CN102788957A - Estimating method of charge state of power battery - Google Patents

Abstract

The invention relates to an estimating method of the charge state of a power battery. The estimating method is characterized by including the following steps: estimating system on chip (SOC) initial value of the battery according to last charge-discharge state of the battery and SOC value and standing time of the battery since last-time use comprehensively, calculating battery inner resistance and polarization voltage of the battery, collecting working current and voltage at two ends of the battery in real time, estimating SOC of the battery according to the SOC comprehensive estimation algorithm based on an expansion Kalman filtering method and an ampere-hour metering method, storing the SOC of the battery and the corresponding working current and voltage at two ends of the battery, storing current time, battery SOC and battery charge-discharge state before the system is closed down and closing the system down. The estimating method has the advantages of being small in SOC initial value, simple in algorithm, small in metering quantity and high in accuracy.

Description

A kind of power battery charged state evaluation method

Technical field

The present invention relates to a kind of evaluation method of battery charge state, be specifically related to a kind of state-of-charge evaluation method of used for electric vehicle electrokinetic cell.

Background technology

The state-of-charge of battery (SOC-State Of Charge) is one of important parameter of reflection battery performance, and it provides important evidence for administering and maintaining of electric battery.For the electric motor car driver, the SOC of the battery just fuel contents gauge as general-utility car is the same, has only the SOC that has known that accurately battery is current, just can judge the continual mileage of automobile.But because the height of battery is non-linear, the factor that influences battery SOC is numerous, and the SOC of estimating battery is more than calculating the many of oil mass complicacy.

In theory, the SOC's of battery is defined as the current battery dump energy Q of electric battery _CWith electric battery rated capacity Q ₀Ratio, can be through following formula statement:

During the battery Full Charge Capacity, definition SOC is 1; When battery discharge finished, definition SOC was 0.

At present, the method for estimating battery SOC has both at home and abroad: internal resistance method, open-circuit voltage method, ampere-hour measurement Law, Kalman filtering method, neural network method etc.Wherein, the internal resistance method through detecting the SOC that the internal resistance of cell comes counting cell, still accurately detects relatively difficulty of the internal resistance of cell according to the funtcional relationship between the internal resistance of cell and the battery SOC, and practicality is not strong; The SOC that the open-circuit voltage method is confirmed battery according to the open-circuit voltage and the funtcional relationship between the battery SOC of battery, but this method need leave standstill battery for a long time, generally is used for laboratory or the battery maintenance stage, can't on real vehicle, be applied; The ampere-hour measurement Law is carried out integration through the charging and discharging currents to battery and is drawn the electric weight inflow battery or that battery is emitted; Ampere-hour measurement Law algorithm is simple, realizes easily, is the algorithm of a kind of estimating battery SOC relatively more commonly used; But also there are some problems in the ampere-hour measurement Law, comprising:

(1) initial SOC that can't counting cell;

(2) have cumulative errors, to the accuracy requirement of current detecting system than higher;

Kalman filtering is a kind of Recursive Linear minimum variance estimate algorithm, utilizes the value in a last moment and the parameter value of measuring in real time to estimate in real time.When Kalman filter was applied to the battery SOC estimation, battery was represented as a linear discrete system, and battery SOC is a state variable of system.Kalman filtering algorithm is insensitive to the error of initial SOC, and the error of current detecting is had correcting action.But during the SOC of application card Thalmann filter estimating battery, need an accurate battery model, and the Kalman filter calculated amount is big, the computing power of master controller is required than higher.

The method of estimation (application number: 200710064294.4) based on the nickel-hydrogen power battery charged state of standard battery model of Tsing-Hua University; The SOC of utilization Kalman filtering algorithm estimating battery; Arithmetic accuracy is high, but the estimating algorithm of SOC initial value is single, does not consider the time of repose of battery.Whole algorithm is complicated, and calculated amount is bigger.

A kind of method for estimating charge state of power cell (application number: 200610167393) of BYD company; Its outstanding characteristics are temperature of having taken all factors into consideration battery, discharge and recharge factors such as number of times, charge-discharge magnification, battery actual capacity; Also different to these parameters of different batteries, need do a large amount of experiments and confirm these parameters, the algorithm practicality is not strong; And the algorithm of the initial SOC of estimating battery also is single open-circuit voltage method, and error is bigger.

The method of the SOC of the battery of the estimation hybrid electric vehicle of LG Chemical Ltd. (application number: 200680016312.5), the SOC of utilization open-circuit voltage method and ampere-hour measurement Law estimating battery, there is cumulative errors in this algorithm, and current detection accuracy is required than higher.

Comprehensively to the inquiry of the Patent data that retrieves, find that the subject matter of present various battery charge state methods of estimation is that the initial SOC algorithm for estimating of battery is single, error is bigger; But whole algorithm has perhaps guaranteed the precision too complex, and perhaps algorithm is simple still can't guarantee precision, does not have the algorithm of a kind of integration algorithm complexity and arithmetic accuracy.

Summary of the invention

The state-of-charge evaluation method that the purpose of this invention is to provide a kind of electrokinetic cell; The comprehensive estimate method of having used open-circuit voltage method, EKF method and ampere-hour measurement Law to combine; To improve the precision of algorithm, shortcut calculation makes algorithm in reality, obtain more applications.

The battery model of selecting for use among the present invention of being shown in Figure 1, wherein E (t) is an ideal voltage source, the open-circuit voltage of expression battery, under the temperature of confirming, the SOC of it and battery has fixing funtcional relationship.R is that battery ohmic internal resistance, Up are that the polarizing voltage, I (t) of battery is the battery voltage for the working current of battery (during charging for just, during discharge for negative), V (t).

According to the circuit structure of this battery model, the battery status spatial model that makes up discrete form is shown in formula (1), and k is the sampling instant point of battery management system, and Ts is two time intervals between the sampled point:

$\left\{\begin{array}{c}{\mathrm{SOC}}_k={\mathrm{SOC}}_{k-1}-\frac{1}{Q}\·I_{k-1}\·\mathrm{Ts}\\ V_k=E_k-R\·I_k+{\mathrm{Up}}_k\end{array}\right.$ Formula (1)

Wherein, Q is the theoretical capacity of electric battery, and when batteries charging, Q is the rated capacity of electric battery; When battery power discharge, the battery theoretical capacity of Q for going out according to the Peukert Equation for Calculating; SOC _kBe the SOC value of battery of this sampling instant, SOC _K-1SOC value of battery for a last sampling instant; I _K-1Be the current value of a last sampling instant, I _kCurrent value for this sampling instant; V _kBattery terminal voltage sampled value for this sampling instant; E _kOpen-circuit voltage for this sampling instant battery; Up _kPolarizing voltage for this sampling instant battery.

According to the discrete state spatial model of battery, set up Kalman filtering algorithm:

$\left\{\begin{array}{c}{X}_{k/k-1}=X_{k-1/k-Q}-\frac{1}{Q}\·I_{k-1}\·\mathrm{Ts}\\ Y_k=F\left(X_{k/k}\right){|}_{X_{k/k=X_{\mathrm{kk}/k-1}}}+R\·I_k+{\mathrm{Up}}_k\\ C_k=\frac{d\left(F\left(X_{k/k}\right)\right)}{d\left(X_{k/k}\right)}{|}_{X_{k/k}=X_{k/k-1}}\\ P_{k/k-1}=P_{k-1/k-1}+D\\ X_k=P_{k/k-1}\·C_k[C_k^2\·P_{k/k-1}+W]\\ X_{k/k}=X_{k/k-1}+K_k\·(V_k-Y_k)\\ P_{k/k}=P_{k/k-1}-C_k\·C_k\·P_{k/k-1}\\ k=\mathrm{1,2,3},\·\·\·\·\·\·\end{array}\right.$ Formula (2)

Wherein, k is the sampling instant point of battery management system; X _K/k-1Predicted value for this sampling instant of battery SOC; X _K/kBe the estimated value of battery SOC in this sampling instant; Y _kBe the battery voltage of this sampling instant through battery model calculating; I _K-1Be the current value of a last sampling instant, I _kCurrent value for this sampling instant; R is the internal resistance of cell; Up _kPolarizing voltage for this sampling instant battery; C _kBe observing matrix, P _K/k-1Predicted value mean square deviation for SOC; P _K/kEstimated value mean square deviation for SOC; D is the system noise variance matrix; K _kBe the systematic error gain; W is a systematic survey noise variance matrix; V _kBe the cell voltage of system in this sampling instant collection; F (X _K/k) function that the SOC of expression battery open circuit voltage E (k) and battery concerns, expression formula is shown in formula (3):

$E_k=F\left(X_{k/k}\right)=a_0\·X_{k/k}^n+a_1\·X_{k/k}^{n-1}+a_2\·X_{k/k}^{n-2}+\·\·\·+a_{n-1}\·X_{k/k}+a_n$ Formula (3)

Wherein, E _kBe the open-circuit voltage of battery, X _K/kBe the SOC of battery, a ₀, a ₁, a ₂..., a _nBe multinomial coefficient, n is a natural number, representes polynomial number of times.Formula (3) draws data by experiment in advance, and draws a through MATLAB data fitting instrument ₀, a ₁, a ₂..., a _nValue with n.

As shown in Figure 2, the invention is characterized in, comprise following steps successively:

Step (1):

In the start moment of battery management system, master controller carries out initialization to following parameter:

1) the theoretical capacity Q of counting cell;

2) systematic survey noise variance matrix W _k, W _kMeasuring error for the battery management system voltage detection module;

3) system noise variance matrix D is taken as 1;

4) the SI Ts of battery management system;

5) system prediction error covariance matrix initial value P _0/0

Step (2):

Initial SOC according to process flow diagram counting cell shown in Figure 3:

Step (2.1) is measured the open-circuit voltage of battery;

Step (2.2) reads the data in the memory block of depositing the last unused time among the Flash, and the initial SOC of counting cell;

Step (2.2.1) is if the data among the Flash are 0xFFFFFF, calculates the initial SOC of this battery through battery open circuit voltage being applied to formula (4) in when start.

Y=p ₀X ⁿ+ p ₁X ^N-1+ ... + p _N-1X+p _nFormula (4)

Wherein, y is the SOC of battery, and x is the open-circuit voltage of battery, p ₀P _nBe multinomial coefficient, n is a natural number, representes polynomial number of times.Formula (4) draws data by experiment in advance, and draws p through MATLAB data fitting instrument ₀P _nAnd the value of n.

Step (2.2.2) then reads the time of this on time and last shutdown if the data among the Flash are not 0xFFFFFF, and the difference that juice is calculated between the two is Δ T;

Step (2.2.2.1) is if Δ T >=6 hour; Data when then reading in the Flash district the last shutdown of expression among the charging and discharging state ex_state of battery; The charging and discharging state of battery open circuit voltage and last battery is applied to the initial SOC of related function counting cell, and wherein said related function is through formula (5) expression:

$\left\{\begin{array}{cc}y={\mathrm{<figure-callout\; id="0"\; label="p"\; filenames="HSA00000506710300041.png"\; state="\{\{state\}\}">p</figure-callout>}}_0\&CenterDot;x^n+p_1\&CenterDot;x^{n-1}+\&CenterDot;\&CenterDot;\&CenterDot;+p_{n-1}\&CenterDot;x+p_n& \mathrm{ex}\_\mathrm{state}=1\\ y={\mathrm{<figure-callout\; id="0"\; label="q"\; filenames="HSA00000506710300041.png"\; state="\{\{state\}\}">q</figure-callout>}}_0\&CenterDot;x^n+q_1\&CenterDot;x^{n-1}+\&CenterDot;\&CenterDot;\&CenterDot;+q_{n-1}\&CenterDot;x+q_n& \mathrm{ex}\_\mathrm{state}=2\end{array}\right.$ Formula (5)

Wherein, y is the state-of-charge of battery, and x is the open-circuit voltage of battery, p ₀P _nAnd q ₀Q _nBe multinomial coefficient, n is a natural number, and ex_state=1 representes that last battery is in charged state, and ex_state=2 representes that last battery is in discharge condition.Formula (5) draws data by experiment in advance, and draws p through MATLAB data fitting instrument ₀P _n, q ₀Q _nAnd the value of n.

Step (2.2.2.2) if 4 hours≤Δ T＜6 hours; Data when then reading the charging and discharging state-ex_state of battery when the expression last time shuts down in the Flash district and representing last the shutdown among the state-of-charge-ex_soc of battery; And the charging and discharging state that battery open circuit voltage and battery is last is applied to the battery SOC that recomputates when formula (5) draws this start, and the battery SOC that ex_soc is calculated during with start is got the initial SOC of average battery when starting shooting as this.

Step (2.2.2.3) is if Δ T＜4 hour, the initial SOC of battery when the state-of-charge-ex_soc of battery starts shooting as this when then reading last shutdown.

Step (3):

The control battery carries out work, working current, the battery that the detects battery battery terminal voltage of moment of starting working, according to formula (6) counting cell internal resistance:

$R=\left|\frac{V-E}{I}\right|$ Formula (6)

Wherein, R is the internal resistance of cell, and V is the start working battery terminal voltage of moment of battery, and E is the open-circuit voltage of battery, and I is the working current of battery.

Step (4):

Step (4.1) is for k=1, and 2,3 ..., 100 sampling instant point, circulation is operated as follows:

Step (4.1.1) battery management system is gathered battery voltage V _kWorking current I with battery _k

Step (4.1.2) according to formula (2), is carried out Kalman filtering algorithm with the initial value of the initial SOC of battery as expanded Kalman filtration algorithm.

Step (4.1.3) is every at a distance from 5s, and battery SOC and corresponding cell voltage, electric current, charging and discharging state that step (4.1.2) is calculated deposit the Flash memory module in, accomplish the calculating and the recording process of a SOC value.

Step (4.2) is for ensuing k=101, and 102,103 ... The sampling instant point, the circulation operate as follows:

Step (4.2.1) battery management system is gathered battery voltage V _kWorking current I with battery _k

Step (4.2.2) according to formula (7), is carried out the ampere-hour measurement Law, the SOC of estimating battery with the SOC value of last sampling instant of the expanded Kalman filtration algorithm initial value as the ampere-hour measurement Law.Wherein, formula (7) is explained through following formula:

${\mathrm{SOC}}_k={\mathrm{SOC}}_{k-1}-\frac{I_{k-1}\&CenterDot;\mathrm{Ts}}{Q}$ Formula (7)

Wherein, SOC _kBe the battery SOC of this sampling instant point, SOC _K-1Be the battery SOC of a last sampling instant point, I _K-1Be the electric current of a last sampling instant point, Ts is a sampling time interval, and Q is the theoretical capacity of battery, and when battery charge, Q is the rated capacity of battery, when battery discharge, and the theoretical capacity of Q for coming out according to the Peukert Equation for Calculating.

Step (4.2.3) is every at a distance from 5s, and battery SOC and corresponding cell voltage, electric current, charging and discharging state that step (4.2.2) is calculated deposit the Flash memory module in, accomplish the calculating and the recording process of a SOC value.

Step (5):

After system provides off signal, with the SOC of current time, battery, battery charging and discharging state storage in Flash memory block, shutdown system then.

According to above-mentioned steps; The initial SOC value of at first accurate estimating battery; Then according to the initial SOC of battery and battery current, the information of voltage gathered in real time, the SOC value of real-time estimating battery, and voltage, electric current, SOC, the charging and discharging state of battery be stored in the Flash memory module; So that when the subsequent examination battery performance, can be as a reference with the data in the Flash memory module.

Advantage of the present invention:

1. accurate battery SOC initial value estimating algorithm.During the initial SOC of estimating battery, taken all factors into consideration the charging and discharging state when battery is last to be used, the last SOC of back battery and the time of repose of battery of using, arithmetic accuracy is high.

2. the real-time estimating algorithm of battery SOC has been taked the comprehensive estimate algorithm based on EKF method and ampere-hour measurement Law, and precision is high, calculated amount is little, and is little to the computing power requirement of master controller, is easy to realize.

3. during with the SOC of EKF method estimating battery, according to the difference of the charging and discharging state of battery, algorithm also has corresponding change, and the result is more accurate in estimation.

Description of drawings

The circuit structure of Fig. 1 battery model

Calculate the algorithm flow chart of battery SOC among Fig. 2 the present invention

The algorithm flow chart of the initial SOC of counting cell among Fig. 3 the present invention

Its open-circuit voltage and SOC relation curve during the charging of Fig. 4 6AH Ni-MH battery

Its open-circuit voltage and SOC relation curve during the discharge of Fig. 5 6AH Ni-MH battery

The practical implementation method

The concrete steps of embodiment of the present invention can be divided into following four steps:

Step (1): battery set charge/discharge experiment;

Step (2): experimental data is handled, drawn Peukert constant and battery open circuit voltage and the SOC relation curve and the mathematic(al) representation of Ni-MH battery;

Step (3): make up complete SOC estimating algorithm;

Step (4): algorithm application is in battery management system.

Be composed in series by 6 joint cells with one below, rated capacity is 6AH, and nominal voltage is that the Ni-MH battery of 7.2V is an example, specifically introduce implementation process of the present invention.

Step (1) is carried out the battery set charge/discharge experiment.

1) constant current charge-discharge experiment: respectively the discharge-rate of battery with 0.2C, 1C, 2C, 5C discharged, draw discharge data as shown in table 1.

Discharge data under the different discharge-rates of table 1

Discharge current/A	Electric weight/the mA.H that emits	Discharge time/min
			1.1012(0.2C)	5570.6	303.4
5.506(1C)	5469.2	59.5
			11.0119(2C)	5353.4	29.2
27.5205(5C)	5274.6	11.5

2) combination discharges and recharges experiment:

At first to battery with low discharging current, the voltage at sampling battery two ends is after the battery voltage is lower than 6V; Think that the electric weight of battery all gives out light, this moment, the SOC of battery was 0, stopped discharge; Left standstill 6 hours, the voltage of measuring the battery two ends is the open-circuit voltage that SOC is 0 o'clock correspondence; Battery with 0.1C (0.6A) current charges, after 0.5 hour, is stopped charging, the voltage at sampling battery two ends, the SI is 1s, after 30 data points of sampling, stops sampling; Battery left standstill 6 hours, and the voltage of measuring the battery two ends is the open-circuit voltage that SOC is 5% o'clock correspondence; Continuation to battery with the current charges of 0.1C 0.5 hour, the voltage at sampling battery two ends, the SI is 1s, after 30 data points of sampling, stops sampling; Left standstill then 6 hours, obtaining SOC is 10% o'clock corresponding battery open circuit voltage; Using such method, obtaining SOC respectively is the battery open circuit voltage of 15%, 20%, 30%, 40%, 50%, 60%, 70%, 75%, 80%, 85%, 90%, 95%, 100% correspondence.Simultaneously, according to the voltage difference before and after leaving standstill, can obtain the size of polarizing voltage of the different SOC points correspondences of battery.

After battery is full of electricity, to the multiplying power discharging of battery, draw the SOC and the battery open circuit voltage relation data of battery when discharging with 0.1C with same method.

As shown in table 2ly discharge and recharge the data that draw of experiment for this combination.

Table 2 combination discharges and recharges experimental data

The processing of step (2) experimental data

1) calculating of Peukert constant

When battery discharge, the Peukert equation is arranged:

I ⁿT=K formula (8)

Wherein, n and K are constants, and for same Battery pack, these two constants are identical.According to the data in the table 1, choose two groups of data arbitrarily, can calculate the Peukert constant.Through calculating, the value of Peukert constant is:

n＝1.0185，K＝5.6052。So the theoretical capacity when battery discharges with different discharge currents is: Q=I ^1-n* K, that is: Q=I ^-0.0185* 5.6052.

2) battery open circuit voltage and SOC relation function obtains

Based on the data in the table 2, simulate the relation curve of battery open circuit voltage and SOC with MATLAB data fitting instrument cftool.Through trial repeatedly; Find with 4 order polynomial matches match open-circuit voltage and SOC relation curve more accurately; Battery open circuit voltage and SOC relation curve when being illustrated in figure 4 as the charging that simulates, battery open circuit voltage and SOC relation curve when Fig. 5 is discharge.Wherein, battery open circuit voltage and SOC relation function expression formula such as formula (9) are said during charging; Battery open circuit voltage and SOC relation function expression formula such as formula (10) are said during discharge.

E _Charge=-1.9632S ⁴+ 6.9125S ³-7.953S ²+ 3.8715S+7.4085 formula (9)

E _Discharge=-0.7394S ⁴+ 4.3622S ³-5.8873S ²+ 3.2522S+7.312 formula (10)

According to same method, we can draw the mathematic(al) representation of expression battery SOC and battery open circuit voltage funtcional relationship.

Formula (4) can specifically be expressed as so:

Y=1.289*x ³-28.041*x ²+ 202.900*x-488.319 formula (11)

Formula (5) can specifically be expressed as:

$\left\{\begin{array}{cc}y=1.289*x^3-28.041*x^2+202.900*x-488.319& \mathrm{ex}\_\mathrm{state}=1\\ y=-2.200*x^3+52.390*x^2-409.500*x+1063.2& \mathrm{ex}\_\mathrm{state}=2\end{array}\right.$ Formula (12)

Step (3) makes up complete battery SOC estimating algorithm

The estimated value computing formula of the measurement amount in the Kalman filtering algorithm is:

$Y_k=F\left(X_{k/k}\right){|}_{X_{k/k}=X_{k/k-1}+R\&CenterDot;I_k+\mathrm{Up}\left(k\right)}$ Formula (13)

Wherein, F (X _K/k) expression formula can obtain through formula (9) and formula (10):

F (X _K/k) _C=-1.9632X _K/k ⁴+ 6.9125X _K/k ³-7.953X _K/k ²+ 3.8715X _K/k+ 7.4085 formula (14)

F (X _K/k) _D=-0.7394X _K/k ⁴+ 4.3622X _K/k ³-5.8873X _K/k ²+ 3.2522X _K/k+ 7.312 formula (15)

When battery in when charging, F (X _K/k) express through formula (14); When battery in when discharge, F (X _K/k) express through formula (15).

Up (k) is the polarizing voltage of battery, and experimental data is made into table, when calculating, through tabling look-up and approach based on linear interpolation calculates the value of Up (k).

Kalman filter observing matrix C _kExpression formula can calculate through formula (9) and (10):

C _k=-7.853X _K/k ³+ 20.738X _K/k ²-15.906X _K/k+ 3.872 formula (16)

C _k=-2.958X _K/k ³+ 13.087X _K/k ²-11.775X _K/k+ 3.252 formula (17)

When battery in when charging, C _kExpression formula express through formula (16); When battery in when discharge, C _kExpression formula express through formula (17).

Step (4) algorithm application is in battery management system

For the present invention, battery management system is gathered the electric current I of battery in real time _k, battery voltage V _k, main control MCU is according to SOC estimating algorithm estimating battery SOC of the present invention and store and show.

SOC estimating algorithm among the present invention has been considered the charging and discharging state that battery is current and the time of repose of previous battery charging and discharging state and battery simultaneously when the initial SOC of estimating battery, the result is more accurate in estimation.When real-time estimating battery SOC, taked comprehensive estimate algorithm based on EKF method and ampere-hour measurement Law, arithmetic accuracy is high, and calculated amount is less, the computing power of master controller is required little, is easy to realization.

The present invention is an example with the electric battery that is composed in series of Ni-MH battery of 6 joint 6AH, has set forth concrete implementation method, the invention is not restricted to battery types and the concrete parameter set forth.

Claims (1)

1. a power battery charged state evaluation method is characterized in that, contains following steps successively:

Step (1) is measured the open-circuit voltage of battery;

Whether for the first time step (2) judges system's operation; Whether for the first time determination methods is: judge system's operation through the data in the Flash memory block of depositing the last unused time; If data are 0xFFFFFF; Then system is operation for the first time, if data are not 0xFFFFFF, then system is not operation for the first time.

Step (2.1) is if system is operation for the first time, and through battery open circuit voltage being applied to the initial SOC of related function battery when calculating this start, wherein, SOC is the abbreviation of English State Of Charge, the state-of-charge of expression electric battery.Wherein, said related function is expressed through following formula:

y＝p ₀·x ⁿ+p ₁·x ^n-1+…+p _n-1·x+p _n

Wherein, y is the SOC of battery, and x is the open-circuit voltage of battery, p ₀P _nBe multinomial coefficient, n is a natural number, representes polynomial number of times.This formula is drawn data and is drawn p through MATLAB data fitting instrument by experiment in advance ₀P _nValue with n.

Step (2.2) then read the time of this on time and last shutdown, and the difference of calculating between the two is Δ T if system is not operation for the first time;

Step (2.2.1) is if Δ T >=6 hour; Data when then reading in the Flash district the last shutdown of expression among the charging and discharging state ex_state of battery; The charging and discharging state of battery open circuit voltage and last battery is applied to the initial SOC of related function counting cell, and wherein said related function is represented through following formula:

$\left\{\begin{array}{cc}y=p_0\&CenterDot;x^n+p_1\&CenterDot;x^{n-1}+\&CenterDot;\&CenterDot;\&CenterDot;+p_{n-1}\&CenterDot;x+p_n& \mathrm{ex}\_\mathrm{state}=1\\ y=q_0\&CenterDot;x^n+q_1\&CenterDot;x^{n-1}+\&CenterDot;\&CenterDot;\&CenterDot;+q_{n-1}\&CenterDot;x+q_n& \mathrm{ex}\_\mathrm{state}=2\end{array}\right.$

Wherein, y is the state-of-charge of battery, and x is the open-circuit voltage of battery, p ₀P _nAnd q ₀Q _nBe multinomial coefficient, n is a natural number, representes polynomial number of times, and ex_state=1 representes that last battery is in charged state, and ex_state=2 representes that last battery is in discharge condition.This formula draws data by experiment in advance, and draws n and p through MATLAB data fitting instrument ₀P _nAnd q ₀Q _nValue.

Step (2.2.2) if 4 hours≤Δ T＜6 hours; Data when then reading the charging and discharging state-ex_state of battery when the expression last time shuts down in the Flash district and representing last the shutdown among the state-of-charge-ex_soc of battery; And the charging and discharging state that battery open circuit voltage and battery is last is applied to the battery SOC that recomputates when correlation formula draws this start, and the battery SOC that ex_soc is calculated during with start is got the initial SOC of average battery when starting shooting as this.Wherein related correlation formula is expressed through following formula:

$\left\{\begin{array}{cc}y=p_0\&CenterDot;x^n+p_1\&CenterDot;x^{n-1}+\&CenterDot;\&CenterDot;\&CenterDot;+p_{n-1}\&CenterDot;x+p_n& \mathrm{ex}\_\mathrm{state}=1\\ y=q_0\&CenterDot;x^n+q_1\&CenterDot;x^{n-1}+\&CenterDot;\&CenterDot;\&CenterDot;+q_{n-1}\&CenterDot;x+q_n& \mathrm{ex}\_\mathrm{state}=2\end{array}\right.$

Wherein, y is the state-of-charge of battery, and x is the open-circuit voltage of battery, p ₀P _nAnd q ₀Q _nBe multinomial coefficient, n is a natural number, representes polynomial number of times, and ex_state=1 representes that last battery is in charged state, and ex_state=2 representes that last battery is in discharge condition.This formula draws data by experiment in advance, and draws n and p through MATLAB data fitting instrument ₀P _nAnd q ₀Q _nValue.

Step (2.2.3) is if Δ T＜4 hour, the initial SOC of battery when the state-of-charge-ex_soc of battery starts shooting as this when then reading last shutdown.

Step (3) control battery carries out work, working current, the battery that the detects battery battery terminal voltage of moment of starting working, according to the computes internal resistance of cell:

$R=\left|\frac{V-E}{I}\right|$

Wherein, R is the internal resistance of cell, and V is the start working battery terminal voltage of moment of battery, and E is the open-circuit voltage of battery, and I is the working current of battery.

Step (4) is according to the SOC of battery and the data of characterizing battery polarizing voltage and battery SOC relation, and through the polarizing voltage of interpolation calculation battery, wherein, the data of the SOC of characterizing battery polarizing voltage and battery relation are obtained by experiment in advance.

Step (5) is with the initial value of the initial SOC of battery as expanded Kalman filtration algorithm, and according to the EKF method, every separated 1s calculates the state-of-charge of one-shot battery, and the execution number of times kalman_count of record expanded Kalman filtration algorithm.

If kalman_count＞100, the battery charge state SOC that then last expanded Kalman filtration algorithm is drawn _kAs the initial value of ampere-hour measurement Law, according to the ampere-hour measurement Law, every state-of-charge that calculates one-shot battery at a distance from 1s.

Step (6) is every at a distance from 5s, and battery SOC and corresponding cell voltage, electric current, charging and discharging state that step (5) is calculated deposit the Flash memory module in, accomplish the calculating and the recording process of a SOC value, and so circulation is until system closedown.

After step (7) system provides off signal, with the SOC of current time, battery, battery charging and discharging state storage in the Flash memory block, shutdown system after accomplishing.

Images