CN105759216A - Method for estimating state of charge of soft package lithium-ion battery - Google Patents

Abstract

The invention relates to a method for estimating the state of charge of a soft package lithium-ion battery. The method comprises a step S1 of acquiring battery surface dynamic stress and a battery working state signal that includes first data for indicating that a battery is in a charging, standing or discharged state; a step S2 of determining whether the battery is in a standing state, if so, executing a step S3, and otherwise, executing a step S4; the step S3 of taking the previous estimation result as the current state of charge of the lithium-ion battery; the step S4 of obtaining battery surface static stress according to the battery surface dynamic stress, and executing a step S5; and the step S5 of estimating the state of charge of the lithium-ion battery by combining a corresponding function of the static stress and the state of charge according to the obtained battery surface static stress. Compared with the prior art, the invention achieves the estimation of the charge of charge of the lithium-ion battery by using a method based on stress measurement, and reduces the system complexity while improving the system accuracy.

Description

A kind of soft bag lithium ionic cell charge state estimation method

Technical field

The present invention relates to a kind of charge states of lithium ion battery method of estimation, especially relate to a kind of soft bag lithium ionic cell charge state estimation method.

Background technology

Traditional battery management system (BatteryManagementSystem, BMS) to inside battery state, when particularly SOC being estimated, it is substantially all the method being based on lithium ion battery open-circuit voltage OCV or battery equivalent circuit, actually, there is a congenital drawback in the method, that is, voltage platform district estimates inaccurate, is particularly acute especially for ferric phosphate lithium cell；Some new methods at present, such as fuzzy logic, artificial neural network and support vector machine are also applied in the estimation of SOC gradually, but above method needs substantial amounts of data supporting, and model is complicated.

In sum, battery SOC method of estimation used in current BMS system is not accuracy is exactly excessively complicated not, and cost is high.Therefore, it is necessary to provide a kind of SOC method of estimation based on battery surface stress measurement under current BMS system architecture.

After 2000, lithium ion battery stress studies is subject to people's attention gradually, inventor have found that lithium ion battery surface stress is main relevant with the embedding lithium of electrode, the embedding lithium degree of electrode then can characterize battery charge state, through deep experiment measuring and theory analysis, inventor have found that soft bag lithium ionic cell surface stress and SOC exist relation one to one, and curve monotonicity is fine, discharge and recharge hysteresis is inconspicuous, and platform area discharge and recharge stress curve essentially coincides, if this discovery is applied in charge states of lithium ion battery estimation, then can solve the method platform area such as equivalent circuit and estimate inaccurate drawback.

Summary of the invention

Defect that the purpose of the present invention is contemplated to overcome above-mentioned prior art to exist and a kind of soft bag lithium ionic cell charge state estimation method is provided.

The purpose of the present invention can be achieved through the following technical solutions:

A kind of soft bag lithium ionic cell charge state estimation method, including step:

S1: gather battery surface dynamic stress and cell operating status signal, wherein, described cell operating status signal includes for indicating battery to be in the first data of charging, standing or discharge condition；

S2: judge whether battery is in static condition, if it has, then perform step S3, if it has not, then perform step S4；

S3: once estimate the result state-of-charge as current lithium ion battery using front；

S4: according to battery surface dynamic stress, obtains battery surface static stress, and performs step S5；

S5: according to the battery surface static stress obtained, in conjunction with static stress and the respective function of state-of-charge, estimation obtains the state-of-charge of lithium ion battery.

The present invention is characterized by based on the one-to-one relationship between battery surface stress and the state-of-charge of lithium ion battery, due to the collection difficulty of static stress, adopts the mode by gathering battery surface dynamic stress, estimates the state-of-charge of lithium ion battery.

Described step S4 specifically includes step:

S41: judge whether battery is in charged state, if it has, then perform step S42, if it has not, then perform step S43；

S42: according to the battery surface dynamic stress collected and trickle charge time, it is thus achieved that battery surface static stress:

DS-S=SS

Wherein: DS is battery surface dynamic stress, S is stress decay value, and SS is battery surface static stress；

S43: using battery surface dynamic stress as battery surface static stress.

Described cell operating status signal also includes the second data for recording the battery trickle charge time,

Stress decay value in described step S42 obtains according to stress decay model, described stress decay model particularly as follows:

$S=y_0+{\mathrm{Ae}}^{-\frac{t_c}{\mathrm{\τ}}}$

Wherein: y₀For primary stress, A is attenuation amplitude, and τ is attenuation constant, t_cFor the battery trickle charge time.

When battery starts to charge up each time, the battery trickle charge time is recalculated.

Described step S42 specifically includes step:

S421: estimate the state-of-charge of lithium ion battery according to battery surface dynamic stress；

S422: select the primary stress of correspondence, attenuation amplitude and attenuation constant according to estimation results, and substitute into stress decay model；

S423: obtain stress decay value according to stress decay model and battery trickle charge time；

S424: obtain battery surface static stress according to stress decay value.

In described step S5, the respective function of stress and state-of-charge is specially the respective function of averaged static stress and state-of-charge, and wherein said averaged static stress is the meansigma methods of static stress during charge and discharge under same state-of-charge,

Step S5 is particularly as follows: using the battery surface static stress that obtains as averaged static stress, the respective function according to stress Yu state-of-charge, estimation obtains the state-of-charge of lithium ion battery.

Compared with prior art, the invention have the advantages that

1) present invention adopts the method based on stress measurement to realize charge states of lithium ion battery, reduces the complexity of system while improving system accuracies.

2) outbound data collection of the present invention is only limitted to surface stress and cell operating status information (status and trickle charge time), it is necessary to data volume less, model amount of calculation is less.

3) state-of-charge of present battery is directly calculated according to stress, all unrelated with the historic state of battery and error, do not result in the accumulation of error.

4) when charging, it is contemplated that stress decay value, it is possible to improve precision.

5) first estimate state-of-charge according to dynamic stress, select the primary stress of correspondence, attenuation amplitude and attenuation constant to substitute into attenuation model further according to state-of-charge, improve estimation precision.

Accompanying drawing explanation

Fig. 1 is the key step schematic flow sheet of the inventive method；

Fig. 2 is the relation schematic diagram of battery dynamic stress and the static stress measuring gained by experiment；

The schematic diagram of Fig. 3 stress decay model application；

Fig. 4 attenuation amplitude and matching figure line thereof；

Fig. 5 static stress and matched curve thereof；

Fig. 6 attenuation constant and matched curve thereof；

Fig. 7 is graph of a relation and the averaged static stress curve of the surveyed battery static stress of experiment and battery SOC；

Fig. 8 is the detailed process schematic diagram of the present invention.

Detailed description of the invention

Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.The present embodiment is carried out premised on technical solution of the present invention, gives detailed embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment.

The application is before application, need to measure the one-to-one relationship obtaining between battery surface static stress and the state-of-charge of lithium ion battery in advance, this one-to-one relationship can adopt experiment to determine, the battery of same type (especially same model), corresponding relation between its static stress and state-of-charge is basically identical, obtain one-to-one relationship, namely obtain static stress and the respective function of electricity condition, this function can simulate a mathematic(al) representation and store, the form that can also adopt synopsis stores, how this corresponding relation concrete obtains, the precision of respective function how, it is not belonging to the category that the application discusses, therefore the application no longer describes in detail, the application only requires, if mode records by experiment, the data set up need to come from the experiments of measuring in cell single loop.

A kind of soft bag lithium ionic cell charge state estimation method, as it is shown in figure 1, include step:

S1: gather battery surface dynamic stress and cell operating status signal, wherein, cell operating status signal includes for indicating battery to be in the first data of charging, standing or discharge condition, measurement difficulty due to battery surface static stress, and dynamic stress can be by what pressure transducer recorded, therefore the application selects to measure battery surface dynamic stress.

S2: judge whether battery is in static condition, if it has, then perform step S3, if it has not, then perform step S4；

S3: once estimate the result state-of-charge as current lithium ion battery using front；

S4: according to battery surface dynamic stress, obtains battery surface static stress, and performs step S5；

Step S4 specifically includes step:

S41: judge whether battery is in charged state, if it has, then perform step S42, if it has not, then perform step S43；

S42: according to the battery surface dynamic stress collected and trickle charge time, it is thus achieved that battery surface static stress:

DS-S=SS

Wherein: DS is battery surface dynamic stress, S is stress decay value, and SS is battery surface static stress；

It is illustrated in figure 2 dynamic stress (DynamicStress, SD), the instantaneous stress on the lithium battery surface namely recorded by pressure transducer.Static stress (StaticStress, SS) is the stress that dynamic stress obtains after sufficient standing, can be generally considered as standing 2 hours, and stress no longer recovers, and stress now is static stress.Stress decay is that after lithium ion battery charge or discharge process stops, the phenomenon recovered occurs stress.There is following relation in three kinds of stress:

SS=DS-S

Battery carrying out intermittent charge and can obtain the static stress change curve with SOC of battery, namely often charging 5%SOC stands 2h, until reaching charge cutoff voltage；Discharge process is also such.As shown below, the figure line of square filled marks is the dynamic stress in charging process, and the figure line of black triangle labelling is the static stress in charging process；The figure line of open triangular markers is the dynamic stress in discharge process, and the figure line of hollow circular labelling is the static stress in discharge process；As shown in Figure 2, the static stress of charging process is significantly less than dynamic stress, and two curves of discharge process almost overlap.This is owing to the stress decay characteristic of charge and discharge process causes, and stress decay during charging is obvious, and almost without stress decay when discharging.

Cell operating status signal also includes the second data for recording the battery trickle charge time, and when battery starts to charge up each time, the battery trickle charge time is recalculated.

Step S42 specifically includes step:

S421: estimate the state-of-charge of lithium ion battery according to battery surface dynamic stress；

S422: select corresponding primary stress, attenuation amplitude and attenuation constant according to estimation results, and substitute into stress decay model and try to achieve stress decay value, stress decay model particularly as follows:

$S=y_0+{\mathrm{Ae}}^{-\frac{t_c}{\mathrm{\τ}}}$

Wherein: y₀For primary stress, A is attenuation amplitude, and τ is attenuation constant, t_cFor the battery trickle charge time.

Owing to the primary stress in stress decay model under different state-of-charges, attenuation amplitude and attenuation constant can be otherwise varied, so the application have employed for different state-of-charges interval, it is proposed that the mode of the parameter group of one group of stress decay model.

Owing to stress decay is relevant with the trickle charge time, therefore the application adds the second data in cell operating status signal, is used for recording the battery trickle charge time, to improve precision.Additionally, due to stress decay model is relevant with SOC (state-of-charge, lower same), namely stress decay is the function of SOC, but when setting up SOC and estimating model, SOC is the amount of being estimated, namely unknown quantity, thus in stress decay model can not directly using SOC as input quantity.So it is accomplished by SOC is carried out pre-estimation, because SOC and charging stress exist dull functional relationship, therefore can use dynamic stress (replacing charging stress) that SOC is carried out pre-estimation.By the SOC that pre-estimation obtains, determine that SOC is interval, SOC be divided into multiple interval by the application, then with each interval stress decay fitting formula corresponding for right endpoint SOC come in this interval of approximate evaluation stress decay a little, the method of the approximate substitution owing to using, so the SOC interval divided is more many, model is more accurate, and consequent error is more little, the problem considering experimental period, being divided into ten intervals in the present embodiment, every 10% as an interval, specifically as shown in table 1

Table 1

SOC	Dynamic stress (kg)	SOC	Dynamic stress (kg)
				0～0.1	0.00～0.99	0.5～0.6	3.72～4.09
0.1～0.2	0.99～1.95	0.6～0.7	4.09～4.57
				0.2～0.3	1.95～2.86	0.7～0.8	4.57～5.11
0.3～0.4	2.86～3.42	0.8～0.9	5.11～5.75
				0.4～0.5	3.42～3.72	0.9～1.0	5.75～6.12

Then different stress decay fitting formula counter stress decay interval for SOC is used to be calculated, as shown in Figure 3.

With (0.5,0.6] interval for example, the Maxwell stress decay model according to lithium ion battery, utilize PolynomialFit instrument that this curve is fitted in Origin,

As shown in Figure 4, along with the growth of trickle charge time, attenuation amplitude also constantly increases.Both functional relationships can use quadratic polynomial matching, goodness of fit R²0.998 can be reached.Write out attenuation amplitude A and trickle charge time t_cBetween function expression: A=0.01503+0.00849t_c-9.43353×10^-5t_c ²

Shown in Fig. 5, along with the growth of trickle charge time, y₀Constantly increasing, both functional relationships can use cubic polynomial matching, goodness of fit R²0.999 can be reached.Primary stress y can be write out₀With trickle charge time t_cBetween function expression:

y₀=4.1707-0.00408t_c-2.6539×10^-4t_c ²+6.43776×10^-6t_c ³

Fig. 6 is attenuation constant τ and matched curve thereof.As seen from the figure, along with the growth of trickle charge time, attenuation constant constantly increases, and both functional relationships can use equation of linear regression matching, goodness of fit R²0.965 can be reached.Attenuation constant τ and trickle charge time t can be write out_cBetween function expression:

τ=707.7+14.19376t_c

Utilize identical experimental technique, record at ten SOC point places of 10%～100%SOC respectively, the stress decay process that the different trickle charge times are corresponding, and in Origin software, each stress decay process is carried out curve fitting, draw stress decay model corresponding to different SOC point place difference trickle charge time, then extract the parameter in each stress decay model, make these parameters (A, y₀, τ) with trickle charge time t_cImage and carry out curve fitting.

The stress decay Estimating The Model Coefficients result at different SOC point place is listed below:

10%:

$\left\{\begin{array}{c}0.060017\\ y0=0.855567\\ \mathrm{\τ}=1502.643\end{array}\right.$

20%:

$\left\{\begin{array}{c}A=0.00528+0.02087x-6.02143\×{10}^{-4}{x}^2\\ y_0=1.720002\\ \mathrm{\τ}=1150.272\end{array}\right.$

30%:

$\left\{\begin{array}{c}A=0.00331+0.01124x-1.1506\×{10}^{-4}{x}^2\\ y_0=2.66753-0.00838x-3.48413\×{10}^{-4}{x}^2+1.44753\×{10}^{-5}{x}^3\\ \mathrm{\τ}=1317.487\end{array}\right.$

40%::

$\left\{\begin{array}{c}A=0.02023+0.00811x-6.92522\×{10}^{-5}{x}^2\\ y_0=3.40641-0.00993x+3.82614\×{10}^{-5}{x}^2+2.25839\×{10}^{-6}{x}^3\\ \mathrm{\τ}=993.45+21.49173x\end{array}\right.$

50%:

$\left\{\begin{array}{c}A=0.01031+0.0105x-1.51042\×{10}^{-4}{x}^2\\ y_0=3.86794-0.0171x+4.86389\×{10}^{-4}{x}^2-5.53241\×{10}^{-6}{x}^3\\ \mathrm{\τ}=846.202+15.95467x\end{array}\right.$

60%:

$\left\{\begin{array}{c}A=0.01503+0.00849x-9.43353\×{10}^{-5}{x}^2\\ y_0=4.17071-0.00408x-2.6539\×{10}^{-4}{x}^2+6.43776\×{10}^{-6}{x}^3\\ \mathrm{\τ}=707.706+14.19376x\end{array}\right.$

70%:

$\left\{\begin{array}{c}A=0.00171+0.00932x-9.6137\×{10}^{-5}{x}^2\\ y_0=4.62703-0.00718x-1.09848\×{10}^{-4}{x}^2+3.22049\×{10}^{-6}{x}^3\\ \mathrm{\τ}=601.861+10.1786x\end{array}\right.$

80%:

$\left\{\begin{array}{c}A=-0.00955+0.01354x-1.21334\×{10}^{-4}{x}^2\\ y_0=5.11562+0.00121x-5.14842\×{10}^{-4}{x}^2+7.49891\×{10}^{-6}{x}^3\\ \mathrm{\τ}=532.467+8.46182x\end{array}\right.$

90%:

$\left\{\begin{array}{c}A=0.09196+0.00673x-8.67725\×{10}^{-6}{x}^2\\ y_0=5.66442+0.00703x-5.13515\×{10}^{-4}{x}^2-5.51339\×{10}^{-6}{x}^3\\ \mathrm{\τ}=1054.135\end{array}\right.$

100%:

$\left\{\begin{array}{c}A=0.1442+0.0005x-2.16857\×{10}^{-5}{x}^2\\ y_0=6.3357+0.01435x-5.13086\×{10}^{-4}{x}^2+4.27917\×{10}^{-6}{x}^3\\ \mathrm{\τ}=1265.894\end{array}\right.$

S423: obtain stress decay value according to stress decay model and battery trickle charge time；

S424: obtain battery surface static stress according to stress decay value.

S43: using battery surface dynamic stress as battery surface static stress.

S5: according to the battery surface static stress obtained, in conjunction with the respective function of stress Yu state-of-charge, estimation obtains the state-of-charge of lithium ion battery,

The respective function of stress therein and state-of-charge can have two, is respectively used to characterize the corresponding relation between charging static stress and state-of-charge, and for characterizing during electric discharge the corresponding relation between static stress and state-of-charge, so can improve precision.But static stress difference during due to discharge and recharge is less, in order to simplify computation complexity, can adopt the respective function of averaged static stress and state-of-charge, averaged static stress therein be charging time battery surface static stress and electric discharge time battery surface static stress meansigma methods.

The corresponding relation between static stress and state-of-charge during due to charging and discharging can be slightly different, but gap is less, specifically as shown in Figure 7, in Fig. 7, solid line represents the respective function image of static stress and state-of-charge during charging, the respective function image of static stress and state-of-charge when long dotted line is electric discharge, short dash line is the respective function image of draw stress and state-of-charge, as shown in the figure, for adopting " averaged static stress " to estimate the SOC systematic error produced, maximum error is less than 3%, and present the feature that platform area error is less, but but can be greatly simplified memory capacity and calculating process.

Shown in Fig. 8, SOC estimates that the input quantity of model is battery status signal and dynamic stress signal, and two signals are to synchronize.Namely battery status signal represents that battery is in charging, the token state standing or discharging, and represents charged state with " 1 " in this model, and " 0 " represents static condition, and "-1 " represents discharge condition.Battery status signal is additionally operable to calculate the trickle charge time, and when model inspection starts to charge up to battery, model starts to calculate the trickle charge time, when model inspection to battery enters other states, stops calculating；When detecting that battery starts to charge up again, model restarts to calculate the trickle charge time.Dynamic stress signal is the battery surface pressure that pressure transducer real-time monitors.

The work process of model is as follows:

Input signal monitored in real time by model, detect that battery is charged state when a certain moment, start to perform stress decay model, SOC is estimated by dynamic stress signal, and counted to get the trickle charge time by battery status signal, estimate SOC and trickle charge time as stress decay mode input gauge calculate obtain static stress, the battery charge state (SOC) obtaining current time of then tabling look-up；When a certain moment model inspection to battery is static condition, model keeps the SOC in a moment automatically；When a certain moment detects that battery is discharge condition, owing to discharge process is substantially not present stress decay, so the dynamic stress of discharge process is approximately equal to static stress, the battery charge state of current time can be obtained either directly through tabling look-up.

Such as, in the present embodiment, static stress and state-of-charge respective function adopt synopsis form, specifically as shown in table 2:

Table 2

Claims (6)

1. a soft bag lithium ionic cell charge state estimation method, it is characterised in that include step:

S1: gather battery surface dynamic stress and cell operating status signal, wherein, described cell operating status signal includes for indicating battery to be in the first data of charging, standing or discharge condition；

S2: judge whether battery is in static condition, if it has, then perform step S3, if it has not, then perform step S4；

S3: once estimate the result state-of-charge as current lithium ion battery using front；

S4: according to battery surface dynamic stress, obtains battery surface static stress, and performs step S5；

S5: according to the battery surface static stress obtained, the respective function according to stress Yu state-of-charge, estimation obtains the state-of-charge of lithium ion battery.

2. a kind of soft bag lithium ionic cell charge state estimation method according to claim 1, it is characterised in that described step S4 specifically includes step:

S41: judge whether battery is in charged state, if it has, then perform step S42, if it has not, then perform step S43；

S42: according to the battery surface dynamic stress collected and trickle charge time, it is thus achieved that battery surface static stress:

DS-S=SS

Wherein: DS is battery surface dynamic stress, S is stress decay value, and SS is battery surface static stress；

S43: using battery surface dynamic stress as battery surface static stress.

3. a kind of soft bag lithium ionic cell charge state estimation method according to claim 2, it is characterised in that described cell operating status signal also includes the second data for recording the battery trickle charge time,

Stress decay value in described step S42 obtains according to stress decay model, described stress decay model particularly as follows:

$S=y_0+{\mathrm{Ae}}^{-\frac{t_c}{\mathrm{\τ}}}$

Wherein: y₀For primary stress, A is attenuation amplitude, and τ is attenuation constant, t_cFor the battery trickle charge time.

4. a kind of soft bag lithium ionic cell charge state estimation method according to claim 3, it is characterised in that when battery starts to charge up each time, the battery trickle charge time is recalculated.

5. a kind of soft bag lithium ionic cell charge state estimation method according to claim 3, it is characterised in that described step S42 specifically includes step:

S421: estimate the state-of-charge of lithium ion battery according to battery surface dynamic stress；

S422: select the primary stress of correspondence, attenuation amplitude and attenuation constant according to estimation results, and substitute into stress decay model；

S423: obtain stress decay value according to stress decay model and battery trickle charge time；

S424: obtain battery surface static stress according to stress decay value.

6. a kind of soft bag lithium ionic cell charge state estimation method according to claim 3, it is characterized in that, in described step S5, the respective function of stress and state-of-charge is specially the respective function of averaged static stress and state-of-charge, wherein said averaged static stress is the meansigma methods of static stress during charge and discharge under same state-of-charge

Step S5 is particularly as follows: using the battery surface static stress that obtains as averaged static stress, the respective function according to stress Yu state-of-charge, estimation obtains the state-of-charge of lithium ion battery.