CN105116337A - Lithium ion battery full charge storage service life evaluation method - Google Patents

Abstract

The invention discloses a lithium ion battery full charge storage service life evaluation method. The method comprises steps: a battery sample for evaluation of lithium ion battery full charge states is stored at a plurality of storage temperature values T for appointed sampling time t, a capacity loss rate Q is obtained, and experiment data are generated; a capacity attenuation aged model is established, and values of a model parameter rho and a model parameter a are determined; the selection rationality of the storage temperature scope is determined based on the situation whether the model parameter a meets the Arrhenius formula at each storage temperature value T; a mean value and a standard deviation standard deviation of a model A and a model B are obtained through statistics; the value of the model parameter rho and the mean value of the models A and B are substituted into the model and fitting goodness determination is carried out; after fitting goodness is met, the service life of the lithium ion battery to be evaluated is predicted at a normal temperature and a service life distribution diagram is obtained. The evaluation method has advantages of sufficient service life evaluation principle, high data reliability, high precision, short evaluation time, wide application scope, is simple and practicable, and is easy to implement.

Description

A kind of full charge storage life assessment method of lithium ion battery

Technical field

The present invention relates to the storage life assessment technology of lithium ion battery, be specifically related to a kind of full charge storage life assessment method of lithium ion battery utilizing high temperature to accelerate.

Background technology

Along with the development of society and the progress of science and technology, lithium ion battery is widely used as energy resource system of new generation.The advantages such as lithium ion battery, as electrical source of power, is compared traditional oils electrical source of power and had high security, pollution-free, make lithium battery have power supply on vehicle application prospect, successfully produce the pure electric automobile based on lithium battery both at home and abroad at present and extensively sell.

But the life-span of lithium ion battery is one of maximum restraining factors limiting its widespread use.Usually, the end of life of battery is defined as 80% of its initial capacity.By external, the electrical source of power requirement of lithium ion battery as electric automobile is proposed, its, target was use 15 years in serviceable life, after circulating battery more than 45000 times, its capacity still remains on more than 80%, and propose to detect by test the life-span analyzing battery in 1 ~ 2 year, current driving force battery is still difficult to reach this life level.

In the life-span of the method rapid evaluating cell accelerated by high temperature, to the research promoting long-life batteries, there is extremely crucial effect.For battery memory requirement under different state-of-charge, battery storage life accelerated deterioration evaluation method conventional both at home and abroad is at present generally empirical method.Thumb rule is based upon on the basis of experience, propose on the basis meeting certain condition, the storing temperature of battery raises 10 DEG C, the capacity attenuation speed increase of battery about one times, this method simply can estimate the lithium ion battery full charge storage life-span, but the confidence level of data is not high, there is the shortcomings such as degree of accuracy is not high, evaluation time is long.Therefore, how accurately estimate the lithium ion battery full charge storage life-span, become lithium ion battery by during wide popularization and application be badly in need of one of key technical problem solved.

Summary of the invention

The technical problem to be solved in the present invention is: for the problems referred to above of prior art, there is provided abundant, the high data reliability of a kind of life assessment principle and high precision, simple, be easy to realize, evaluation time is short, applied range, can provide the lithium ion battery of reliable technical support full charge storage life assessment method to the improvement of the storage life of lithium ion battery.

In order to solve the problems of the technologies described above, the technical solution used in the present invention is:

A kind of full charge storage life assessment method of lithium ion battery, step comprises:

1) according to storage minimum temperature and the electrolyte decomposition temperature determination storage temperature range of lithium ion battery to be evaluated, multiple storing temperature value T is chosen according to the temperature interval of presetting in storage temperature range, the battery sample evaluating the full state of charge of lithium ion battery is stored t sample time specified under multiple storing temperature value T, under each storing temperature value T each sample time t as a sampling spot, each sampling spot at least one battery sample corresponding, obtains the capacitance loss rate Q of battery sample under each sampling spot and generates experimental data;

2) set up capacity attenuation Ageing Model shown in formula (1), experimental error is considered on the basis of described experimental data, and utilize the value of capacity attenuation Ageing Model Confirming model parameter ρ shown in formula (1); On the basis of the value of Confirming model parameter ρ, utilize the nonlinear curve function of origin software to carry out matching to capacity attenuation Ageing Model formula (1) Suo Shi, obtain the value of simulation curve under each storing temperature value T and model parameter a;

$\left\{\begin{array}{c}Q={\mathrm{at}}^{\mathrm{\ρ}}\\ a=A-B/T\end{array}\right.---\left(1\right)$

In formula (1), Q represents the capacitance loss rate of battery sample, and T represents storing temperature value, and t represents sample time, and a, ρ, A, B are model parameter to be solved;

3) whether meet based on the model parameter a under each storing temperature value T the selection rationality that Arrhenius formula judges storage temperature range, if rationally, determine capacitance loss rate Q in experimental data, storing temperature value T, sample time t error range and redirect performs step 4); If unreasonable, then cast out the number of sampling certificate of a highest storing temperature value T from experimental data, redirect performs step 2);

4) by capacitance loss rate Q, storing temperature value T in experimental data, sample time t and model parameter ρ value substitute into the value that formula (2) illustrated equation obtains model parameter A, B, and utilize MonteCarlo method based on capacitance loss rate Q, storing temperature value T in described experimental data, sample time t error range draw the value of n group model parameter A, B, the value of n group model parameter A, B is added up to the average and standard deviation that draw model parameter A, B;

x＝(R ^TR) ^-1R ^TQ(2)

In formula (2), R ^tthe transposed matrix of representing matrix R, the expression formula of matrix Q is such as formula shown in (3), and the expression formula of matrix R is such as formula shown in (4), and the expression formula of matrix x is such as formula shown in (5);

$Q=\left(\begin{array}{c}{\mathrm{lnQ}}_{11}\\ \·\\ \·\\ \·\\ {\mathrm{lnQ}}_{mn}\end{array}\right)---\left(3\right)$

In formula (3), Q ₁₁represent first storing temperature value T ₁₁lower first sample time t ₁₁corresponding capacitance loss rate, Q _mnrepresent m storing temperature value T _mnlower n-th sample time t _mncorresponding capacitance loss rate;

$R=\left(\begin{array}{ccc}1& 1/T_{11}& {\mathrm{lnt}}_{11}\\ M& O& M\\ 1& 1/T_{mn}& {\mathrm{lnt}}_{mn}\end{array}\right)\text{---}\left(4\right)$

In formula (4), T ₁₁represent first storing temperature value, T _mnrepresent m storing temperature value, t ₁₁represent first storing temperature value T ₁₁under first sample time t ₁₁, t _mnrepresent m storing temperature value T _mnunder the n-th sample time;

$x=\left(\begin{array}{c}A\\ B\\ \mathrm{\ρ}\end{array}\right)---\left(5\right)$

In formula (5), the matrix in x expression (2), ρ, A, B are model parameter to be solved;

5) average of the value of model parameter ρ and model parameter A, B is substituted into capacity attenuation Ageing Model shown in formula (1), and utilize origin curve to carry out curve fitting, judge that matching obtains the coefficient of determination R of curve ²whether be greater than default threshold value, if be greater than default threshold value, judge that the goodness of fit meets the demands, redirect performs step 6); Otherwise judge that the goodness of fit does not meet the demands, adjustment capacitance loss rate Q, storing temperature value T, sample time t error range, redirect performs step 4);

6) according to the interval of the average of described model parameter A, B and standard deviation Confirming model parameter A, B, in described interval, the value of model parameter A, B is solved confidence interval of mean mode under the degree of confidence preset according to known standard deviation and carries out random value, obtain the value of many group models parameter A, B; By the method for MonteCarlo, the value of the value of model parameter ρ and described many group models parameter A, B is substituted into capacity attenuation Ageing Model shown in formula (1) to predict the lithium ion battery to be evaluated life-span at normal temperatures, obtain the life-span distribution plan of lithium ion battery to be evaluated.

Preferably, described step 1) after also comprise revision sample time t step, concrete steps comprise: in judgment experiment data, under each storing temperature value T, whether the capacitance loss rate Q of battery sample exceedes default loss-rate threshold, if the capacitance loss rate Q of battery sample exceedes loss-rate threshold under some storing temperature value T, then t sample time under this storing temperature value T is carried out revising and obtain new sampling spot, and the battery sample storing full state of charge is placed based on new sampling spot, obtain the capacitance loss rate Q of battery sample under new sampling spot;

Preferably, described step 1) in the calculation expression of capacitance loss rate Q such as formula shown in (6);

Q＝1-Q ₁/Q ₀(6)

In formula (6), Q is the capacitance loss rate of battery sample under some sampling spots, Q ₀represent the capacity of full electric charge battery sample, Q ₁represent the capacity of battery sample when sampling spot.

Preferably, described step 1) in corresponding three battery samples of each sampling spot, and after generation experimental data, also comprise the step that data are rejected and benefit is surveyed, detailed step comprises: for three capacitance loss rate Q of three battery samples of each sampling spot, first maximal value is wherein obtained, minimum value and intermediate value, then by intermediate value respectively and maximal value, minimum value compares, if the error of intermediate value and mxm., intermediate value and the error with minimum value are all in predetermined threshold value, be then maximal value by the capacitance loss rate Q value of this sampling spot, minimum value, the average of intermediate value three, if only the error of maximal value and intermediate value exceedes predetermined threshold value in maximal value, minimum value, then rejecting maximal value, is intermediate value and the average with minimum value by the capacitance loss rate Q value of this sampling spot, if only the error of minimum value and intermediate value exceedes predetermined threshold value in maximal value, minimum value, then rejecting minimum value, is intermediate value and the average with maximal value by the capacitance loss rate Q value of this sampling spot, if the error of maximal value, both minimum value and intermediate value all exceedes predetermined threshold value, then reappose battery sample for this sampling spot and mend and survey capacitance loss rate Q.

Preferably, described step 3) detailed step as follows:

3.1) for the value of the variable a under each storing temperature value T in experimental data, utilize origin linear curve fit lna value to 1/T mapping matched curve;

3.2) coefficient of determination R of curve is obtained according to matching ²whether judge that matching obtains curve is straight line, if it is straight line that matching obtains curve, then judge that each storing temperature value T drag parameter a meets Arrhenius formula, the selection of storage temperature range is reasonable, determine capacitance loss rate Q in experimental data, storing temperature value T, sample time t error range and redirect performs step 4); If it is not straight line that matching obtains curve, judge that each storing temperature value T drag parameter a does not meet Arrhenius formula, the selection of storage temperature range is unreasonable, casts out the number of sampling certificate of a highest storing temperature value T from experimental data, and redirect performs step 2).

Preferably, described step 4) in n group model parameter A, B value in n value be greater than 1000.

Preferably, described step 4) in the function expression that adopts when the average drawing model parameter A, B is added up to the value of n group model parameter A, B specifically such as formula shown in (7);

$\stackrel{\‾}{X}=\frac{1}{n}\underset{i=1}{\overset{n}{\Σ}}{X}_i---\left(7\right)$

In formula (7), represent the average calculated, X _irepresent i-th value in n group model parameter A or B, n represents the quantity of model parameter A or B.

Preferably, described step 4) in the function expression that adopts when the standard deviation drawing model parameter A, B is added up to the value of n group model parameter A, B specifically such as formula shown in (8);

$S=\sqrt{\frac{1}{n}\underset{i=1}{\overset{n}{\Σ}}{(X_i-\stackrel{\‾}{X})}^2}---\left(8\right)$

In formula (8), S represents the standard deviation calculated, X _irepresent i-th value in n group model parameter A or B, n represents the quantity of model parameter A or B, represent the average of n group model parameter A or B.

Preferably, described step 6) in when the value of model parameter A, B is solved that confidence interval of mean mode carries out random value under the degree of confidence preset according to known standard deviation, the fiducial interval of random value is specifically such as formula shown in (9);

$\[\stackrel{\‾}{X}-{\mathrm{\μ}}_{1-\mathrm{\α}/2}\frac{S}{\sqrt{n}},\stackrel{\‾}{X}+{\mathrm{\μ}}_{1-\mathrm{\α}/2}\frac{S}{\sqrt{n}}\]---\left(9\right)$

In formula (9), represent the average of n group model parameter A or B, S represents the standard deviation of n group model parameter A or B, and n represents the quantity of model parameter A or B, and μ is fractile, and 1-α is fiducial interval.

The full charge storage life assessment method of lithium ion battery of the present invention has following advantage:

1, life assessment principle is abundant.The present invention is by step 1) battery sample evaluating the full state of charge of lithium ion battery is stored under multiple storing temperature value T the t generation sample time experimental data of specifying, and after setting up capacity attenuation Ageing Model shown in formula (1), the selection rationality that Arrhenius formula judges storage temperature range whether is met based on each storing temperature value T drag parameter a, therefore the present invention is that Corpus--based Method Epidemiological Analysis and Arrhenius formula carry out Acceleration study analysis to the ageing process of battery, make full use of statistical principle to simulate capacity attenuation in ageing process, derive the life-span of battery, there is the sufficient advantage of life assessment principle.

2, high data reliability and high precision.The general method empirically of traditional life assessment method obtains the speedup factor k (namely store under acceleration temperature T be equivalent to for a day storage at room temperature under k days) of battery, its error and degree of accuracy poor.And step 6 of the present invention) according to the interval of the average of described model parameter A, B and standard deviation Confirming model parameter A, B, in described interval, the value of model parameter A, B is solved confidence interval of mean mode under the degree of confidence preset according to known standard deviation and carries out random value, obtain the value of many group models parameter A, B; By the method for MonteCarlo, the value of the value of model parameter ρ and described many group models parameter A, B is substituted into capacity attenuation Ageing Model shown in formula (1) to predict the lithium ion battery to be evaluated life-span at normal temperatures, obtain the life-span distribution plan of lithium ion battery to be evaluated, therefore in the degree of confidence of Corpus--based Method, the rate of decay of the battery that the basis of the statistical analysis such as distributed area draws, therefore has high data reliability and high-precision advantage.

3, simple, be easy to realize: the full charge storage life assessment method of lithium ion battery of the present invention is simple, and simple to operate on the basis controlling to evaluate environment precision, the program of its simulation process obtains and is also easy to realize.

4, evaluation time is short.The single high temperature of conventional batteries life assessment method general accelerates the method in conjunction with normal temperature checking, and due to battery, the life-span is generally several years at normal temperatures, and therefore the correctness of verification model and speedup factor is for a long time consuming time.The full charge storage life assessment method of lithium ion battery of the present invention adopts the method for many temperature experiments to utilize the correctness of room temperature and other higher temperatures point contrast verification model and speedup factor simultaneously, can the correctness of verification model and speedup factor in the short period of time.

5, the full charge storage life assessment method of lithium ion battery of the present invention is evaluated the storage life of lithium ion battery under full charge condition, can evaluate to fast and reliable the life-span of lithium battery, the evaluation method of the storage life of lithium ion battery under full charge condition can be provided for the producer and user, ensure production and the application of product, reliable technical support can be provided to the improvement of the storage life of lithium ion battery.

Accompanying drawing explanation

Fig. 1 is the basic procedure schematic diagram of the embodiment of the present invention one (embodiment two).

Fig. 2 is the embodiment of the present invention one step 2) curve (aging tendency matched curve) that obtains of matching.

Fig. 3 is the embodiment of the present invention one step 5) curve that obtains of matching and real data contrast figure.

Fig. 4 is the embodiment of the present invention one step 6) the life-span distribution plan that obtains.

Fig. 5 is the embodiment of the present invention one step 3) in by lna value, figure is judged to the 1/T energy of activation obtained of mapping.

Fig. 6 is the embodiment of the present invention two step 2) curve (aging tendency matched curve) that obtains of matching.

Fig. 7 is the embodiment of the present invention two step 3) in by lna value, figure is judged to the 1/T energy of activation obtained of mapping.

Fig. 8 is the embodiment of the present invention two step 5) curve that obtains of matching and real data contrast figure.

Fig. 9 is the embodiment of the present invention two step 6) the life-span distribution plan that obtains.

Embodiment

Hereafter will so that abroad certain producer 18650 battery and certain producer 18650 battery domestic, as embodiment, are further detailed the full charge storage life assessment method of lithium ion battery of the present invention.

Embodiment one:

As shown in Figure 1, the step of the full charge storage life assessment method of the present embodiment lithium ion battery comprises:

1) determine that experiment condition carries out experiment and generates experimental data: according to storage minimum temperature and the electrolyte decomposition temperature determination storage temperature range of lithium ion battery to be evaluated, multiple storing temperature value T is chosen according to the temperature interval of presetting in storage temperature range, the battery sample evaluating the full state of charge of lithium ion battery is stored t sample time specified under multiple storing temperature value T, under each storing temperature value T each sample time t as a sampling spot, each sampling spot at least one battery sample corresponding, the capacitance loss rate Q obtaining battery sample under each sampling spot generates experimental data.

At present generally by room temperature mathematic(al) expectation for the requirement condition in lithium ion life-span.In the present embodiment, the storage minimum temperature of certain producer 18650 battery external is designed to 298K, general lithium-ion battery electrolytes decomposition temperature is at about 353K, therefore setting maximum temperature is 343K and gets interval temperature 10K, the storing temperature value T chosen is respectively 298K, 313K, 323K, 333K, 343K, t sample time specified is designed to 30 days, 60 days, 90 days, 150 days, 210 days, 270 days, 360 days, corresponding three battery samples of each sampling spot, the testing scheme of each sampling spot therefore designed is as shown in table 1.

Table 1: the testing scheme of each sampling spot of embodiment one.

In the present embodiment, step 1) after also comprise revision sample time t step, concrete steps comprise: in judgment experiment data, under each storing temperature value T, whether the capacitance loss rate Q of battery sample exceedes default loss-rate threshold, if the capacitance loss rate Q of battery sample exceedes loss-rate threshold under some storing temperature value T, then t sample time under this storing temperature value T is carried out revising and obtain new sampling spot, and the battery sample storing full state of charge is placed based on new sampling spot, obtain the capacitance loss rate Q of battery sample under new sampling spot; It should be noted that, the step of aforesaid revision t sample time is not necessary step, such as, when storage temperature range value is reasonable, then can not need revision t sample time completely.

In the present embodiment, the capacitance loss rate Q storing 30 days battery samples under 343K is about 18%, close to battery life terminal (20%), therefore needs to revise t sample time under 343K, t sample time under revised 343K is respectively 1/2 week, 1 week, 2 weeks, 3 weeks ... (namely 3 days, 7 days, 14 days ...), after therefore the present embodiment is revised t sample time under 343K, the final test scheme of each sampling spot is as shown in table 2 in fact.

Table 2: the final test scheme of each sampling spot of embodiment one.

2) set up capacity attenuation Ageing Model shown in formula (1), experimental error is considered on the basis of described experimental data, and utilize the value of capacity attenuation Ageing Model Confirming model parameter ρ shown in formula (1); On the basis of the value of Confirming model parameter ρ, utilize the nonlinear curve function of origin software to carry out matching to capacity attenuation Ageing Model formula (1) Suo Shi, obtain the value of simulation curve under each storing temperature value T and model parameter a;

$\left\{\begin{array}{c}Q={\mathrm{at}}^{\mathrm{\ρ}}\\ a=A-B/T\end{array}\right.---\left(1\right)$

In formula (1), Q represents the capacitance loss rate of battery sample, and T represents storing temperature value, and t represents sample time, and a, ρ, A, B are model parameter to be solved;

In the present embodiment, by considering that experimental error utilizes capacity attenuation Ageing Model shown in formula (1) to obtain the value of model parameter ρ between 0.47 ~ 0.55, for simplifying subsequent calculations, the value of delivery shape parameter ρ is 0.5 (ρ=0.5).After getting ρ=0.5, the simulation curve under utilize the nonlinear curve function of origin software to carry out different temperatures that matching obtains as shown in Figure 2.See Fig. 2, curve is respectively the curve value of battery at the storage different time capacitance loss rate Q of 298K, 313K, 323K, 333K, 343K from the bottom to top, and on curve, data point is respectively the capacitance loss rate Q of sampling spot reality.Right part a1 ~ a5 represents value and the error thereof of the model parameter a at 298K, 313K, 323K, 333K, 343K temperature respectively, R ²represent the coefficient of determination.

3) whether meet based on the model parameter a under each storing temperature value T the selection rationality that Arrhenius formula judges storage temperature range, if rationally, determine capacitance loss rate Q in experimental data, storing temperature value T, sample time t error range and redirect performs step 4); If unreasonable, then cast out the number of sampling certificate of a highest storing temperature value T from experimental data, redirect performs step 2).

4) by capacitance loss rate Q, storing temperature value T in experimental data, sample time t and model parameter ρ value substitute into the value that formula (2) illustrated equation obtains model parameter A, B, and utilize MonteCarlo method based on capacitance loss rate Q, storing temperature value T in described experimental data, sample time t error range draw the value of n group model parameter A, B, the value of n group model parameter A, B is added up to the average and standard deviation that draw model parameter A, B;

x＝(R ^TR) ^-1R ^TQ(2)

In formula (2), R ^tthe transposed matrix of representing matrix R, the expression formula of matrix Q is such as formula shown in (3), and the expression formula of matrix R is such as formula shown in (4), and the expression formula of matrix x is such as formula shown in (5);

$Q=\left(\begin{array}{c}{\mathrm{lnQ}}_{11}\\ \·\\ \·\\ \·\\ {\mathrm{lnQ}}_{mn}\end{array}\right)---\left(3\right)$

In formula (3), Q ₁₁represent first storing temperature value T ₁₁lower first sample time t ₁₁corresponding capacitance loss rate, Q _mnrepresent m storing temperature value T _mnlower n-th sample time t _mncorresponding capacitance loss rate;

$R=\left(\begin{array}{ccc}1& 1/T_{11}& {\mathrm{lnt}}_{11}\\ M& O& M\\ 1& 1/T_{mn}& {\mathrm{lnt}}_{mn}\end{array}\right)---\left(4\right)$

In formula (4), T ₁₁represent first storing temperature value, T _mnrepresent m storing temperature value, t ₁₁represent first storing temperature value T ₁₁under first sample time t ₁₁, t _mnrepresent m storing temperature value T _mnunder the n-th sample time;

$x=\left(\begin{array}{c}A\\ B\\ \mathrm{\ρ}\end{array}\right)---\left(5\right)$

In formula (5), the matrix in x expression (2), ρ, A, B are model parameter to be solved.

In the present embodiment, the average of model parameter A, B and standard deviation A=11.5 (0.76), B=3605 (243.7), the numerical value before its bracket is average, and the numerical value in bracket is the standard deviation of parameter.

5) average of the value of model parameter ρ and model parameter A, B is substituted into capacity attenuation Ageing Model shown in formula (1), and utilize origin curve to carry out curve fitting, the curve obtained and real data comparison diagram as shown in Figure 3, judge that matching obtains the coefficient of determination R of curve ²whether be greater than default threshold value, if be greater than default threshold value, judge that the goodness of fit meets the demands, redirect performs step 6); Otherwise judge that the goodness of fit does not meet the demands, adjustment capacitance loss rate Q, storing temperature value T, sample time t error range, redirect performs step 4).

The threshold value preset in the present embodiment is 0.9, if R ²>0.9, then think that matching degree can accept, if R ²<0.9 then thinks that fitting degree is unacceptable, then need to jump to step 4), so as expand or reduction capacitance loss rate Q, storing temperature value T, sample time t error range basis on again calculate the value of A, B.In the present embodiment, matching obtains the coefficient of determination R of curve ²≈ 0.97, therefore judge that the goodness of fit meets the demands, redirect performs step 7).

6) according to the interval of the average of described model parameter A, B and standard deviation Confirming model parameter A, B, in described interval, the value of model parameter A, B is solved confidence interval of mean mode under the degree of confidence preset according to known standard deviation and carries out random value, obtain the value of many group models parameter A, B; The value of the value of model parameter ρ and described many group models parameter A, B substituted into capacity attenuation Ageing Model shown in formula (1) by the method for MonteCarlo to predict the life-span of lithium ion battery to be evaluated under normal temperature (298K) (the present embodiment with capacitance loss 20% for end of life), obtain the life-span distribution plan of lithium ion battery to be evaluated as shown in Figure 4.

As can be seen from the statistics of Fig. 4, the life-span distributed area of battery is 2.5-4.8, and in 1000 MonteCarlo method simulations, about has the life-span of the battery calculated for 100 times below 3 years, have the life-span calculated for 900 times more than 3 years.This just illustrates that the confidence level of battery life more than 3 years is 90%.In conjunction with the curve of life-span distribution plan, according in real life to the evaluation of battery life, can think that the mean lifetime of battery is about 3.5 years, namely the evaluation result of the storage life of this lithium ion battery under full charge condition is 3.5 years.

In the present embodiment, step 1) in specifically refer to the capacitance loss rate Q obtaining battery sample under each sampling spot according to formula (6);

Q＝1-Q ₁/Q ₀(6)

In formula (6), Q is the capacitance loss rate of battery sample under some sampling spots, Q ₀represent the capacity of full electric charge battery sample, Q ₁represent the capacity of battery sample when sampling spot.

In the present embodiment, step 1) in corresponding three battery samples of each sampling spot, step 2) after also comprise data reject and mend survey step, detailed step comprises: for three capacitance loss rate Q of three battery samples of each sampling spot, first maximal value is wherein obtained, minimum value and intermediate value, then by intermediate value respectively and maximal value, minimum value compares, if the error of intermediate value and mxm., intermediate value and the error with minimum value are all in predetermined threshold value (in the present embodiment, concrete value is 5%), be then maximal value by the capacitance loss rate Q value of this sampling spot, minimum value, the average of intermediate value three, if only the error of maximal value and intermediate value exceedes predetermined threshold value in maximal value, minimum value, then rejecting maximal value, is intermediate value and the average with minimum value by the capacitance loss rate Q value of this sampling spot, if only the error of minimum value and intermediate value exceedes predetermined threshold value in maximal value, minimum value, then rejecting minimum value, is intermediate value and the average with maximal value by the capacitance loss rate Q value of this sampling spot, if the error of maximal value, both minimum value and intermediate value all exceedes predetermined threshold value, then reappose battery sample for this sampling spot and mend and survey capacitance loss rate Q.Because the consistance of the preparation of battery may go wrong, and consistency problem may be extended under the high temperature conditions, therefore need to carry out the rejecting to experimental result appropriateness and benefit survey, effectively can eliminate the accidental error that the consistance inequality prepared due to battery produces, the accuracy of experimental data can be improved.

In the present embodiment, described step 4) detailed step as follows:

4.1) for the value of the variable a under each storing temperature value T in experimental data, utilize origin linear curve fit lna value to 1/T mapping matched curve (as shown in Figure 3);

4.2) coefficient of determination R of curve is obtained according to matching ²whether judge that matching obtains curve is straight line, if it is straight line that matching obtains curve, then judge that each storing temperature value T drag parameter a meets Arrhenius formula, the selection of storage temperature range is reasonable, determine capacitance loss rate Q in experimental data, storing temperature value T, sample time t error range and redirect performs step 5); If it is not straight line that matching obtains curve, judge that each storing temperature value T drag parameter a does not meet Arrhenius formula, the selection of storage temperature range is unreasonable, casts out the number of sampling certificate of a highest storing temperature value T from experimental data, and redirect performs step 3).See data point in Fig. 5, Fig. 5 be respectively utilize different temperatures under the lna value that calculates of a value obtained and 1/T value, utilize origin linear curve fit lna value to 1/T mapping in the present embodiment, the coefficient of determination R of matched curve ²=0.99, this illustrative graph is straight line in whole temperature range lna value to 1/T, therefore can judge that a value and temperature meet Arrhenius formula, therefore this sample meets Arrhenius formula in this experimental temperature scope (298K-343K), namely in whole temperature range, energy of activation is consistent.

In the present embodiment, described step 5) in n group model parameter A, B value in n value be greater than 1000.The number of times that n value in the value of n group model parameter A, B is simulated for MonteCarlo method, when n value is greater than 1000, the data of MonteCarlo method simulation make structure more accurate.

In the present embodiment, described step 5) in the function expression that adopts when the average drawing model parameter A, B is added up to the value of n group model parameter A, B specifically such as formula shown in (7);

$\stackrel{\&OverBar;}{X}=\frac{1}{n}\underset{i=1}{\overset{n}{\&Sigma;}}{X}_i---\left(7\right)$

In formula (7), represent the average calculated, X _irepresent i-th value in n group model parameter A or B, n represents the quantity of model parameter A or B.

Preferably, described step 5) in the function expression that adopts when the standard deviation drawing model parameter A, B is added up to the value of n group model parameter A, B specifically such as formula shown in (8);

$S=\sqrt{\frac{1}{n}\underset{i=1}{\overset{n}{\&Sigma;}}{(X_i-\stackrel{\&OverBar;}{X})}^2}---\left(8\right)$

In formula (8), S represents the standard deviation calculated, X _irepresent i-th value in n group model parameter A or B, n represents the quantity of model parameter A or B, represent the average of n group model parameter A or B.

In the present embodiment, described step 7) in by the value of model parameter A, B according to known standard deviation solve preset degree of confidence under confidence interval of mean mode carry out random value time, the fiducial interval of random value is specifically such as formula shown in (9);

$\&lsqb;\stackrel{\&OverBar;}{X}-{\mathrm{\&mu;}}_{1-\mathrm{\&alpha;}/2}\frac{S}{\sqrt{n}},\stackrel{\&OverBar;}{X}+{\mathrm{\&mu;}}_{1-\mathrm{\&alpha;}/2}\frac{S}{\sqrt{n}}\&rsqb;---\left(9\right)$

In formula (9), represent the average of n group model parameter A or B, S represents the standard deviation of n group model parameter A or B, and n represents the quantity of model parameter A or B, and μ is fractile, and 1-α is fiducial interval.In the present embodiment, the fractile μ value that the α in fiducial interval is corresponding is specifically as shown in table 3.

Table 3:, the fractile μ value table that the α in fiducial interval is corresponding.

α	0.90	0.95	0.975	0.99	0.995	0.999
							μ	1.282	1.646	1.960	2.326	2.576	3.090

Embodiment two:

The present embodiment is identical with the basic step of embodiment one, and its main difference point is:

In the present embodiment step 1) in, in view of the experiment starting stage finds that this example battery capacity attenuation under 343K is too fast, the highest acceleration temperature 333K that this example is taked, getting interval temperature range is 5K, get four interval temperature ranges, therefore storing temperature is set as 298K, 318K, 323K, 328K, 333K, and 15 days will be set to sample time, 30 days, 45 days, 60 days, 90 days.Simultaneously, because the capacity difference of the cell of example 2 battery is larger, therefore increase sampling number to 5, the final test scheme finally obtaining each sampling spot is as shown in table 4, in addition in the present embodiment also to capacitance loss rate Q, storing temperature value T in experimental data, sample time t error range expand, equally also excessive to battery error point carries out suitable deletion and mends surveying.

Table 4: the final test scheme of each sampling spot of embodiment two.

In the present embodiment step 2) in, by considering that experimental error utilizes model to obtain the scope of ρ value between 0.42 ~ 0.57, for simplifying subsequent calculations, get ρ=0.5 equally.Meanwhile, after getting ρ=0.5, obtain simulation curve at different temperatures and a value.Step 2) in utilize the nonlinear curve function matching of origin software to obtain curve as shown in Figure 6.See Fig. 6, curve is respectively the storage different time capacity attenuation curve value of battery at 298K, 318K, 323K, 328K, 333K from the bottom to top, on curve, data point is respectively actual capacity attenuation mean value, right part a1 ~ a5 represents a value and the error thereof of 298K, 318K, 323K, 328K, 333K respectively, R ²represent the coefficient of determination.

In the present embodiment step 3) in, after determining ρ=0.5, utilize origin linear curve fit lna value to map the curve that obtains of matched curve as shown in Figure 7 to 1/T.Data results shows the coefficient of determination R of the curve of matching ²=0.9963, this illustrative graph is straight line in whole temperature range lna value to 1/T, namely a value and temperature meet Arrhenius formula, therefore this sample meets Arrhenius formula in this experimental temperature scope (298K-333K), and namely in whole temperature range, energy of activation is consistent.Wherein in Fig. 7 data point be respectively utilize different temperatures under the relation of lna and 1/T that calculates of a value obtained.

In the present embodiment step 4) in, the average of model parameter A, B and standard deviation A=4.85 (0.53), B=1483 (138.7), the numerical value before its bracket is average, and the numerical value in bracket is the standard deviation of parameter.

In the present embodiment step 5) in, the average of the value 0.5 of model parameter ρ and model parameter A, B is substituted into capacity attenuation Ageing Model shown in formula (1), and the relation of the curve utilizing origin curve to carry out curve fitting to obtain and actual data point as shown in Figure 8.In Fig. 8, the data of the actual acquisition in data point position and error range thereof, curve is the curve that matching obtains, and formula is the fitting formula of whole temperature range, and by the coefficient of determination (R can be known after the matching of Origin software ²) ≈ 0.96, this illustrative graph fitting degree is fine.

In the present embodiment step 6) in, the life-span distribution plan of lithium ion battery to be evaluated is obtained as shown in Figure 9 based on 1000 MonteCarlo method simulations, as can be seen from the statistics of Fig. 9, the life-span distributed area of battery is 2.2 ~ 2.7 years, and in 1000 MonteCarlo method simulations, about there is the life-span of the battery calculated for 100 times below 2.3 years, have the life-span calculated for 900 times more than 2.3 years.This just illustrates that the confidence level of battery life more than 2.3 years is 90%.The mean lifetime of battery is about 2.5 years.In conjunction with life-span distribution curve, according in real life to the evaluation of battery life, can think that the mean lifetime of battery is about 2.5 years, namely the evaluation result of the storage life of this kind of lithium ion battery under full charge condition is 2.5 years.

The above is only the preferred embodiment of the present invention, protection scope of the present invention be not only confined to above-described embodiment, and all technical schemes belonged under thinking of the present invention all belong to protection scope of the present invention.It should be pointed out that for those skilled in the art, some improvements and modifications without departing from the principles of the present invention, these improvements and modifications also should be considered as protection scope of the present invention.

Claims (9)

1. the full charge storage life assessment method of lithium ion battery, is characterized in that step comprises:

1) according to storage minimum temperature and the electrolyte decomposition temperature determination storage temperature range of lithium ion battery to be evaluated, multiple storing temperature value T is chosen according to the temperature interval of presetting in storage temperature range, the battery sample evaluating the full state of charge of lithium ion battery is stored t sample time specified under multiple storing temperature value T, under each storing temperature value T each sample time t as a sampling spot, each sampling spot at least one battery sample corresponding, obtains the capacitance loss rate Q of battery sample under each sampling spot and generates experimental data;

2) set up capacity attenuation Ageing Model shown in formula (1), experimental error is considered on the basis of described experimental data, and utilize the value of capacity attenuation Ageing Model Confirming model parameter ρ shown in formula (1); On the basis of the value of Confirming model parameter ρ, utilize the nonlinear curve function of origin software to carry out matching to capacity attenuation Ageing Model formula (1) Suo Shi, obtain the value of simulation curve under each storing temperature value T and model parameter a;

In formula (1), Q represents the capacitance loss rate of battery sample, and T represents storing temperature value, and t represents sample time, and a, ρ, A, B are model parameter to be solved;

3) whether meet based on the model parameter a under each storing temperature value T the selection rationality that Arrhenius formula judges storage temperature range, if rationally, determine capacitance loss rate Q in experimental data, storing temperature value T, sample time t error range and redirect performs step 4); If unreasonable, then cast out the number of sampling certificate of a highest storing temperature value T from experimental data, redirect performs step 2);

4) by capacitance loss rate Q, storing temperature value T in experimental data, sample time t and model parameter ρ value substitute into the value that formula (2) illustrated equation obtains model parameter A, B, and utilize MonteCarlo method based on capacitance loss rate Q, storing temperature value T in described experimental data, sample time t error range draw the value of n group model parameter A, B, the value of n group model parameter A, B is added up to the average and standard deviation that draw model parameter A, B;

x＝(R ^TR) ^-1R ^TQ(2)

In formula (2), R ^tthe transposed matrix of representing matrix R, the expression formula of matrix Q is such as formula shown in (3), and the expression formula of matrix R is such as formula shown in (4), and the expression formula of matrix x is such as formula shown in (5);

In formula (3), Q ₁₁represent first storing temperature value T ₁₁lower first sample time t ₁₁corresponding capacitance loss rate, Q _mnrepresent m storing temperature value T _mnlower n-th sample time t _mncorresponding capacitance loss rate;

In formula (4), T ₁₁represent first storing temperature value, T _mnrepresent m storing temperature value, t ₁₁represent first storing temperature value T ₁₁under first sample time t ₁₁, t _mnrepresent m storing temperature value T _mnunder the n-th sample time;

In formula (5), the matrix in x expression (2), ρ, A, B are model parameter to be solved;

5) average of the value of model parameter ρ and model parameter A, B is substituted into capacity attenuation Ageing Model shown in formula (1), and utilize origin curve to carry out curve fitting, judge that matching obtains the coefficient of determination R of curve ²whether be greater than default threshold value, if be greater than default threshold value, judge that the goodness of fit meets the demands, redirect performs step 6); Otherwise judge that the goodness of fit does not meet the demands, adjustment capacitance loss rate Q, storing temperature value T, sample time t error range, redirect performs step 4);

6) according to the interval of the average of described model parameter A, B and standard deviation Confirming model parameter A, B, in described interval, the value of model parameter A, B is solved confidence interval of mean mode under the degree of confidence preset according to known standard deviation and carries out random value, obtain the value of many group models parameter A, B; By the method for MonteCarlo, the value of the value of model parameter ρ and described many group models parameter A, B is substituted into capacity attenuation Ageing Model shown in formula (1) to predict the lithium ion battery to be evaluated life-span at normal temperatures, obtain the life-span distribution plan of lithium ion battery to be evaluated.

2. the full charge storage life assessment method of lithium ion battery according to claim 1, it is characterized in that, described step 1) after also comprise revision sample time t step, concrete steps comprise: in judgment experiment data, under each storing temperature value T, whether the capacitance loss rate Q of battery sample exceedes default loss-rate threshold, if the capacitance loss rate Q of battery sample exceedes loss-rate threshold under some storing temperature value T, then t sample time under this storing temperature value T is carried out revising and obtain new sampling spot, and the battery sample storing full state of charge is placed based on new sampling spot, obtain the capacitance loss rate Q of battery sample under new sampling spot.

3. the full charge storage life assessment method of lithium ion battery according to claim 2, is characterized in that, described step 1) in the calculation expression of capacitance loss rate Q such as formula shown in (6);

Q＝1-Q ₁/Q ₀(6)

In formula (6), Q is the capacitance loss rate of battery sample under some sampling spots, Q ₀represent the capacity of full electric charge battery sample, Q ₁represent the capacity of battery sample when sampling spot.

4. the full charge storage life assessment method of lithium ion battery according to claim 3, it is characterized in that, described step 1) in corresponding three battery samples of each sampling spot, and after generation experimental data, also comprise the step that data are rejected and benefit is surveyed, detailed step comprises: for three capacitance loss rate Q of three battery samples of each sampling spot, first maximal value is wherein obtained, minimum value and intermediate value, then by intermediate value respectively and maximal value, minimum value compares, if the error of intermediate value and mxm., intermediate value and the error with minimum value are all in predetermined threshold value, be then maximal value by the capacitance loss rate Q value of this sampling spot, minimum value, the average of intermediate value three, if only the error of maximal value and intermediate value exceedes predetermined threshold value in maximal value, minimum value, then rejecting maximal value, is intermediate value and the average with minimum value by the capacitance loss rate Q value of this sampling spot, if only the error of minimum value and intermediate value exceedes predetermined threshold value in maximal value, minimum value, then rejecting minimum value, is intermediate value and the average with maximal value by the capacitance loss rate Q value of this sampling spot, if the error of maximal value, both minimum value and intermediate value all exceedes predetermined threshold value, then reappose battery sample for this sampling spot and mend and survey capacitance loss rate Q.

5. the full charge storage life assessment method of the lithium ion battery according to claim 1 or 2 or 3 or 4, is characterized in that, described step 3) detailed step as follows:

3.1) for the value of the variable a under each storing temperature value T in experimental data, utilize origin linear curve fit lna value to 1/T mapping matched curve;

3.2) coefficient of determination R of curve is obtained according to matching ²whether judge that matching obtains curve is straight line, if it is straight line that matching obtains curve, then judge that each storing temperature value T drag parameter a meets Arrhenius formula, the selection of storage temperature range is reasonable, determine capacitance loss rate Q in experimental data, storing temperature value T, sample time t error range and redirect performs step 4); If it is not straight line that matching obtains curve, judge that each storing temperature value T drag parameter a does not meet Arrhenius formula, the selection of storage temperature range is unreasonable, casts out the number of sampling certificate of a highest storing temperature value T from experimental data, and redirect performs step 2).

6. the full charge storage life assessment method of lithium ion battery according to claim 5, is characterized in that: described step 4) in n group model parameter A, B value in n value be greater than 1000.

7. the full charge storage life assessment method of lithium ion battery according to claim 6, is characterized in that: described step 4) in the function expression that adopts when the average drawing model parameter A, B is added up to the value of n group model parameter A, B specifically such as formula shown in (7);

In formula (7), represent the average calculated, X _irepresent i-th value in n group model parameter A or B, n represents the quantity of model parameter A or B.

8. the full charge storage life assessment method of lithium ion battery according to claim 7, is characterized in that: described step 4) in the function expression that adopts when the standard deviation drawing model parameter A, B is added up to the value of n group model parameter A, B specifically such as formula shown in (8);

In formula (8), S represents the standard deviation calculated, X _irepresent i-th value in n group model parameter A or B, n represents the quantity of model parameter A or B, represent the average of n group model parameter A or B.

9. the full charge storage life assessment method of lithium ion battery according to claim 8, it is characterized in that, described step 6) in by the value of model parameter A, B according to known standard deviation solve preset degree of confidence under confidence interval of mean mode carry out random value time, the fiducial interval of random value is specifically such as formula shown in (9);

In formula (9), represent the average of n group model parameter A or B, S represents the standard deviation of n group model parameter A or B, and n represents the quantity of model parameter A or B, and μ is fractile, and 1-α is fiducial interval.