Battery life prediction method and apparatus, electronic device, and readable storage medium

Abstract

The application provides a battery life prediction method, device, electronic equipment and readable storage medium. The battery life prediction method comprises the following steps: obtaining capacity attenuation data of a battery, wherein the capacity attenuation data comprises a cycle number of the battery and a discharge capacity of the battery at each cycle; performing model training according to the capacity attenuation data as training set data to obtain model parameters of a battery life prediction model, wherein the model parameters comprise a non-periodic change trend parameter, a periodic change trend parameter and a fluctuation change trend parameter of the battery capacity; establishing the battery life prediction model according to the model parameters; and inputting capacity attenuation data of a battery to be predicted into the battery life prediction model to obtain prediction data of the battery life. The embodiment of the application can improve the prediction accuracy of the battery life.

Description

TECHNICAL FIELD

[0001] Embodiments of the present application relate to the technical field of deep learning, and particularly relate to a battery life prediction method and device, electronic equipment and readable storage medium. BACKGROUND

[0002] With the development of new energy technology, electric vehicles are also more and more popular with users. Electric vehicles are usually powered by batteries, and the reliability and cycle life parameters of the batteries affect the performance of the vehicles, so they also attract people's attention. The existing battery life prediction methods are usually based on battery mechanism models or statistical models. However, the existing battery life prediction methods have relatively low accuracy in predicting battery life. SUMMARY

[0003] Embodiments of the present application provide a battery life prediction method, device, electronic equipment and readable storage medium to solve the problem of relatively low accuracy of existing battery life prediction methods in predicting battery life.

[0004] To solve the above problems, the present application is implemented as follows:

[0005] In a first aspect, embodiments of the present application provide a battery life prediction method, comprising the following steps:

[0006] Obtaining capacity attenuation data of the battery, the capacity attenuation data comprising the number of cycles of the battery and the discharge capacity of the battery at each cycle;

[0007] Model training is performed according to the capacity attenuation data as training set data to obtain model parameters of the battery life prediction model, wherein the model parameters comprise non-periodic change trend parameters, periodic change trend parameters and fluctuation change trend parameters of the battery capacity;

[0008] Establishing a battery life prediction model according to the model parameters;

[0009] Inputting the capacity attenuation data of the battery to be predicted into the battery life prediction model to obtain prediction data of the battery life.

[0010] In some embodiments, after obtaining the capacity attenuation data of the battery, the method further comprises:

[0011] Calculating the mean of each capacity attenuation data;

[0012] Calculating the variance of the capacity attenuation data according to the mean;

[0013] Standardizing the capacity attenuation data according to the mean and the variance, wherein the standardization is performed by the following formula:

[0014]

[0015] wherein, is a standardized result of x i , C is a set of capacity fade data, x i represents the i-th capacity fade data in C, μ C is a mean value of C, is a variance of C, and ε is a preset adjustment factor.

[0016] In some embodiments, the model training according to the capacity fade data as training set data comprises:

[0017] determining a particle speed update parameter of a particle swarm optimization algorithm;

[0018] determining a particle position update parameter of the particle swarm optimization algorithm according to the speed update amount;

[0019] taking the capacity fade data as training set data, taking a target parameter as a particle, and using the particle swarm optimization algorithm to iteratively update the particle speed update parameter and the particle position update parameter according to a preset loss function, to obtain an optimal value of the target parameter as a model parameter, wherein the target parameter comprises a non-periodic variation trend parameter, a periodic variation trend parameter, and a fluctuation variation trend parameter of the battery capacity.

[0020] In some embodiments, the determination of the particle speed update parameter of the particle swarm optimization algorithm comprises:

[0021] taking the following formula as a speed update formula of the target parameter:

[0022] v id (t+1) = ωv id (t) + c1r id1 (t)(p id (t) - X id (t)) + c2r id2 (t)(p gd (t) - X id (t));

[0023] wherein, v id (t+1) is a speed of the i-th target parameter in the d-th dimension in the t+1-th iteration, ω is a preset inertia weight, v id (t) is a speed of the i-th target parameter in the d-th dimension in the t-th iteration, X id (t) is a position of the i-th target parameter in the d-th dimension in the t-th iteration, c1 and c2 are preset acceleration constants, r id1 (t) and r id2 (t) are preset random numbers, and pid (t) is the value of the individual historical extreme value of the ith target parameter at time t in the dth dimension, p gd (t) is the value of the global extreme value of the ith target parameter at time t in the dth dimension;

[0024] The particle position update parameter of the target parameter is determined according to the speed update amount, including:

[0025] The following formula is taken as the position update formula of the target parameter:

[0026] X id (t+1)=X id (t)+v id (t+1);

[0027] Wherein, X id (t+1) is the position of the ith target parameter in the dth dimension in the t+1th iteration.

[0028] In some embodiments, the battery life prediction model is established according to the model parameter, including:

[0029] The following formula is taken as the battery life prediction model:

[0030] y(t)=g(t)+s(t)+h(t);

[0031] Wherein, g(t) represents the trend of change of the time series in the non-period, s(t) represents the trend of change of the time series in the period, and h(t) is the fluctuation change of the time series.

[0032] Wherein,

[0033] Wherein, C(t) is the carrying capacity, k is the growth rate, m is the offset, t is the cycle number, a(t) is an indicator function, and a(t) satisfies a(t)∈{0,1} S , δ is the growth rate change amount on the time stamp, and γ is the offset change amount on the time stamp;

[0034]

[0035] Wherein, P represents the period of the time series, N is the number of time periods, β=(a1, b1, …, a N , b N ) T , β~Normal(0, σ 2 ), Normal represents normal distribution, and σ 2 is a seasonal influence parameter;

[0036]

[0037] wherein, D i is a fluctuation time range parameter, L is a total number of fluctuation time range parameters, κ=(κ1,κ2,……,κ L ) T , and κ ~ Normal(0, v 2 ), v 2 is a fluctuation influence parameter.

[0038] In some embodiments, the inputting the capacity attenuation data into the battery life prediction model obtains predicted data of battery life, comprising:

[0039] inputting a first data set into the battery life prediction model to obtain output second data, wherein the first data set is the last group of discharge capacity data of the capacity attenuation data, and the data in the first data set is arranged in time sequence;

[0040] deleting the first data in the first data set, and taking the second data as the last data in the first data set to obtain an updated first data set;

[0041] iteratively updating the first data set until the capacity of the battery reaches a preset value;

[0042] taking the cycle number of the battery when the capacity of the battery reaches the preset value as the predicted data of the battery life.

[0043] In some embodiments, the deleting the first data in the first data set, and taking the second data as the last data in the first data set to obtain an updated first data set, comprises:

[0044] taking the second data as the last data in the first data set after the second data is de-normalized.

[0045] In a second aspect, the embodiments of the present application also provide a battery life prediction device, comprising:

[0046] a obtaining module, configured to obtain capacity attenuation data of a battery, the capacity attenuation data comprising cycle number of the battery and discharge capacity of the battery at each cycle;

[0047] a parameter training module, configured to perform model training according to the capacity attenuation data as training set data to obtain model parameters of a battery life prediction model, wherein the model parameters comprise non-periodic variation trend parameters and periodic variation trend parameters of the battery capacity;

[0048] a model establishing module, configured to establish the battery life prediction model according to the model parameters.

[0049] a prediction module, configured to input the last set of discharge capacity data of the capacity attenuation data of the battery to be predicted into the battery life prediction model to obtain prediction data of the battery life.

[0050] In a third aspect, an electronic device is provided, which includes a transceiver, a memory, a processor, and a program stored in the memory and capable of running on the processor; the processor is configured to read the program in the memory to implement the steps in the method in the first aspect.

[0051] In a fourth aspect, a readable storage medium is provided, which is used to store a program, and the program is executed by a processor to implement the steps in the method in the first aspect.

[0052] In the method provided by the embodiments of the present application, the capacity attenuation data of the battery is obtained, the capacity attenuation data includes the cycle number of the battery and the discharge capacity of the battery in each cycle; the model parameters of the battery life prediction model are obtained by performing model training according to the capacity attenuation data as training set data, wherein the model parameters include the non-periodic change trend parameter, the periodic change trend parameter and the fluctuation change trend parameter of the battery capacity; the battery life prediction model is established according to the model parameters; and the prediction data of the battery life is obtained by inputting the capacity attenuation data into the battery life prediction model. In this way, the embodiments of the present application can improve the prediction accuracy of the battery life by introducing the non-periodic change trend parameter, the periodic change trend parameter and the fluctuation change trend parameter of the battery capacity, and the technical solution of the embodiments of the present application is based on the capacity attenuation data of the battery, which can match different types of batteries and improve the versatility. BRIEF DESCRIPTION OF DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the description of the embodiments of the present application will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0054] Figure 1 is a flowchart of the battery life prediction method provided by the embodiments of the present application;

[0055] Figure 2 is another flowchart of the battery life prediction method provided by the embodiments of the present application;

[0056] Figure 3 is a structural schematic diagram of the battery life prediction device provided by the embodiments of the present application;

[0057] Figure 4 is a structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0059] The terms "first", "second", and the like in the embodiments of the present application are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or device including a series of steps or units does not necessarily have to include only those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to the process, method, product, or device. In addition, "and / or" is used in the present application to represent at least one of the connected objects, for example, A and / or B and / or C represents 7 cases including A alone, B alone, C alone, A and B both exist, B and C both exist, A and C both exist, and A, B and C all exist.

[0060] The embodiments of the present application provide a battery life prediction method.

[0061] As shown in FIG. 1, in one embodiment, the method comprises the following steps: Figure 1

[0062] Step 101: Obtain capacity attenuation data of the battery, wherein the capacity attenuation data comprises the cycle number of the battery and the discharge capacity of the battery in each cycle.

[0063] In the embodiment, first, the capacity attenuation data of the battery is obtained. Here, the battery analyzed can be a lithium ion battery or other types of batteries, which are not limited further herein.

[0064] It should be understood that, with the continuous charging and discharging cycles of the battery, the total amount of electricity that the battery can accommodate will decrease, which is called the attenuation of the battery capacity. In the embodiment, first, a certain amount of capacity attenuation data of the battery is obtained. The obtained data should be the data of the battery under normal working conditions, and exemplarily, the capacity attenuation data at normal temperature. The capacity attenuation data specifically comprises the cycle number of the battery and the discharge capacity of the battery in each cycle.

[0065] ​In this embodiment, the collected capacity decay data can be set as needed. For example, it can be set to a specified quantity or a specified time period. In one embodiment, the condition is set to a certain percentage of the battery's theoretical cycle life, for example, it can be different percentages such as 40%, 45%, or 50%.

[0066] For example, the lifespan of a battery is defined as reaching 80% of its initial capacity, at which point it is considered to have reached its service life. Theoretically, it would reach its service life after 1000 cycles, and capacity decay data corresponding to 400 cycles would be collected. Obviously, the above data is only used as an example for illustration. In implementation, it can be set as needed, and no further limitations or descriptions are made in this embodiment.

[0067] In some embodiments, after acquiring the capacity decay data, the capacity decay data is also normalized.

[0068] Specifically, after step 101, the method further includes:

[0069] Calculate the mean of each of the aforementioned capacity decay data;

[0070] Calculate the variance of the capacity decay data based on the mean;

[0071] The capacity decay data is standardized based on the mean and the variance.

[0072] In this embodiment, the set of capacity decay data is denoted as C = {x1, x2, x3, ..., x...} m}, where m is the number of data points. First, the mean μ of C is calculated using formula (1). C .

[0073]

[0074] Next, the variance of C is calculated using formula (2).

[0075]

[0076] Furthermore, C is standardized according to formula (3).

[0077]

[0078] in, For data x in C i The standardized result, x i μ represents the i-th capacity decay data in C. C Let C be the mean. C is a variance of C, and ε is a preset adjustment factor. It should be understood that the variance of C can be 0, and by setting the preset adjustment factor, it can be avoided that the denominator is equal to 0 when the variance is 0.

[0079] Step 102: model training is performed according to the capacity attenuation data as training set data to obtain model parameters of the battery life prediction model, wherein the model parameters include a non-periodic change trend parameter, a periodic change trend parameter and a fluctuation change trend parameter of the battery capacity.

[0080] Next, the model parameters of the battery life prediction model are obtained through model training. In some embodiments, the model parameters of the battery life prediction model are obtained through model training by a particle swarm optimization algorithm.

[0081] In one embodiment, the step 102 specifically includes:

[0082] determining a particle velocity update parameter of the particle swarm optimization algorithm;

[0083] determining a particle position update parameter of the particle swarm optimization algorithm according to the velocity update amount;

[0084] taking the capacity attenuation data as training set data, taking a target parameter as a particle, and using the particle swarm optimization algorithm to iteratively update the particle velocity update parameter and the particle position update parameter according to a preset loss function to obtain an optimal value of the target parameter as a model parameter, wherein the target parameter includes a non-periodic change trend parameter, a periodic change trend parameter and a fluctuation change trend parameter of the battery capacity.

[0085] In some embodiments, the determination of the particle velocity update parameter of the particle swarm optimization algorithm includes:

[0086] the following formula (4) is taken as a velocity update formula of the target parameter:

[0087] v id (t+1) = ωv id (t) + c1r id1 (t)(p id (t) - X id (t)) + c2r id2 (t)(p gd (t) - X id (t)) …… (4);

[0088] In formula (4), v id (t+1) is a velocity of the i-th target parameter in the d-th dimension in the t+1-th iteration, ω is a preset inertia weight, which can be set to different values as needed, and exemplary values can be 0.5, 0.6, etc., v id(t) the velocity of the i-th target parameter in the d-th dimension in the t-th iteration, X id (t) the position of the i-th target parameter in the d-th dimension in the t-th iteration, c1 and c2 are preset acceleration constants, r id1 (t) and r id2 (t) is a preset random number, which can be in the range of (0, 1), p id (t) is the value of the individual historical extreme value of the i-th target parameter at time t in the d-th dimension, p gd (t) is the value of the global extreme value of the i-th target parameter at time t in the d-th dimension.

[0089] In some embodiments, the particle position update parameter of the target parameter is determined according to the velocity update amount, including:

[0090] The following formula (5) is used as the position update formula of the target parameter:

[0091] X id (t+1) = X id (t) + v id (t+1) …… (5).

[0092] In formula (5), X id (t+1) is the position of the i-th target parameter in the d-th dimension in the t+1-th iteration, the time unit is 1, so that v id (t+1) actually corresponds to the position change amount of the d-th dimension in the t+1-th iteration.

[0093] In some embodiments, the mean square error loss MSE is used as the loss function for model training, and specifically, the loss function is shown in formula (6).

[0094]

[0095] In formula (6), n is the number of groups corresponding to the battery capacity and cycle number data in the training set, y i is the true value of the battery discharge capacity in the training set data, is the predicted value of the battery discharge capacity.

[0096] In implementation, first, the value range of the target parameter is set, then a part of the data in the training set, for example, 1 / 4 or 1 / 5 or different proportions of the training set data, is used as the loss function of formula (6), and the optimal value of the target parameter is obtained in the set range by continuously updating the above formula (4) and (5), here, the optimal value specifically includes the historical optimal position of a single particle and the global optimal solution of all particles, and finally, the optimal value is used as the model parameter of the battery life prediction model.

[0097] The optimal solution can be obtained when the loss function converges, or when the number of iterations reaches a certain number, for example, after 30 iterations, the obtained value is taken as the optimal value.

[0098] The particle swarm optimization algorithm is used to optimize the parameters in the range, to obtain the historical optimal position of a single particle and the optimal position of the group, thereby avoiding the blindness and randomness caused by artificial setting of model initialization parameters.

[0099] In some embodiments, the battery life prediction model is established according to the model parameters, comprising:

[0100] The following formula is taken as the battery life prediction model:

[0101] y(t)=g(t)+s(t)+h(t)……(7);

[0102] Wherein, g(t) represents the trend of the time series on the non-periodic, s(t) represents the trend of the time series on the periodic, and h(t) is the fluctuation of the time series;

[0103] Wherein,

[0104] In formula (8), C(t) is the carrying capacity, k is the growth rate, m is the offset, t is the cycle number, a(t) is an indicator function, a(t)=(a1(t), a2(t), …, a S (t)) T , a(t) satisfies a(t)∈{0,1} S ; δ is the growth rate change amount on the timestamp, δ=(δ1, δ2, …, δ S ) T ; γ is the offset change amount on the timestamp, γ=(γ1, γ2, …, γ S ) T , γ t =-s t δ t .

[0105] The above a(t) is an indicator function, which is defined as: when t satisfies the value range S, the value of a(t) is 1, when t does not satisfy the value range S, the value of a(t) is 0, S is a time range set according to needs, specifically, the timestamp of the position of the growth rate change point is s j , when 1≤j≤S,

[0106]

[0107] Wherein, P represents a period of time series, and N is a number of time periods.

[0108] β = (a1, b1, …, a N , b N ) T , β ~ Normal (0, σ 2 ), and σ 2 is a seasonal influence parameter.

[0109]

[0110] In formula (10), D i is a fluctuation time range parameter, L is a total number of fluctuation time range parameters, κ = (κ1, κ2, …, κ L ) T , and κ ~ Normal (0, v 2 ), v 2 is a fluctuation influence parameter.

[0111] For example, the fluctuation influence can refer to an influence of a non-periodic and irregular factor such as a holiday on the use of the battery.

[0112] In some embodiments, an error term may be further added in formula (7), so that formula (7) can be further expressed as formula (11).

[0113]

[0114] The error indicates that there can be an error, which is generally a parameter satisfying a normal distribution.

[0115] In implementation, first, the parameters are initialized to obtain an initialized optimal position, further, a value range of each model parameter is set, each target parameter is taken as a particle, and an optimal value is determined by a particle swarm optimization algorithm, and the optimal value is taken as a model parameter of the battery life prediction model.

[0116] Step 103: establishing the battery life prediction model according to the model parameters.

[0117] The model parameters obtained in step 102 are substituted into the model, so that the battery life prediction model for predicting the battery life is obtained.

[0118] The established battery life prediction model can be a Prophet model (a kind of model), the Prophet model can automatically add abnormal detection logic in the model, reduce the influence of noise on the prediction result, can avoid the influence of the periodic capacity jump existing in the capacity fading process of some batteries on the prediction result, and improve the prediction accuracy of the cycle life of the battery.

[0119] Meanwhile, the technical scheme of the embodiment does not need to understand the physical and chemical mechanism of battery degradation, and is good in adaptability for cycle life prediction of different types of batteries from the data point of view.

[0120] Step 104: inputting the capacity attenuation data of the battery to be predicted into the battery life prediction model to obtain prediction data of the battery life.

[0121] The capacity attenuation data is input into the battery life prediction model, and the prediction data of the battery life can be obtained.

[0122] In some embodiments, the step 104 specifically includes:

[0123] Inputting the first data set into the battery life prediction model to obtain output second data, wherein the first data set is the last group of discharge capacity data of the capacity attenuation data, and the data in the first data set is arranged in time sequence;

[0124] Deleting the first data in the first data set, and taking the second data as the last data in the first data set to obtain an updated first data set;

[0125] Iteratively updating the first data set until the capacity of the battery reaches a preset value;

[0126] Taking the cycle number of the battery when the capacity of the battery reaches the preset value as the prediction data of the battery life.

[0127] In the embodiment, the last group of discharge capacity data of the capacity attenuation data is denoted as C * , specifically,

[0128]

[0129] Here, is the first data set, each data in the first data set is arranged in time sequence, H is the length of the batch data, and the first data set is input into the battery life prediction model to obtain the output result According to the standardization process, the first data set is inversely standardized to obtain x m+1 , x m+1 ​That is, the second data described above. That is, the step specifically includes the second data after the standardization as the last data in the first data group.

[0130] The x m+1 As The last data, while deleting the first data in the first data group The updated first data group As

[0131] In this way, the updating process for the first data group is performed in a loop until the capacity of the battery reaches the preset value.

[0132] For example, the preset value is set to 80%, and through the above process, when the battery capacity reaches 80%, the number of cycles elapsed is recorded.

[0133] In implementation, the preset value can be set to the capacity corresponding to the life of the battery, so that the life of the battery can be predicted.

[0134] As Figure 2 shown, the technical solution of the embodiment can be summarized as follows: first, the particle position and velocity are initialized to obtain the optimal position in the initialized state. Next, model training is performed to calculate the historical optimal position of a single particle and the global optimal solution under the constraint of a preset loss function, and the position and velocity of the particle are constantly updated according to the calculation result. The process of model training is iteratively performed until the model training condition is met, and a battery life prediction model meeting the requirements is obtained. Based on the battery life prediction model, the prediction result of the battery life can be obtained.

[0135] The embodiment of the application also provides a battery life prediction device.

[0136] As Figure 3 shown, in one embodiment, the battery life prediction device 300 includes:

[0137] The acquisition module 301 is configured to acquire capacity attenuation data of the battery, the capacity attenuation data including the number of cycles of the battery and the discharge capacity of the battery at each cycle;

[0138] The parameter training module 302 is configured to perform model training according to the capacity attenuation data as training set data to obtain model parameters of the battery life prediction model, wherein the model parameters include a non-periodic change trend parameter, a periodic change trend parameter and a fluctuation change trend parameter of the battery capacity.

[0139] The model establishment module 303 is configured to establish the battery life prediction model according to the model parameters.

[0140] The prediction module 304 is configured to input the capacity attenuation data of the battery to be predicted into the battery life prediction model to obtain prediction data of the battery life.

[0141] In some embodiments, the method further comprises:

[0142] The mean calculation module is configured to calculate a mean value of each of the capacity attenuation data.

[0143] The variance calculation module is configured to calculate a variance of the capacity attenuation data according to the mean value.

[0144] The standardization processing module is configured to perform standardization processing on the capacity attenuation data according to the mean value and the variance, wherein the standardization processing is performed according to the following formula:

[0145]

[0146] wherein, is the standardization result of x i , C is a set of the capacity attenuation data, x i represents the i-th capacity attenuation data in C, μ C is the mean value of C, is the variance of C, and ε is a preset adjustment factor.

[0147] In some embodiments, the parameter training module 302 comprises:

[0148] The speed update parameter determination submodule is configured to determine a particle speed update parameter of a particle swarm optimization algorithm.

[0149] The position update parameter determination submodule is configured to determine a particle position update parameter of the particle swarm optimization algorithm according to the speed update amount.

[0150] The training submodule is configured to use the capacity attenuation data as training set data, use a target parameter as a particle, and use a particle swarm optimization algorithm to iteratively update the particle speed update parameter and the particle position update parameter according to a preset loss function, so as to obtain an optimal value of the target parameter as a model parameter, wherein the target parameter comprises a non-periodic change trend parameter, a periodic change trend parameter, and a fluctuation change trend parameter of the battery capacity.

[0151] In some embodiments, the speed update parameter determination submodule is specifically configured to use the following formula as a speed update formula of the target parameter:

[0152] v id (t+1) = ωv id (t) + c1r id1 (t)(p id (t) - x id (t)) + c2rid2 (t)(p gd (t)-X id (t));

[0153] wherein, v id (t+1) is the speed of the i th target parameter in the d th dimension in the t+1 th iteration, ω is a preset inertia weight, v id (t) is the speed of the i th target parameter in the d th dimension in the t th iteration, X id (t) is the position of the i th target parameter in the d th dimension in the t th iteration, c1 and c2 are preset acceleration constants, r id1 (t) and r id2 (t) is a preset random number, p id (t) is the value of the individual historical extreme value of the i th target parameter at the t th moment in the d th dimension, p gd (t) is the value of the global extreme value of the i th target parameter at the t th moment in the d th dimension.

[0154] The position updating parameter determination sub-module is specifically configured to use the following formula as the position updating formula of the target parameter:

[0155] X id (t+1) = X id (t) + v id (t+1);

[0156] wherein, X id (t+1) is the position of the i th target parameter in the d th dimension in the t+1 th iteration.

[0157] In some embodiments, the model establishing module 303 is specifically configured to use the following formula as the battery life prediction model:

[0158] y(t) = g(t) + s(t) + h(t);

[0159] wherein, g(t) represents the trend of the time series in the non-periodic direction, s(t) represents the trend of the time series in the periodic direction, and h(t) is the fluctuation of the time series.

[0160] wherein,

[0161] wherein, C(t) is the carrying capacity, k is the growth rate, m is the offset, t is the cycle number, a(t) is an indicator function, and a(t) satisfies a(t) ∈ {0, 1} S , δ is the growth rate change amount on the time stamp, and γ is the offset change amount on the time stamp.

[0162]

[0163] Where P represents the period of the time series, N is the number of time periods, and β = (a1, b1, ..., a N b N ) T ,β~Nomal(0,σ 2 ), σ 2 Parameters related to seasonal influence;

[0164]

[0165] in, D i Here, L represents the time range of fluctuation parameters, and L represents the total number of time range of fluctuation parameters. κ=(κ1,κ2,……,κ L ) T , and κ~Nomal(0, v 2 ), v 2 The parameter is affected by fluctuations.

[0166] In some embodiments, the prediction module 304 includes:

[0167] The input submodule is used to input the first data group into the battery life prediction model to obtain the output second data, wherein the first data group is the last set of discharge capacity data of the capacity decay data, and the data in the first data group are arranged in chronological order.

[0168] The update submodule is used to delete the first data in the first data group and use the second data as the last data in the first data group to obtain the updated first data group;

[0169] The iterative submodule is used to iteratively update the first data group until the battery capacity reaches a preset value;

[0170] The recording submodule is used to use the number of battery cycles when the battery capacity reaches a preset value as predicted data for battery life.

[0171] In some embodiments, the update submodule is specifically used to: denormalize the second data and use it as the last data in the first data group.

[0172] The battery life prediction device 300 of this embodiment can implement all the steps of the above-described battery life prediction method embodiment and achieve essentially the same technical effect, which will not be repeated here.

[0173] This invention also provides an electronic device. Please refer to [link to relevant documentation]. Figure 4The electronic device can include a processor 401, a memory 402, and a program 4021 stored on the memory 402 and executable on the processor 401.

[0174] In the case of the electronic device being a terminal, the program 4021, when executed by the processor 401, can implement Figure 1 Any step in the corresponding method embodiments and the same beneficial effects can be achieved, and details are not repeated here.

[0175] In the case of the electronic device being a network side device, the program 4021, when executed by the processor 401, can implement Figure 4 Any step in the corresponding method embodiments and the same beneficial effects can be achieved, and details are not repeated here.

[0176] A person of ordinary skill in the art can understand that all or part of the steps of the above-mentioned embodiment methods can be completed by program instructions related to hardware, and the program can be stored in a readable medium.

[0177] The embodiment of the present application also provides a readable storage medium, and the readable storage medium stores a computer program, and the computer program is executable by a processor to implement the above Figure 1 Any step in the corresponding method embodiments and the same beneficial effects can be achieved, and details are not repeated here.

[0178] The storage medium, such as a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0179] The above is the preferred embodiment of the present application, it should be noted that for those of ordinary skill in the art, without departing from the principles of the present application, can make a number of improvements and refinements, these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A method for predicting battery life, characterized in that, Includes the following steps: Obtain battery capacity decay data, which includes the number of battery cycles and the battery discharge capacity during each cycle. The battery life prediction model is trained using the capacity decay data as the training set data to obtain the model parameters, wherein the model parameters include non-periodic change trend parameters, periodic change trend parameters, and fluctuation change trend parameters of battery capacity. A battery life prediction model is established based on the model parameters. Input the capacity degradation data of the battery to be predicted into the battery life prediction model to obtain the predicted battery life data. The step of establishing a battery life prediction model based on the model parameters includes: The following formula is used as a battery life prediction model: y(t)=g(t)+s(t)+h(t); Where g(t) represents the non-periodic trend of the time series, s(t) represents the periodic trend of the time series, and h(t) represents the fluctuation of the time series. in, ; Where c(t) is the carrying capacity, k is the growth rate, m is the offset, t is the number of cycles, and a(t) is the indicator function, satisfying a(t)∈{0,1}. S , This represents the change in growth rate at the timestamp. This represents the change in offset on the timestamp; ； Where P represents the period of the time series, and N is the number of time periods. , , Parameters related to seasonal influence; ； in, , Here, L represents the time range of fluctuation parameters, and L represents the total number of time range of fluctuation parameters. , ; ,and , The parameter is affected by fluctuations.

2. The method according to claim 1, characterized in that, After acquiring the battery capacity degradation data, the method further includes: Calculate the mean of each of the aforementioned capacity decay data; Calculate the variance of the capacity decay data based on the mean; The capacity decay data is standardized based on the mean and variance, wherein the standardization is performed using the following formula: ； in, for The standardized results, where C is the set of capacity decay data. This represents the i-th capacity decay data in C. Let C be the mean. Let C be the variance. This is the preset adjustment factor.

3. The method according to claim 1, characterized in that, The step of training the model using the capacity decay data as training set data includes: Determine the particle velocity update parameters for the particle swarm optimization algorithm; The particle position update parameters of the particle swarm optimization algorithm are determined based on the velocity update amount. Using the capacity decay data as training set data and the target parameter as particles, the particle swarm optimization algorithm iteratively updates the particle velocity update parameter and the particle position update parameter according to a preset loss function to obtain the optimal value of the target parameter as the model parameter. The target parameter includes the non-periodic change trend parameter, the periodic change trend parameter, and the fluctuation change trend parameter of the battery capacity.

4. The method according to claim 3, characterized in that, The determination of particle velocity update parameters for the particle swarm optimization algorithm includes: The following formula is used as the velocity update formula for the target parameter: v id (t+1)=ωv id (t)+c1r id1 (t)(p id (t)-X id (t))+c2r id2 (t)(p gd (t)-X id (t)); Among them, v id (t+1) represents the velocity of the i-th objective parameter in the d-th dimension during the (t+1)-th iteration, ω is the preset inertia weight, and v id (t) represents the velocity of the i-th objective parameter in the d-th dimension during the t-th iteration, X id (t) represents the position of the i-th objective parameter in the d-th dimension during the t-th iteration, c1 and c2 are preset acceleration constants, and r id1 (t) and r id2 (t) is a preset random number, p id (t) represents the value of the individual historical extreme value of the i-th objective parameter at time t in the d-th dimension, p gd (t) represents the global extremum of the i-th objective parameter at time t in the d-th dimension; The step of determining the particle position update parameters of the target parameters based on the velocity update amount includes: Use the following formula as the position update formula for the target parameter: X id (t+1)=X id (t)+v id (t+1); Among them, X id (t+1) represents the position of the i target parameters in the d-th dimension during the (t+1)-th iteration.

5. The method according to any one of claims 1 to 4, characterized in that, The step of inputting the capacity degradation data into the battery life prediction model to obtain predicted battery life data includes: The first data set is input into the battery life prediction model to obtain the second output data, wherein the first data set is the last set of discharge capacity data of the capacity decay data, and the data in the first data set are arranged in chronological order. Delete the first data in the first data group and use the second data as the last data in the first data group to obtain the updated first data group; Iteratively update the first data group until the battery capacity reaches a preset value; The number of battery cycles when the battery capacity reaches a preset value is used as the predicted data for battery life.

6. The method according to claim 5, characterized in that, The step of deleting the first data in the first data group and using the second data as the last data in the first data group to obtain the updated first data group includes: The second data is denormalized and then used as the last data in the first data group.

7. A battery life prediction device, characterized in that, include: The acquisition module is used to acquire battery capacity decay data, which includes the number of battery cycles and the battery discharge capacity during each cycle. The parameter training module is used to train the model based on the capacity decay data as training set data to obtain the model parameters of the battery life prediction model, wherein the model parameters include non-periodic change trend parameters and periodic change trend parameters of battery capacity. The model building module is used to build a battery life prediction model based on the model parameters. The prediction module is used to input the last set of discharge capacity data of the battery to be predicted into the battery life prediction model to obtain the predicted battery life data. The model building module is used for: The following formula is used as a battery life prediction model: y(t)=g(t)+s(t)+h(t); Where g(t) represents the non-periodic trend of the time series, s(t) represents the periodic trend of the time series, and h(t) represents the fluctuation of the time series. in, ; Where C(t) is the carrying capacity, k is the growth rate, m is the offset, t is the number of cycles, and a(t) is the indicator function, satisfying a(t)∈{0,1}. S δ represents the change in growth rate at the timestamp, and γ represents the change in offset at the timestamp; ； Where P represents the period of the time series, and N is the number of time periods. , , Parameters related to seasonal influence; ； in, , Here, L represents the time range of fluctuation parameters, and L represents the total number of time range of fluctuation parameters. , ; ,and , The parameter is affected by fluctuations.

8. An electronic device, comprising: A memory, a processor, and a program stored in the memory and executable on the processor; characterized in that the processor is configured to read the program in the memory to implement the steps in the battery life prediction method as described in any one of claims 1 to 6.

9. A readable storage medium for storing a program, characterized in that, When the program is executed by the processor, it implements the steps in the battery life prediction method as described in any one of claims 1 to 6.

Images