CN107544033B - Digital-analog fusion prediction method for remaining service life of lithium ion battery - Google Patents
Digital-analog fusion prediction method for remaining service life of lithium ion battery Download PDFInfo
- Publication number
- CN107544033B CN107544033B CN201710791239.9A CN201710791239A CN107544033B CN 107544033 B CN107544033 B CN 107544033B CN 201710791239 A CN201710791239 A CN 201710791239A CN 107544033 B CN107544033 B CN 107544033B
- Authority
- CN
- China
- Prior art keywords
- lithium battery
- service life
- temperature
- residual service
- predicted
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Charge And Discharge Circuits For Batteries Or The Like (AREA)
- Tests Of Electric Status Of Batteries (AREA)
Abstract
The invention discloses a digital-analog fusion prediction method for the remaining service life of a lithium battery, which comprises the steps of carrying out model training by using low-temperature capacity degradation data of the lithium battery in a lithium battery historical database, the low-temperature remaining service life of the lithium battery and the low-temperature remaining service life of the lithium battery to obtain a prediction model for the low-temperature remaining service life of the lithium battery; inputting part of low-temperature capacity degradation data of the lithium battery to be tested into the lithium battery low-temperature remaining service life prediction model, and predicting the low-temperature remaining service life of the lithium battery to be tested to obtain a predicted value of the low-temperature remaining service life of the lithium battery; determining a prediction equation of the high-temperature residual service life of the lithium battery to be tested by using the predicted value of the low-temperature residual service life of the lithium battery; and calculating the high-temperature predicted residual service life of the lithium battery by using the determined high-temperature residual service life prediction equation of the lithium battery to be detected.
Description
Technical Field
The invention relates to a method for predicting the remaining service life of a lithium battery based on the fusion of data driving and model driving, which combines the advantages of the data driving method and the model driving method to complete the prediction of the remaining service life of the lithium battery at different temperatures. The method is suitable for the fields of lithium ion service life prediction and the like.
Background
A lithium ion battery (commonly called as a lithium battery), which is a secondary battery (a rechargeable battery) and mainly works by moving lithium ions between a positive electrode and a negative electrode, has the advantages of high energy density, high open-circuit voltage, large output power, no memory effect, low self-discharge, wide working temperature range, high charging and discharging speed and the like, and is widely applied to the electronic product industries of notebook computers, mobile phones and the like; the vehicle industries such as electric bicycles, electric automobiles, and the like; aerospace and other national defense and military industries; and in the medical industry, such as hearing aids, cardiac pacemakers and other non-life supporting devices. As a core component, the loss due to ignition of the lithium ion battery is enormous. If can just change its battery before the lithium cell is ageing, should play certain positive role to the prevention of lithium cell accident, but blindly change the lithium cell in advance, must lead to the fact huge waste. Therefore, the significance of accurately predicting the residual service life of the lithium battery is great.
At present, the prediction methods for the remaining service life of the lithium ion battery are mainly classified into three types based on model driving, data driving and fusion type methods. The method based on model driving can model the failure mechanism of the lithium battery, can well reflect the physical and electrochemical characteristics of the battery, but has the problems of difficult modeling or difficult model parameter identification; the data-driven method is simple and practical, can be realized only by the support of test data and state monitoring data, is easily influenced by data uncertainty and incompleteness, and has poor robustness and adaptability. The fusion-based method integrates multiple methods, makes up for the defect of a single model, gives full play to the advantages of different models, and can obtain better performance. However, most of the existing fusion strategies are only fused in a decision layer, and only simple result integration is performed, so that the advantages of a fusion type method cannot be fully exerted. According to the invention, through integrating the advantages of the data driving method and the model driving method, the prediction of the residual service life of the lithium ion battery at different temperatures is completed, so that the service life management of the lithium ion battery in actual use is assisted.
Disclosure of Invention
The invention aims to provide a digital-analog fusion prediction method for the residual service life of a lithium ion battery, which combines the advantages of a data driving method and a model driving method based on a data driving and model driving fusion method to complete the service life prediction of the lithium ion battery at 25 ℃, 45 ℃ and 60 ℃.
According to an embodiment of the invention, the digital-analog fusion prediction method for the remaining service life of the lithium battery provided by the invention comprises the following steps:
performing model training by using the low-temperature capacity degradation data of the lithium battery, the low-temperature remaining service life of the lithium battery and the low-temperature remaining service life of the lithium battery in the historical database of the lithium battery to obtain a low-temperature remaining service life prediction model of the lithium battery;
inputting part of low-temperature capacity degradation data of the lithium battery to be tested into the lithium battery low-temperature remaining service life prediction model, and predicting the low-temperature remaining service life of the lithium battery to be tested to obtain a predicted value of the low-temperature remaining service life of the lithium battery;
determining a prediction equation of the high-temperature residual service life of the lithium battery to be tested by using the predicted value of the low-temperature residual service life of the lithium battery;
and calculating the high-temperature predicted residual service life of the lithium battery by using the determined high-temperature residual service life prediction equation of the lithium battery to be detected.
According to another embodiment of the invention, the digital-analog fusion prediction method for the remaining service life of the lithium battery provided by the invention comprises the following steps:
performing model training and verification processing by using the low-temperature capacity degradation data of the lithium battery, the low-temperature remaining service life of the lithium battery and the low-temperature remaining service life of the lithium battery in the historical database of the lithium battery to obtain a low-temperature remaining service life prediction model of the lithium battery and a high-temperature remaining service life correction prediction equation of the lithium battery;
inputting part of low-temperature capacity degradation data of the lithium battery to be tested into the lithium battery low-temperature remaining service life prediction model, and predicting the low-temperature remaining service life of the lithium battery to be tested to obtain a predicted value of the low-temperature remaining service life of the lithium battery;
determining a prediction equation of the high-temperature residual service life of the lithium battery to be tested by using the predicted value of the low-temperature residual service life of the lithium battery;
and calculating the high-temperature predicted residual service life of the lithium battery by using the determined high-temperature residual service life prediction equation of the lithium battery to be detected, and correcting the high-temperature predicted residual service life of the lithium battery by using the high-temperature residual service life correction prediction equation of the lithium battery to obtain the actual high-temperature residual service life prediction value of the lithium battery.
Preferably, the obtaining of the prediction model of the low-temperature remaining service life of the lithium battery and the correction prediction equation of the high-temperature remaining service life of the lithium battery include:
carrying out residual service life prediction training on a linear prediction model with good linear prediction effect by using low-temperature capacity degradation data of the lithium battery in a lithium battery historical database to obtain a low-temperature residual service life prediction model of the lithium battery;
determining a prediction equation of the high-temperature residual service life of the lithium battery by utilizing the low-temperature residual service life of the lithium battery predicted by the low-temperature residual service life prediction model of the lithium battery and the Allen-nies model;
the high-temperature residual service life correction prediction equation of the lithium battery is obtained by verifying and processing the actual high-temperature residual service life value of the lithium battery in the lithium battery historical database, the high-temperature predicted residual service life value of the lithium battery predicted by the high-temperature residual service life prediction equation of the lithium battery and the low-temperature predicted residual service life value of the lithium battery predicted by the low-temperature residual service life prediction model of the lithium battery.
Preferably, the low-temperature actual remaining service life of the lithium battery comprises an actual remaining service life of the lithium battery at 25 ℃ and an actual remaining service life of the lithium battery at 45 ℃; the low-temperature predicted residual service life of the lithium battery comprises the predicted residual service life of the lithium battery at 25 ℃ and the predicted residual service life of the lithium battery at 45 ℃; the actual residual service life of the lithium battery at high temperature is the actual residual service life of the lithium battery at 60 ℃; the predicted residual service life of the lithium battery at the high temperature is the predicted residual service life of the lithium battery at 60 ℃.
Preferably, the training of the predicted value of the remaining service life of the linear prediction model by using the low-temperature capacity degradation data of the lithium battery in the lithium battery historical database comprises the following steps:
and training an autoregressive integral sliding average model ARIMA by using the 25 ℃ capacity degradation data and the 45 ℃ capacity degradation data of the lithium battery and the 25 ℃ residual service life and the 45 ℃ residual service life of the lithium battery in a lithium battery historical database to obtain a trained low-temperature residual service life prediction model for predicting the 25 ℃ residual service life and the 45 ℃ residual service life of the lithium battery.
Preferably, the determining the prediction equation of the high-temperature residual service life of the lithium battery by using the low-temperature residual service life of the lithium battery and the arrhenius model comprises:
determining coefficients in an Arrhenius model by using the residual service life of the lithium battery at 25 ℃ and the residual service life of the lithium battery at 45 ℃;
and using the determined coefficients of the Arrhenius model as a prediction equation of the high-temperature residual service life of the lithium battery.
Preferably, determining the coefficients in the arrhenius model using the remaining service life of the lithium battery at 25 ℃ and the remaining service life of the lithium battery at 45 ℃ comprises:
simplifying an Arrhenius model into an equation ln L ═ a + b/T, wherein L is a residual life value, T is temperature, a is a first coefficient, and b is a second coefficient;
the method comprises the steps of solving a first coefficient a and a second coefficient b in lnL ═ a + b/T by using a residual service life value of the lithium battery at 25 ℃ and a residual service life value of the lithium battery at 45 ℃, so as to obtain an Allen model simplified equation for predicting the residual service life of the lithium battery at 60 ℃, and further calculate a predicted value of the residual service life of the lithium battery to be tested at 60 ℃ by using an equation ln L ═ a + b/T.
Preferably, the equation for obtaining the corrected and predicted high-temperature remaining service life of the lithium battery comprises:
obtaining a ratio curve corresponding to a series of ratios of a series of actual service life values of the high temperature lithium battery to a series of high temperature predicted remaining service life values of the lithium battery by utilizing a series of actual service life values of the high temperature lithium battery in a lithium battery historical database and a series of high temperature predicted remaining service life values of the lithium battery calculated by a lithium battery high temperature remaining service life prediction equation;
obtaining a difference curve corresponding to a series of differences between the predicted residual service life values of the series of lithium batteries at 25 ℃ and the predicted residual service life values of the series of lithium batteries at 45 ℃ by utilizing a series of differences between the predicted residual service life values of the series of lithium batteries at 25 ℃ and the predicted residual service life values of the series of lithium batteries at 45 ℃ predicted by the low-temperature residual service life prediction model of the lithium batteries;
and fitting a lithium battery high-temperature residual service life correction prediction equation which enables the difference value of the predicted residual service life value of the lithium battery at 25 ℃ and the predicted residual service life value of the lithium battery at 45 ℃ to be equal to the ratio of the lithium battery high-temperature actual service life value and the lithium battery high-temperature predicted residual service life value according to the ratio curve and the difference value curve.
Preferably, the correcting the high-temperature remaining service life of the lithium battery by using the correction prediction equation comprises:
and after the high-temperature predicted residual service life of the lithium battery is calculated by using the determined prediction equation of the high-temperature residual service life of the lithium battery to be detected, multiplying the calculation result of the correction prediction equation of the high-temperature residual service life of the lithium battery by the calculated value of the high-temperature predicted residual service life of the lithium battery.
Preferably, the high-temperature remaining service life correction prediction equation of the lithium battery is as follows: ratio ═ f (difference); wherein, the ratio is the ratio of the actual high-temperature service life value of the lithium battery to the predicted residual high-temperature service life value of the lithium battery; (difference) is a fitting function with the variable being the difference between the predicted remaining life value of the lithium battery at 25 ℃ and the predicted remaining life value of the lithium battery at 45 ℃.
Drawings
The foregoing features of the invention will be more readily understood from the following detailed description taken in conjunction with the accompanying drawings, in which:
FIG. 1 shows a flow of a method for predicting the remaining service life of a lithium battery based on digital-analog fusion according to the present invention;
FIG. 2 shows a capacity degradation curve at 25 deg.C, 45 deg.C, 60 deg.C for a lithium battery of the present invention;
FIG. 3 shows the ARIMA life prediction effect of a group A of certain batteries in the lithium battery historical database at 25 ℃.
FIG. 4 shows a fitting relationship between the difference between the predicted lifetime at 25 ℃ and the predicted lifetime at 45 ℃ in the historical data of the present invention and the ratio between the actual lifetime at 60 ℃ and the extrapolated lifetime of the Allen model at 60 ℃;
the numbers, symbols and codes in the figures are explained as follows:
ARIMA: autoregressive Integrated moving average model (see FIG. 1)
difference between predicted service life at 25 ℃ and predicted service life at 45 ℃; ratio: ratio of actual lifetime at 60 ℃ to extrapolated lifetime of the Allen model at 60 DEG C
Detailed Description
In a first embodiment of the present invention, a method for predicting remaining service life of a lithium battery by digital-analog fusion includes:
performing model training by using the low-temperature capacity degradation data of the lithium battery, the low-temperature remaining service life of the lithium battery and the low-temperature remaining service life of the lithium battery in the historical database of the lithium battery to obtain a low-temperature remaining service life prediction model of the lithium battery;
inputting part of low-temperature capacity degradation data of the lithium battery to be tested into the lithium battery low-temperature remaining service life prediction model, and predicting the low-temperature remaining service life of the lithium battery to be tested to obtain a predicted value of the low-temperature remaining service life of the lithium battery;
determining a prediction equation of the high-temperature residual service life of the lithium battery to be tested by using the predicted value of the low-temperature residual service life of the lithium battery;
and calculating the high-temperature predicted residual service life of the lithium battery by using the determined high-temperature residual service life prediction equation of the lithium battery to be detected.
In a second embodiment of the present invention, the digital-analog fusion prediction method for remaining service life of a lithium battery provided by the present invention includes:
performing model training and verification processing by using the low-temperature capacity degradation data of the lithium battery, the low-temperature remaining service life of the lithium battery and the low-temperature remaining service life of the lithium battery in the historical database of the lithium battery to obtain a low-temperature remaining service life prediction model of the lithium battery and a high-temperature remaining service life correction prediction equation of the lithium battery;
inputting part of low-temperature capacity degradation data of the lithium battery to be tested into the lithium battery low-temperature remaining service life prediction model, and predicting the low-temperature remaining service life of the lithium battery to be tested to obtain a predicted value of the low-temperature remaining service life of the lithium battery;
determining a prediction equation of the high-temperature residual service life of the lithium battery to be tested by using the predicted value of the low-temperature residual service life of the lithium battery;
and calculating the high-temperature predicted residual service life of the lithium battery by using the determined high-temperature residual service life prediction equation of the lithium battery to be detected, and correcting the high-temperature predicted residual service life of the lithium battery by using the high-temperature residual service life correction prediction equation of the lithium battery to obtain the actual high-temperature residual service life prediction value of the lithium battery.
The accuracy of the predicted value of the high-temperature remaining service life of the lithium battery predicted by the method of the first embodiment is lower than that of the predicted value of the high-temperature remaining service life of the lithium battery predicted by the method of the second embodiment, so that the method of the first embodiment is suitable for scenes with less computing resources and low prediction accuracy.
The obtained lithium battery low-temperature residual service life prediction model and the lithium battery high-temperature residual service life correction prediction equation comprise:
carrying out residual service life prediction training on a linear prediction model with good linear prediction effect by using low-temperature capacity degradation data of the lithium battery in a lithium battery historical database to obtain a low-temperature residual service life prediction model of the lithium battery;
determining a prediction equation of the high-temperature residual service life of the lithium battery by utilizing the low-temperature residual service life of the lithium battery predicted by the low-temperature residual service life prediction model of the lithium battery and the Allen-nies model;
the high-temperature residual service life correction prediction equation of the lithium battery is obtained by verifying and processing the actual high-temperature residual service life value of the lithium battery in the lithium battery historical database, the high-temperature predicted residual service life value of the lithium battery predicted by the high-temperature residual service life prediction equation of the lithium battery and the low-temperature predicted residual service life value of the lithium battery predicted by the low-temperature residual service life prediction model of the lithium battery.
The low-temperature actual remaining service life of the lithium battery comprises the actual remaining service life of the lithium battery at 25 ℃ and the actual remaining service life of the lithium battery at 45 ℃; the low-temperature predicted residual service life of the lithium battery comprises the predicted residual service life of the lithium battery at 25 ℃ and the predicted residual service life of the lithium battery at 45 ℃; the actual residual service life of the lithium battery at high temperature is the actual residual service life of the lithium battery at 60 ℃; the predicted residual service life of the lithium battery at the high temperature is the predicted residual service life of the lithium battery at 60 ℃.
The training of the predicted value of the remaining service life of the linear prediction model by using the low-temperature capacity degradation data of the lithium battery in the lithium battery historical database comprises the following steps:
and training an autoregressive integral sliding average model ARIMA by using the 25 ℃ capacity degradation data and the 45 ℃ capacity degradation data of the lithium battery and the 25 ℃ residual service life and the 45 ℃ residual service life of the lithium battery in a lithium battery historical database to obtain a trained low-temperature residual service life prediction model for predicting the 25 ℃ residual service life and the 45 ℃ residual service life of the lithium battery.
The method for determining the high-temperature residual service life prediction equation of the lithium battery by using the low-temperature residual service life of the lithium battery and the Arrhenius model comprises the following steps:
determining coefficients in an Arrhenius model by using the residual service life of the lithium battery at 25 ℃ and the residual service life of the lithium battery at 45 ℃;
and using the determined coefficients of the Arrhenius model as a prediction equation of the high-temperature residual service life of the lithium battery.
Preferably, determining the coefficients in the arrhenius model using the remaining service life of the lithium battery at 25 ℃ and the remaining service life of the lithium battery at 45 ℃ comprises:
simplifying an Arrhenius model into an equation ln L ═ a + b/T, wherein L is a residual life value, T is temperature, a is a first coefficient, and b is a second coefficient;
the method comprises the steps of solving a first coefficient a and a second coefficient b in ln L ═ a + b/T by using a residual service life value of the lithium battery at 25 ℃ and a residual service life value of the lithium battery at 45 ℃, so as to obtain an Allen model simplified equation for predicting the residual service life of the lithium battery at 60 ℃, and further calculate the predicted value of the residual service life of the lithium battery to be tested at 60 ℃ by using the equation ln L ═ a + b/T.
The method for obtaining the high-temperature residual service life correction prediction equation of the lithium battery comprises the following steps:
obtaining a ratio curve corresponding to a series of ratios of a series of actual service life values of the high temperature lithium battery to a series of high temperature predicted remaining service life values of the lithium battery by utilizing a series of actual service life values of the high temperature lithium battery in a lithium battery historical database and a series of high temperature predicted remaining service life values of the lithium battery calculated by a lithium battery high temperature remaining service life prediction equation;
obtaining a difference curve corresponding to a series of differences between the predicted residual service life values of the series of lithium batteries at 25 ℃ and the predicted residual service life values of the series of lithium batteries at 45 ℃ by utilizing a series of differences between the predicted residual service life values of the series of lithium batteries at 25 ℃ and the predicted residual service life values of the series of lithium batteries at 45 ℃ predicted by the low-temperature residual service life prediction model of the lithium batteries;
and fitting a lithium battery high-temperature residual service life correction prediction equation which enables the difference value of the predicted residual service life value of the lithium battery at 25 ℃ and the predicted residual service life value of the lithium battery at 45 ℃ to be equal to the ratio of the lithium battery high-temperature actual service life value and the lithium battery high-temperature predicted residual service life value according to the ratio curve and the difference value curve.
Wherein, using the lithium battery high temperature remaining service life correction prediction equation to correct the lithium battery high temperature remaining service life comprises:
and after the high-temperature predicted residual service life of the lithium battery is calculated by using the determined prediction equation of the high-temperature residual service life of the lithium battery to be detected, multiplying the calculation result of the correction prediction equation of the high-temperature residual service life of the lithium battery by the calculated value of the high-temperature predicted residual service life of the lithium battery.
The high-temperature remaining service life correction prediction equation of the lithium battery is as follows: ratio ═ f (difference); wherein, the ratio is the ratio of the actual high-temperature service life value of the lithium battery to the predicted residual high-temperature service life value of the lithium battery; (difference) is a fitting function with the variable being the difference between the predicted remaining life value of the lithium battery at 25 ℃ and the predicted remaining life value of the lithium battery at 45 ℃.
The above-mentioned method of the present invention will be described in detail with reference to the accompanying drawings.
The invention provides a method for predicting the remaining service life of a lithium battery based on data drive and model drive fusion, which considers different temperatures, and a specific design flow chart is shown in figure 1.
Because the linear degree of the degradation curve of the lithium battery at different temperatures is different. As shown in fig. 2, generally, the linearity is high at low temperatures and low at high temperatures. Therefore, the service life prediction at different temperatures needs to be distinguished, and the respective advantages of the data driving method and the model driving method are utilized to carry out targeted solution. The capacity degradation curves at 25 ℃ and 45 ℃ are highly linear and therefore predicted using a data-driven approach. The capacity degradation curve at 60 ℃ is low in linearity, and therefore prediction is performed using an Arrhenius model. However, the accuracy of the extrapolation life of the Arrhenius model is not high, and a result correction coefficient is needed for auxiliary prediction.
And (3) establishing a result correction coefficient equation:
the method comprises the following steps: an Autoregressive Integrated Moving Average (ARIMA) model with a good linear prediction effect in a data driving method is used for predicting the residual service life of the battery in the lithium battery historical database at 25 ℃ and 45 ℃.
Step two: and (3) identifying two parameters in the Arrhenius model by using the predicted values of the residual service life of the lithium battery at 25 ℃ and 45 ℃ obtained in the step one to obtain the final Arrhenius model.
Step three: and predicting the residual service life at 60 ℃ by using an Arrhenius model to obtain a predicted value of the residual service life of the battery at 60 ℃ in a historical database of the lithium battery.
Step four: and repeating the first step to the third step to complete the prediction of the service life of all the batteries in the lithium battery historical database at 25 ℃, 45 ℃ and 60 ℃. And obtaining a result correction coefficient equation according to the prediction result.
Predicting the service life of the lithium battery to be predicted:
the method comprises the following steps: repeating the steps from the first step to the third step in the step of establishing an 'result correction coefficient equation' on the battery to be predicted, correcting the cycle life at the temperature of 60 ℃ by using the result correction coefficient equation obtained in the step four, and giving a final prediction result at the temperature of 25 ℃, 45 ℃ and 60 ℃.
The ARIMA model in the first step of the establishment of the result correction coefficient equation is specifically applied as follows:
firstly, the parameters are ordered. ARIMA (p, d, q) has three parameters of an autoregressive parameter p, the difference number d and a moving average parameter q. Capacity degradation data of the lithium battery in the first 500 cycles is used as input, and the data is differentiated for d times to ensure that a sequence after differentiation is stable; then, an autoregressive equation (ACF) and a Partial autoregressive equation (PACF) of the sequence after the difference are solved, and an autoregressive parameter p and a moving average parameter q are determined by observing the trailing condition of the ACF and the PACF.
And secondly, parameter fitting. Through regression analysis, a parameter α of an Autoregressive (AR) term and a parameter β of a Moving regression (MA) term are fitted, and a significance test and a residual randomness test are performed on the parameters α and β.
And thirdly, checking and applying. The obtained ARIMA (p, d, q) model was checked for agreement with historical data. And if the predicted value is matched with the historical data, the model is applied to predict the capacity degradation condition, and finally the predicted value of the residual service life is obtained.
Wherein, the "Arrhenius model" in the step two of the "establishment of the result correction coefficient equation" is established in the following process:
in 1989, Arrhenius concluded that the reaction of acid-catalyzed hydrolysis and conversion of sucrose was carried out at a temperature that was studied: the rate of degradation of a product's performance is inversely proportional to the index of activation energy and inversely proportional to the index of the inverse temperature, which can be expressed as:
wherein M is the degradation amount of a certain characteristic value of the product;is the degradation rate at temperature T (thermodynamic temperature), which is a linear function of time T; k is Boltzmann's constant of 8.617 × 10-5eV/DEG C; t is the absolute temperature, equal to 273.15 degrees Celsius plus; a. the0Is a constant; t is the reaction time; Δ E is the failure mechanism activation energy, in eV, constant for the same failure mode of the same type of device.
The amount of degradation of the product in the initial state is set to M1Corresponding to time t1(ii) a The amount of degradation of the other state is M2Corresponding to time t2. Then, when T is constant, from T1To t2The cumulative amount of degradation of (a):
then there is
M2-M1＝A0·exp[-ΔE/(kT)]·(t2-t1) (3)
Let t be (t)2-t1) Then there is
When the amount of degeneration M2Reaches a certain threshold value MpWhen it is, the product is considered to be out of service, the time difference (t) isp-t1) That is the product is from t1Beginning extended life L. Namely, it is
Order:
then there is
ln L＝a+b/T (6)
Wherein a and b are undetermined parameters. Therefore, knowing any two temperatures and the remaining service life of the lithium battery at the temperatures, the values of the two parameters a and b can be identified by substituting the values into equation (6), and the remaining service life at any temperature can be extrapolated.
In the four steps, the method for establishing the result correction coefficient equation comprises the following steps:
according to the analysis of the test data, a fitting relation is found between the difference between the predicted service life at 25 ℃ and the predicted service life at 45 ℃ and the ratio between the actual service life at 60 ℃ and the extrapolated service life of the Allenix model at 60 ℃. By fitting the historical data, a result correction coefficient can be obtained by fitting an equation.
ratio＝f(difference) (7)
(1) Advantages of the invention
① the method combines the advantages of the data driving method and the model driving method, solves the problem of the remaining service life of the lithium battery at the temperature of 25 ℃, 45 ℃ and 60 ℃ in a targeted manner, and has high prediction accuracy.
In one embodiment of the invention, the feasibility and the effectiveness of the proposed lithium ion battery life prediction method are verified by adopting test data of Ningde time New energy science and technology Co., Ltd (note that the battery used in the test is a soft package battery specially used in the product design stage, which is different from the battery used in the real product of the company).
According to the difference of design of anode materials, electrolyte and the like, the batch of lithium batteries are divided into 10 groups A-J, wherein the groups A-I are used as a historical database to perform result correction regression equation regression, the group J is used for verification of a final prediction method, and specific information of data used in the example is shown in Table 1. It is noted that the batch of cells is the test cell at the design stage and not the final cell product.
TABLE 1 lithium cell information
The method comprises the following specific steps:
and (3) establishing a result correction coefficient equation:
the method comprises the following steps: specifically, the a-group 25 ℃ data is used as the target, and the original data is as shown in fig. 4, ARIMA (p, d, q) is applied to the first 500 cycles of the group data, so as to obtain the model ARIMA (1,1,11) with the model parameters p being 1, d being 1, and q being 11. Parameter regression is carried out to obtain a parameter alpha of an autoregressive term with a constant value of-8.865E-51Parameter β of moving average term, 0.3991＝0.767，β110.129. The model obtained finally is as follows:
(Ct-Ct-1)＝-8.865×10-5+0.399(Ct-1-Ct-2)+0.767μt-1+0.129μt-11(8)
wherein C istIs the battery capacity at the t-th cycle, μtIs the t-th value of the white noise sequence. It can be seen that the model can well predict the long-term cycle life of the battery, and the final prediction result is shown in fig. 3.
And repeating the first step to obtain the service life prediction results of other batteries in the lithium battery historical database at 25 ℃ and 45 ℃, as shown in Table 2.
TABLE 225 ℃ and 45 ℃ predicted Life
Step two: taking group A as an example, taking group A as the predicted lifetime at 25 ℃ as 3022 and group A as the predicted lifetime at 45 ℃ as 2362, as shown in Table 2, T1＝25+273.15；L13022 and T2＝45+273.15；L12362 is substituted into the equation lnL a + b/T, the parameters a 4.0939 and b 1168.7 are calculated, and the arrhenius model is finally obtained as follows:
lnL＝4.0939+1168.7/T (9)
step three: substituting T60 +273.15 into equation 9 resulted in a predicted lifetime of 2002 cycles at 60 ℃.
Step four: the predicted results at 25 ℃ and 45 ℃ from group a to group I in table 2 were combined and used as input to perform step two and step three to obtain the extrapolated life results for the individual groups of cells at 60 ℃ in the arrhenius model, as shown in table 3.
TABLE 360 ℃ Allenis model extrapolated lifetime
From group A to group I, the difference between the predicted lifetime at 25 ℃ and the predicted lifetime at 45 ℃ and the ratio of the actual lifetime at 60 ℃ to the extrapolated lifetime of the Arrhenius model at 60 ℃ ratio, are shown in Table 4 for the regression data of the result correction equation in Table 4
Note: two columns of values in Table 4 correspond one-to-one with commas as delimiters
Regression analysis is performed on the two columns of data in table 4 to obtain the following regression equation:
ratio＝2.87e-06difference1.61+0.4629 (10)
predicting the service life of the lithium battery to be predicted:
the method comprises the following steps: the predicted life of the J-cell was 2713 cycles at 25 c, 1363 cycles at 45 c, and 1350 cycles apart, as shown in table 1. The difference is substituted into equation (10) to obtain a resultant correction coefficient of 0.7775. Substituting both into the Arrhenius model gave an extrapolated lifetime of 859 cycles at 60 ℃ as shown in Table 3. The extrapolated life times the resulting correction factor gives a predicted life at 60 ℃ of 859 × 0.7775 ≈ 668 cycles. As shown in table 3, the actual life of the J-pack cell at 60 ℃ was 703 cycles. The difference between the predicted life and the actual life is 668-703 ═ 35 cycles, and the prediction precision is 1- |668-703|/703 × 100 | -95.02%.
Although the present invention has been described in detail hereinabove, the present invention is not limited thereto, and various modifications can be made by those skilled in the art in light of the principle of the present invention. Thus, modifications made in accordance with the principles of the present invention should be understood to fall within the scope of the present invention.
Claims (8)
1. A digital-analog fusion prediction method for the remaining service life of a lithium battery comprises the following steps:
carrying out residual service life prediction training on a linear prediction model with good linear prediction effect by using low-temperature capacity degradation data of the lithium battery in a lithium battery historical database to obtain a low-temperature residual service life prediction model of the lithium battery;
determining a coefficient of an Arrhenius model by using a predicted value of the low-temperature residual service life of the lithium battery predicted by a lithium battery low-temperature residual service life prediction model, and taking the Arrhenius model with the determined coefficient as a lithium battery high-temperature residual service life prediction equation;
verifying and processing the actual high-temperature residual service life value of the lithium battery in the lithium battery historical database, the high-temperature predicted residual service life value of the lithium battery predicted by the high-temperature residual service life prediction equation of the lithium battery and the low-temperature predicted residual service life value of the lithium battery predicted by the low-temperature residual service life prediction model of the lithium battery to obtain a high-temperature residual service life correction prediction equation of the lithium battery;
inputting part of low-temperature capacity degradation data of the lithium battery to be tested into the lithium battery low-temperature remaining service life prediction model, and predicting the low-temperature remaining service life of the lithium battery to be tested to obtain a predicted value of the low-temperature remaining service life of the lithium battery;
determining the coefficient of an Arrhenius model by using the predicted value of the low-temperature residual service life of the lithium battery, and taking the Arrhenius model with the determined coefficient as a prediction equation of the high-temperature residual service life of the lithium battery to be tested;
and calculating the high-temperature predicted residual service life of the lithium battery by using the high-temperature residual service life prediction equation of the lithium battery to be detected, and correcting the high-temperature predicted residual service life of the lithium battery by using the high-temperature residual service life correction prediction equation of the lithium battery to obtain the actual high-temperature residual service life prediction value of the lithium battery.
2. The method of claim 1, wherein the low temperature remaining life of the lithium battery comprises an actual remaining life of the lithium battery at 25 ℃ and an actual remaining life of the lithium battery at 45 ℃; the predicted value of the low-temperature residual service life of the lithium battery comprises a predicted value of the residual service life of the lithium battery at 25 ℃ and a predicted value of the residual service life of the lithium battery at 45 ℃; the actual predicted value of the high-temperature remaining service life of the lithium battery is the predicted value of the actual remaining service life of the lithium battery at 60 ℃; the predicted residual service life of the lithium battery at the high temperature is the predicted residual service life of the lithium battery at 60 ℃.
3. The method of claim 2, wherein the training of the linear prediction model with good prediction effect by using the low-temperature capacity degradation data of the lithium batteries in the lithium battery historical database for residual service life prediction comprises:
and training an autoregressive integral sliding average model ARIMA by using the 25 ℃ capacity degradation data and the 45 ℃ capacity degradation data of the lithium battery and the 25 ℃ residual service life and the 45 ℃ residual service life of the lithium battery in a lithium battery historical database to obtain a trained low-temperature residual service life prediction model for predicting the 25 ℃ residual service life and the 45 ℃ residual service life of the lithium battery.
4. The method according to claim 3, wherein the step of determining the coefficients of the Arrhenius model by using the predicted value of the low-temperature residual service life of the lithium battery, and the step of using the Arrhenius model with the determined coefficients as a prediction equation of the high-temperature residual service life of the lithium battery comprises the following steps:
determining coefficients in an Arrhenius model by using the residual service life of the lithium battery at 25 ℃ and the residual service life of the lithium battery at 45 ℃;
and using the determined coefficients of the Arrhenius model as a prediction equation of the high-temperature residual service life of the lithium battery.
5. The method of claim 4, wherein determining the coefficients in the Arrhenius model using the remaining service life of the lithium battery at 25 ℃ and the remaining service life of the lithium battery at 45 ℃ comprises:
simplifying an Arrhenius model into an equation ln L ═ a + b/T, wherein L is a residual life value, T is temperature, a is a first coefficient, and b is a second coefficient;
the method comprises the steps of solving a first coefficient a and a second coefficient b in ln L ═ a + b/T by using a residual service life value of the lithium battery at 25 ℃ and a residual service life value of the lithium battery at 45 ℃, so as to obtain an Allen model simplified equation for predicting the residual service life of the lithium battery at 60 ℃, and further calculate the predicted value of the residual service life of the lithium battery to be tested at 60 ℃ by using the equation ln L ═ a + b/T.
6. The method of claim 2, wherein the obtaining the modified prediction equation of the high-temperature remaining service life of the lithium battery comprises:
obtaining a ratio curve corresponding to a series of ratios of a series of actual service life values of the high temperature lithium battery to a series of high temperature predicted remaining service life values of the lithium battery by utilizing a series of actual service life values of the high temperature lithium battery in a lithium battery historical database and a series of high temperature predicted remaining service life values of the lithium battery calculated by a lithium battery high temperature remaining service life prediction equation;
obtaining a difference curve corresponding to a series of differences between the predicted residual service life values of the series of lithium batteries at 25 ℃ and the predicted residual service life values of the series of lithium batteries at 45 ℃ by utilizing a series of differences between the predicted residual service life values of the series of lithium batteries at 25 ℃ and the predicted residual service life values of the series of lithium batteries at 45 ℃ predicted by the low-temperature residual service life prediction model of the lithium batteries;
and fitting a lithium battery high-temperature residual service life correction prediction equation which enables the difference value of the predicted residual service life value of the lithium battery at 25 ℃ and the predicted residual service life value of the lithium battery at 45 ℃ to be equal to the ratio of the lithium battery high-temperature actual service life value and the lithium battery high-temperature predicted residual service life value according to the ratio curve and the difference value curve.
7. The method of claim 6, wherein correcting the high temperature remaining useful life correction prediction equation for the lithium battery comprises:
and after the high-temperature predicted residual service life of the lithium battery is calculated by using the determined prediction equation of the high-temperature residual service life of the lithium battery to be detected, multiplying the calculation result of the correction prediction equation of the high-temperature residual service life of the lithium battery by the calculated value of the high-temperature predicted residual service life of the lithium battery.
8. The method of claim 7, wherein the modified prediction equation for the high temperature remaining life of the lithium battery is as follows: ratio ═ f (difference);
wherein, the ratio is the ratio of the actual high-temperature service life value of the lithium battery to the predicted residual high-temperature service life value of the lithium battery; (difference) is a fitting function with the variable being the difference between the predicted remaining life value of the lithium battery at 25 ℃ and the predicted remaining life value of the lithium battery at 45 ℃.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610801168 | 2016-09-05 | ||
| CN2016108011681 | 2016-09-05 |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107544033A CN107544033A (en) | 2018-01-05 |
| CN107544033B true CN107544033B (en) | 2020-01-10 |
Family
ID=60959280
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710791239.9A Active CN107544033B (en) | 2016-09-05 | 2017-09-05 | Digital-analog fusion prediction method for remaining service life of lithium ion battery |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107544033B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110058177B (en) * | 2019-05-06 | 2021-08-31 | 奇瑞新能源汽车股份有限公司 | Power battery electric quantity SOC correction method |
| CN110658462B (en) * | 2019-10-25 | 2020-12-15 | 北京航空航天大学 | Lithium battery online service life prediction method based on data fusion and ARIMA model |
| CN110888077A (en) * | 2019-10-30 | 2020-03-17 | 无锡市产品质量监督检验院 | Accelerated lithium ion battery life evaluation method based on ARIMA time sequence |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104714189A (en) * | 2015-04-02 | 2015-06-17 | 奇瑞汽车股份有限公司 | Method for predicting cycle life of battery pack for electric car |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0227278A (en) * | 1988-07-15 | 1990-01-30 | Fujitsu Ltd | Battery life checker |
| WO2010001605A1 (en) * | 2008-07-02 | 2010-01-07 | パナソニック株式会社 | Life span estimation method for a lead acid battery and a power supply system |
| JP2010117256A (en) * | 2008-11-13 | 2010-05-27 | Toshiba Toko Meter Systems Co Ltd | Battery lifetime prediction apparatus |
| CN102778653A (en) * | 2012-06-20 | 2012-11-14 | 哈尔滨工业大学 | Data-driven lithium ion battery cycle life prediction method based on AR (Autoregressive) model and RPF (Regularized Particle Filtering) algorithm |
| CN103344923B (en) * | 2013-08-01 | 2015-07-08 | 哈尔滨工业大学 | EKF (extended Kalmar filter)-method and NSDP-AR-model-based method for predicting cycling life of fusion-type lithium ion battery |
| CN103399281B (en) * | 2013-08-01 | 2016-01-20 | 哈尔滨工业大学 | Based on the ND-AR model of cycle life deterioration stage parameter and the cycle life of lithium ion battery Forecasting Methodology of EKF method |
| CN103399279B (en) * | 2013-08-01 | 2015-12-09 | 哈尔滨工业大学 | Based on EKF method and AR Model Fusion type cycle life of lithium ion battery Forecasting Methodology |
| CN103954913B (en) * | 2014-05-05 | 2017-06-30 | 哈尔滨工业大学深圳研究生院 | Electric automobile power battery life-span prediction method |
-
2017
- 2017-09-05 CN CN201710791239.9A patent/CN107544033B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104714189A (en) * | 2015-04-02 | 2015-06-17 | 奇瑞汽车股份有限公司 | Method for predicting cycle life of battery pack for electric car |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107544033A (en) | 2018-01-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Lipu et al. | A review of state of health and remaining useful life estimation methods for lithium-ion battery in electric vehicles: Challenges and recommendations | |
| Xiong et al. | An electrochemical model based degradation state identification method of Lithium-ion battery for all-climate electric vehicles application | |
| Dai et al. | A novel estimation method for the state of health of lithium-ion battery using prior knowledge-based neural network and Markov chain | |
| Yang et al. | Extreme learning machine-based thermal model for lithium-ion batteries of electric vehicles under external short circuit | |
| Wang et al. | A method for state-of-charge estimation of Li-ion batteries based on multi-model switching strategy | |
| Yu et al. | OCV-SOC-temperature relationship construction and state of charge estimation for a series–parallel lithium-ion battery pack | |
| Deng et al. | Sensitivity analysis and joint estimation of parameters and states for all-solid-state batteries | |
| CN110888057B (en) | Power lithium ion battery electrochemical parameter identification method and system | |
| Hussein | Adaptive artificial neural network-based models for instantaneous power estimation enhancement in electric vehicles’ Li-ion batteries | |
| CN110941929A (en) | Battery health state assessment method based on ARMA and Elman neural network combined modeling | |
| CN107544033B (en) | Digital-analog fusion prediction method for remaining service life of lithium ion battery | |
| Pei et al. | An Equivalent Circuit Model for Lithium Battery of Electric Vehicle considering Self‐Healing Characteristic | |
| CN112986830A (en) | Lithium battery capacity estimation method based on convolution time memory neural network | |
| CN106777786A (en) | A kind of lithium ion battery SOC estimation method | |
| WO2023030024A1 (en) | Electrochemical model-based method and system for estimating state of solid-state lithium battery | |
| CN113794254B (en) | Thermal management strategy configuration method and device, computer equipment and storage medium | |
| Baek et al. | Capacity fade modeling of a lithium-ion battery for electric vehicles | |
| CN107895175A (en) | A kind of method degenerated based on Nonhomogeneous Markov Chains model prediction capacity of lithium ion battery | |
| Zhang et al. | Identification of dynamic model parameters for lithium-ion batteries used in hybrid electric vehicles | |
| De Sousa et al. | Comparison of different approaches for lead acid battery state of health estimation based on artificial neural networks algorithms | |
| Zhang et al. | A data-driven based state of energy estimator of lithium-ion batteries used to supply electric vehicles | |
| Binelo et al. | Mathematical modeling and parameter estimation of battery lifetime using a combined electrical model and a genetic algorithm | |
| Zhang et al. | Real-time estimation of negative electrode potential and state of charge of lithium-ion battery based on a half-cell-level equivalent circuit model | |
| Uddin et al. | Characterising Li-ion battery degradation through the identification of perturbations in electrochemical battery models | |
| CN113721151B (en) | Battery capacity estimation model and method based on double-tower deep learning network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |