold model based on Monte Carlo simulation analysis is performed: for the impact-threshold model constructed in step S70, its reliability is analyzed by Monte Carlo simulation to provide a probability distribution basis for predicting the remaining life of the lithium battery.

[0215] In some embodiments, the specific method of step S80 is as follows:

[0216] By modeling the nonhomogeneous composite Poisson impact damage, the impact velocity v2 = 0.2651x 0.3387Furthermore, it is known that the impact of each shock follows a normal distribution. The cumulative distribution function and probability density function of its non-homogeneous composite Poisson process contain nested infinite series and normal cumulative functions. Given the high complexity of its mathematical expression, traditional analytical methods struggle to efficiently and accurately obtain the relevant characteristics and parameter estimates of this distribution. Simulation methods, however, can obtain approximate but reliable results within a reasonable timeframe by simulating numerous real-world scenarios, thus providing strong support for the research. Therefore, simulation is used for prediction.

[0217] The Monte Carlo simulation method was used, with N = 10,000 simulations. In each simulation, the number of events N(t) was simulated based on the impact rate of the NHPP process. For each impact event, a random impact force was generated according to a normal distribution, and then accumulated to obtain the total impact effect S(t). After multiple simulations, the obtained S(t) data were statistically analyzed to approximate the probability density function and cumulative distribution function of the non-homogeneous composite Poisson distribution. The probability density and cumulative distribution of the non-homogeneous composite Poisson distribution at different time points were obtained. In addition, by setting the threshold of its ST model to 16Ah, the reliability curve can be obtained.

[0218] Example 2:

[0219] This application provides a method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes, such as... Figure 2 As shown, the method includes the following steps S1-S7.

[0220] S1: Establishing an inherent degradation model for lithium batteries:

[0221] X(t) = x0 + at b +δ B B(t)

[0222] in:

[0223] at b The power function drift term (a>0, b∈[0.3,1.2]) describes the nonlinear characteristics of capacity decay (goodness of fit R²>0.95 when b=1.3);

[0224] δ B B(t) is the diffusion term (δ) B =0.003), representing individual differences and measurement noise;

[0225] B(t) represents standard Brownian motion.

[0226] Parameter initialization: Drift parameter a ~ N(0.12, 0.022), diffusion coefficient δ B=0.003, shape parameter b=1.3 (calibrated based on NASA B0005 battery degradation data).

[0227] S2: Construction of the non-homogeneous composite Poisson model.

[0228] The external impact damage model is established as follows:

[0229]

[0230] The number of impact events N(t) follows an intensity function v2(t) = at b The NHPP process.

[0231] S3: Data input.

[0232] Using real-vehicle fast charging event time series, an event with a current ≥55A was defined as an impact event. Statistical analysis of real-vehicle data yielded the following data on the number of fast charges and total number of charges for the vehicle battery over two and a half years of operation. For ease of analysis, the time is statistically analyzed by month; for example, January of the second year is considered the 13th month. Figure 3 As shown.

[0233] S4: Impact strength parameter identification.

[0234] 1) Nonlinear least squares fitting.

[0235] Objective function:

[0236]

[0237] Output the fitting parameters P = [a, b], and plot the fitting curve and residual plot, as follows. Figure 4 As shown.

[0238] Predict the number of fast charges per month and plot the NHPP prediction results against the actual values, such as... Figure 5 As shown.

[0239] Calculate the model indicators.

[0240] S5: Definition of dual-threshold failure mechanism.

[0241] 1) Failure threshold setting:

[0242] The soft failure threshold w = 80% × Q init (capacity decays to 80% of the initial capacity);

[0243] Hard failure threshold D ~ N(0.15, 0.022) (cumulative impact damage threshold).

[0244] 2) Failure determination criteria:

[0245] The battery is considered to be faulty when Y(t) = X(t) + S(t) ≥ w or S(t) ≥ D.

[0246] S6: Constructing a degradation-threshold model:

[0247]

[0248] S7: Dynamic parameter update mechanism based on Bayesian update theory.

[0249] Lifespan of lithium batteries (t) k Time, based on t k Previous degradation data and t k The data observed at each step can be used to update the drift parameter. According to Bayesian theory, the posterior distribution of the drift parameter μ is as follows:

[0250] p(μ|x 1:k )∝p(x 1:k |μ)π0(μ)

[0251] In the formula p(μ|x 1:k —The posterior distribution of the random parameter μ;

[0252] p(x 1:k |μ)——The likelihood function of degenerate data under the assumption that the drift parameter μ is known;

[0253] π0(μ) — the prior distribution of the drift parameter μ.

[0254] Based on the distribution of μ and the properties of Brownian motion, the likelihood function of degenerate data can be obtained as follows:

[0255]

[0256] Therefore, based on the likelihood function and the posterior distribution of the drift parameter μ of the degraded data, the updated... The mean and variance of the parameters can be obtained using the following formulas:

[0257]

[0258] As can be seen from the above formula, once at time t... k By observing the degradation capacity of the lithium battery, the drift parameter μ can be updated by inputting this data.

[0259] Analysis of degradation-threshold model results.

[0260] To demonstrate the effectiveness of the proposed method, the dataset used is the data from vehicles 8, 11, and 20 in the real vehicle data set as the training dataset for the model. This better reflects the differences between individuals in the nonlinear Wiener degradation process. For vehicle 1, the parameters in the prediction model are still identified using MLE, and the iterative process is as follows: Figure 7 As shown in Table 1, the parameter estimation results are as follows.

[0261] Table 1 Parameter estimates

[0262] Model parameters u <![CDATA[σ a ]]> <![CDATA[σ B ]]> b Parameter estimates 0.1923 0.66e-3 3.6117e-18 0.7455 Bayesian parameter final value 0.0443 1.3387e-4 2.1470e-23 1.0367

[0263] Substituting its parameters and data into the degradation expression of the DT model yields the predicted RUL values ​​of lithium-ion batteries for both the nonlinear Wiener degradation model and the DT degradation model, such as... Figure 8 As shown.

[0264] Impact threshold model construction.

[0265] Based on the S2-step NHPP model for predicting the number of impacts, this embodiment uses the non-homogeneous composite Poisson model as the ST model to predict the battery number. Assuming N... i (i = 1, 2, 3, ...) represents the number of impacts in a given month, N j (j = 1, 2, 3, ...) represents the number of impacts up to a certain month. Therefore, the expression for the total damage up to a certain month and up to that month is:

[0266]

[0267] For a non-homogeneous composite Poisson distribution, the impact force follows a normal distribution. The probability that the total damage S²(t) is less than a certain threshold w is:

[0268]

[0269] In the formula, N(t) represents the number of fast charging cycles.

[0270] W(t) — Damage caused by each fast charge.

[0271] It can be seen that if the battery has not experienced any impact events until its capacity degrades to 80% of its original capacity, the reliability value is 1, which is represented as:

[0272] R(t|N(t)=0)=1

[0273] If the number of fast charging events that occur during this period is n, i.e., N(t) = n > 0, then the reliability can be expressed as:

[0274]

[0275] The cumulative distribution function of a nonhomogeneous composite Poisson process, derived from the Poisson distribution formula, is:

[0276]

[0277] The corresponding probability density function is:

[0278]

[0279] Impact-threshold model simulation analysis.

[0280] Based on step S2, the impact velocity v2 = 0.2651x 0.3387 Furthermore, it is known that the impact of each shock follows a normal distribution, and its equations (5-41) and (5-42) contain nested infinite series and normal cumulative functions. Given the high complexity of its mathematical expression, simulation is used for prediction.

[0281] The Monte Carlo simulation method was adopted, and the number of simulations was set to N = 10,000. In each simulation, the number of events N(t) was simulated based on the impact rate of the NHPP process. For each impact event, a random impact force was generated according to a normal distribution, and then the total impact effect S(t) was obtained by summing them up.

[0282] After multiple simulations, statistical analysis was performed on the obtained S(t) data to approximate the probability density function and cumulative distribution function of the non-homogeneous composite Poisson distribution. The approximate results of the probability density and cumulative distribution of the non-homogeneous composite Poisson distribution at different time points are as follows: Figure 9 and Figure 10 As shown, by setting the threshold of its ST model to 16Ah, the reliability curve can be obtained as follows. Figure 11 As shown.

[0283] Figure 9 and Figure 10 The document presents PDF and CDF graphs showing battery damage caused by impact in a specific month, as well as damage up to a specific month. The PDF graphs show the damage in months 8, 16, and 24, illustrating the probability density of different damage levels at that particular point in time. The CDF graphs show the fluctuating trend of the cumulative probability of overall battery damage over time. These curves can be used to assess the probability of a certain degree of battery damage at different time points, providing crucial reference for battery reliability analysis, lifespan prediction, and maintenance strategy planning.

[0284] Figure 11This is a reliability curve showing the lifespan of a lithium battery due to impact interference from month 1 to month 27. By using the reliability curve, the failure trend of lithium batteries under fast charging conditions can be accurately observed, effectively improving the safety and stability of lithium battery applications.

[0285] Example 3:

[0286] This application also provides a device for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes, such as... Figure 12 As shown, the apparatus for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes includes:

[0287] The nonlinear degradation modeling module 1201 is configured to: establish an inherent degradation model of lithium battery based on a nonlinear Wiener process, wherein the nonlinear Wiener process uses drift parameters in the form of power functions to characterize the nonlinear characteristics of lithium battery degradation;

[0288] The impact damage modeling module 1202 is configured to: construct a lithium battery impact damage model based on a non-homogeneous Poisson process, wherein the impact intensity of the non-homogeneous Poisson process is a function of time to characterize the non-homogeneous characteristics of the impact frequency changing with time.

[0289] The impact strength parameter identification module 1203 is configured to optimize and estimate the strength function parameters of the nonhomogeneous Poisson process by using the nonlinear least squares fitting method and the method of minimizing the negative log-likelihood function estimation.

[0290] The dual-threshold failure mechanism definition module 1204 is configured to: set a soft failure threshold w and a hard failure threshold D, where the soft failure threshold w is the amount of degradation corresponding to the capacity decay to 80% of the initial capacity, and the hard failure threshold D is the critical value of cumulative impact damage. When the total degradation of the lithium battery is ≥ w or the cumulative impact damage is ≥ D, the lithium battery is determined to be in failure.

[0291] The lithium battery degradation-threshold model construction module 1205 is configured to: couple the inherent degradation amount of the nonlinear Wiener process with the impact damage amount of the nonhomogeneous Poisson process to obtain the expression for the total degradation amount and construct the lithium battery degradation-threshold model.

[0292] The parameter dynamic update module 1206 is configured to: based on Bayesian theory, use historical degradation data and real-time observation data of lithium batteries to update the drift parameters of the degradation model. The update process calculates the posterior distribution by using the likelihood function of the degradation data and the prior distribution of the parameters, thereby realizing the dynamic correction of the parameters.

[0293] The lithium battery impact-threshold model construction module 1207 is configured as follows: based on non-homogeneous composite

[0294] The Poisson process combines the number of impacts with the normal distribution characteristics of single impact damage to construct a model relating impact damage to a threshold.

[0295] The reliability analysis module 1208 is configured to: perform reliability analysis on the impact-threshold model using the Monte Carlo simulation method; by simulating the number of impact events and the impact force, statistically obtain the probability density function, cumulative distribution function and reliability curve of total impact damage; and predict the remaining life and reliability of the lithium battery.

[0296] This application provides an electronic device. The electronic device may include a processor and a memory, wherein the processor and the memory can communicate; exemplarily, the processor and the memory communicate via a communication bus.

[0297] The processor executes computer execution instructions stored in memory, causing the processor to perform the scheme in the above embodiments. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.

[0298] The communication bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into address bus, data bus, control bus, etc. Transceivers are used to enable communication between database access devices and other computers (e.g., clients, read-write libraries, and read-only libraries). Memory may include random access memory (RAM) and may also include non-volatile memory.

[0299] The electronic device provided in this application embodiment can be the terminal device described in the above embodiments.

[0300] This application also provides a computer-readable storage medium storing computer instructions. When the computer instructions are executed on a computer, the computer performs the technical solution of the lithium battery degradation threshold-impact model construction method based on the Wiener and Possion processes described in the above embodiments.

[0301] This application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. At least one processor can read the computer program from the computer-readable storage medium. When the at least one processor executes the computer program, it can implement the technical solution of the lithium battery degradation threshold-impact model construction method based on the Wiener and Possion processes in the above embodiments.

[0302] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.

[0303] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to implement the solution of this embodiment according to actual needs.

[0304] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.

[0305] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application.

[0306] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0307] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk or optical disc, etc.

[0308] Buses can be Industry Standard Architecture (ISA) buses, Peripheral Component Interconnect (PCI) buses, or Extended Industry Standard Architecture (EISA) buses, etc. Buses can be categorized into address buses, data buses, control buses, etc.

[0309] The aforementioned storage medium can be implemented from any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium accessible to general-purpose or special-purpose computers.

[0310] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. The processor and storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic control unit or main control device.

[0311] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0312] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes, characterized in that, The method includes: Nonlinear degradation modeling includes: establishing an inherent degradation model of lithium batteries based on a nonlinear Wiener process, wherein the nonlinear Wiener process uses drift parameters in the form of power functions to characterize the nonlinear characteristics of lithium battery degradation; Impact damage modeling includes: constructing a lithium battery impact damage model based on a non-homogeneous Poisson process, wherein the impact intensity of the non-homogeneous Poisson process is a function of time to characterize the non-homogeneous characteristics of the impact frequency changing with time. Impact strength parameter identification includes: optimizing and estimating the strength function parameters of the nonhomogeneous Poisson process using nonlinear least squares fitting and minimizing negative log-likelihood function estimation methods; The dual-threshold failure mechanism is defined as follows: setting a soft failure threshold w and a hard failure threshold D, where the soft failure threshold w is the amount of degradation corresponding to the capacity decay to 80% of the initial capacity, and the hard failure threshold D is the critical value of cumulative impact damage. When the total degradation of the lithium battery is ≥ w or the cumulative impact damage is ≥ D, the lithium battery is determined to be in failure. The lithium battery degradation-threshold model is constructed by coupling the inherent degradation amount of the nonlinear Wiener process with the impact damage amount of the nonhomogeneous Poisson process to obtain the expression for the total degradation amount and construct the lithium battery degradation-threshold model. Dynamic parameter updates include: based on Bayesian theory, using historical degradation data and real-time observation data of lithium batteries, updating the drift parameters of the degradation model. The update process calculates the posterior distribution by using the likelihood function of the degradation data and the prior distribution of the parameters, thereby achieving dynamic correction of the parameters. The lithium battery impact-threshold model is constructed, including: based on the non-homogeneous composite Poisson process, combined with the normal distribution characteristics of the number of impacts and single impact damage, a model is constructed to correlate impact damage with the threshold. Reliability analysis includes: using Monte Carlo simulation to perform reliability analysis on the impact-threshold model, simulating the number of impact events and impact intensity, statistically obtaining the probability density function, cumulative distribution function and reliability curve of total impact damage, and predicting the remaining life and reliability of lithium batteries; By coupling the inherent degradation of the nonlinear Wiener process with the impact damage of the nonhomogeneous Poisson process, an expression for the total degradation is obtained, and a lithium battery degradation-threshold model is constructed, including: When the inherent degradation and impact damage are independent, the total degradation energy and impact damage are determined by the following formula: In the formula, Y ( t )for t Total degenerative energy at any given moment S ( t )for t The amount of impact damage at any given moment. X ( t )for t The inherent degradation at any given moment, N ( t () represents the number of impacts. The damage value corresponding to the impact. t i This refers to the specific time point when the i-th event occurs; The degradation expression for the lithium battery degradation-threshold model is determined as follows: In the formula, x 0 represents the initial state value of the system. a , is the scaling factor for nonlinear trends. The diffusion coefficient varies with time t and is used to quantify the dynamic amplitude of random fluctuations during degradation.

2. The method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes as described in claim 1, characterized in that, A lithium battery inherent degradation model is established based on the nonlinear Wiener process, including: A nonlinear Wiener degradation model is established using drift parameters in power function form as an intrinsic degradation model; wherein, the intrinsic degradation model is expressed as: In the formula, X ( t (t) represents the amount of lithium battery degradation at time t. The drift coefficient during the degradation process; This represents the initial degradation level of the lithium battery. θ The parameter vector represents a nonlinear function, characterizing the nonlinear features of lithium battery degradation. B ( t () is Brownian motion, reflecting dynamic uncertainty; Will Represented as a power function The form is obtained as follows: In the formula, a The scaling factor represents the nonlinear degradation trend and determines the overall amplitude and rate of the degradation curve. b The term is an exponential term for time t, controlling the nonlinear characteristics of the degradation rate. Let be the diffusion coefficient of Brownian motion (B(t)), which quantifies the intensity of random fluctuations; The inherent degradation model crosses the failure threshold. The probability density function of the first arrival time Represented as: In the formula, T Let be a continuous random variable, representing the remaining lifespan of the lithium battery, and exp be an exponential function with the natural constant as its base.

3. The method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes as described in claim 1, characterized in that, A lithium battery impact damage model is constructed based on the non-homogeneous Poisson process, including: Set impact strength v It is time t The function is defined such that the counting process satisfies the following conditions: Increment Obtain the parameter as Poisson distribution: In the formula, Indicates the interval Impact intensity function of lithium battery impact event The integral represents the integral over the interval The average number of internal shock events, where [a,b] is a continuous time interval. t 1 is the starting point of the observation window. t 2 is the end point of the observation window. n For discrete variables, e It is a natural constant. In order to be in The probability that the total number of events occurring within the interval is n; When n independent normally distributed random variables are added together Follows a normal distribution The probability density function is determined as follows: In the formula, f This represents the probability density function, describing the distribution of the sum of n independent and identically distributed random variables. x This represents the independent variable in the formula. Y i Indicates the first i A random variable, i Indicates the specific position in the sequence of random variables. u Represents the expected value of a single random variable. This represents the variance of the sum of random variables; The impact strength function is expressed as: 。 4. The method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes according to claim 1, characterized in that, Based on Bayesian theory, this study updates the drift parameters of the degradation model using historical degradation data and real-time observation data of lithium batteries. The update process calculates the posterior distribution by combining the likelihood function of the degradation data with the prior distribution of the parameters, thus achieving dynamic parameter correction. This includes: Based on historical degradation detection data of lithium batteries To estimate the drift parameter μ, based on Bayesian theory, the posterior distribution of the drift parameter μ is as follows: In the formula, Let μ be the posterior distribution of the drift parameter μ. The likelihood function of the degraded data when the drift parameter μ is known; Let μ be the prior distribution of the drift parameter μ. Based on the distribution of the drift parameter μ and the properties of Brownian motion, the likelihood function of the degenerate data is determined by the following formula: In the formula, n For discrete-time indexing, k The total number of observation points. t n Let n be the end time of the nth time interval. t n-1 This represents the starting time of the (n-1)th time interval. x n In order to be in t n The observed value at time, x n-1 In order to be in t n-1 The observed value at time, This is a time-dependent intensity function describing the occurrence rate of an event at time t, with the change pattern adjusted by the parameter η. The shape parameter of the intensity function. The diffusion coefficient is denoted as . Based on the likelihood function of degraded data, the updated The mean and variance of the parameters are obtained using the following formulas: In the formula, The posterior estimate of the drift parameter at the current time. These are the posterior estimates of the drift parameters from the previous time step. The standard deviation of the estimated value at the current time. The standard deviation of the estimated value at the previous time step. b It is a non-linear degradation index. t k For the current observation time, t k-1 For the previous observation time, x k The observed value at the current moment, x k-1 This is the observation value from the previous moment.

5. The method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes according to claim 1, characterized in that, Based on the non-homogeneous composite Poisson process, and combining the normal distribution characteristics of the number of impacts and single impact damage, a model is constructed to correlate impact damage with a threshold, including: Determine total damage The probability of being less than the threshold w is: In the formula, N ( t () represents the number of impacts. R ( t Let t represent the reliability of the system at time t, and P be the probability operator. W ( T i ) for at time T i The instantaneous damage caused by the i-th event; If no impact event occurs during the process of battery capacity degradation to 80% of its original capacity, the reliability value is 1, which is represented as: If a battery experiences a shock event during the process of its capacity degrading to 80% of its original capacity, the reliability is expressed as follows: In the formula, n represents the number of impact events. Y D This is a dynamic failure threshold; The cumulative distribution function of the nonhomogeneous composite Poisson process is determined as follows: In the formula, F S ( s , t Let be the cumulative distribution function of the composite random process S at time t. S The total amount of dynamic damage. S i Let be the random damage amount of the i-th event. This is a function for the cumulative event occurrence rate. s Let be the independent variable of the distribution function. nu Let represent the mean of the total damage from n events. This represents the amplified term of the standard deviation of random damage; The corresponding probability density function is determined as follows: In the formula, f S ( s , t Let be the probability density function of the total dynamic damage S at time t. v 2( t ) is the impact strength function.

6. The method for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes according to claim 1, characterized in that, The Monte Carlo simulation method was used to perform reliability analysis on the impact-threshold model. By simulating the number of impact events and the impact force, the probability density function, cumulative distribution function, and reliability curve of the total impact damage were statistically obtained, including: The Monte Carlo simulation method was used to simulate the number of events based on the impact rate of the NHPP process. For each impact event, random impact forces are generated according to a normal distribution, and the total impact damage is obtained by accumulating them. After multiple simulations, the total impact damage is statistically analyzed to obtain the probability density function and cumulative distribution function of the non-homogeneous composite Poisson distribution. The probability density and cumulative distribution of the non-homogeneous composite Poisson distribution at different time points are obtained.

7. A device for constructing a lithium battery degradation threshold-impact model based on the Wiener and Possion processes, used to implement the method as described in any one of claims 1-6, characterized in that, The device includes: The nonlinear degradation modeling module is configured to: establish an inherent degradation model of lithium battery based on the nonlinear Wiener process, wherein the nonlinear Wiener process uses drift parameters in the form of power functions to characterize the nonlinear characteristics of lithium battery degradation; The impact damage modeling module is configured to: construct a lithium battery impact damage model based on a non-homogeneous Poisson process, wherein the impact intensity of the non-homogeneous Poisson process is a function of time to characterize the non-homogeneous characteristics of the impact frequency changing with time. The impact strength parameter identification module is configured to optimize and estimate the strength function parameters of the non-homogeneous Poisson process using a nonlinear least squares fitting method. The dual-threshold failure mechanism definition module is configured to: set a soft failure threshold w and a hard failure threshold D, where the soft failure threshold w is the amount of degradation corresponding to the capacity decay to 80% of the initial capacity, and the hard failure threshold D is the critical value of cumulative impact damage. When the total degradation of the lithium battery is ≥ w or the cumulative impact damage is ≥ D, the lithium battery is determined to be in failure. The lithium battery degradation-threshold model construction module is configured to: couple the inherent degradation amount of the nonlinear Wiener process with the impact damage amount of the nonhomogeneous Poisson process to obtain the expression for the total degradation amount and construct the lithium battery degradation-threshold model. The parameter dynamic update module is configured to: based on Bayesian theory, use historical degradation data and real-time observation data of lithium batteries to update the drift parameters of the degradation model. The update process calculates the posterior distribution by using the likelihood function of the degradation data and the prior distribution of the parameters, thereby realizing the dynamic correction of the parameters. The lithium battery impact-threshold model construction module is configured to: construct a model relating impact damage to a threshold based on a non-homogeneous composite Poisson process, combined with the normal distribution characteristics of the number of impacts and single impact damage; The reliability analysis module is configured to: use Monte Carlo simulation to perform reliability analysis on the impact-threshold model, and obtain the probability density function, cumulative distribution function and reliability curve of total impact damage by simulating the number of impact events and impact intensity, and predict the remaining life and reliability of lithium battery.

8. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes the computer execution instructions stored in the memory to implement the lithium battery degradation threshold-impact model construction method based on the Wiener and Possion processes as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the lithium battery degradation threshold-impact model construction method based on the Wiener and Possion processes as described in any one of claims 1-6.

Images