CN104502858A - Power battery SOC estimation method based on backward difference discrete model and system thereof - Google Patents

Abstract

Provided is a power battery SOC estimation method based on backward difference discrete model and a system thereof; the method comprises the following steps: step one, establishing a backward difference discrete model of a power battery, identifying parameters of the backward difference discrete model by a least square method including forgetting factors; step two, on the basis of the backward difference discrete model of the power battery obtained in step one, using self-adaptive extended Kalman filter in combination with a non-linear relationship between an open-circuit voltage and the SOC to complete an effective estimation of the power battery SOC. In the system, voltage and current sensors connected with the power battery are connected with an embedded microcontroller via an analog-digital conversion module. The microcontroller comprises a low-pass filter pre-processing module, a backward difference discrete battery model parameter online identification module, and an AEKF algorithm SOC estimation module. The obtained SOC result is sent to a CAN network of a display device. The power battery SOC estimation system based on backward difference discrete model is simple in structure; the parameter identification speed and precision are increased; the affection to the identification caused by history data is reduced; the calculation is convenient; and the SOC estimation precision is high.

Description

Based on electrokinetic cell SOC method of estimation and the system of backward difference discrete model

Technical field

The present invention relates to the state of charge estimation technique field of automobile power cell, be specially the electrokinetic cell SOC method of estimation based on backward difference discrete model and system.

Background technology

In recent years, as electrokinetic cell, lithium battery compared with traditional lead-acid battery, Ni-MH battery, have that energy density is high, memory-less effect, have extended cycle life, the feature such as environmental friendliness, so lithium battery has become the main body of electric automobile power battery.

In electric automobile, the state of charge SOC (state-of-charge) of accurate estimation battery, i.e. battery dump energy is prerequisite and the key of the good operation of cell management system of electric automobile BMS (battery management system).

The accurate estimation of battery state of charge SOC mainly comprises two parts: the estimation of parameter identification and state of charge SOC.Conventional parameter identification method has genetic algorithm, least square method, double card Kalman Filtering etc.; The evaluation method of conventional state of charge SOC has open-circuit voltage method, ampere-hour integral method, Kalman filtering, EKF, Unscented kalman filtering, adaptive Kalman filter etc.

The combination that this two parts are different, forms different SOC method of estimation, and the speed and the precision that obtain SOC have notable difference.Current use more, more representationally to have: 1. the battery charge state of Unscented kalman filtering estimates experimental study, the combination of least square method and Unscented kalman filtering; 2. the combination of least square method and open-circuit voltage method; 3. the application of adaptive Kalman filter in vehicle lithium-ion power battery SOC estimation, the combination of genetic algorithm and adaptive Kalman filter; 4. the lithium battery SOC based on RLS and EKF algorithm estimates, the combination of least square method and EKF; 5. utilize two Kalman filtering algorithm to estimate the internal state of electric vehicle lithium-ion-power cell, utilize two Kalman filtering algorithm to have estimated SOC and the inner parameter change of lithium ion power storage battery used for electric vehicle simultaneously.

But these SOC methods of estimation existing are that precision is not high a bit, some is that calculation of complex speed is slow, all fails to meet automobile power cell---the online of lithium battery state of charge is accurately estimated.

Summary of the invention

The object of the invention is openly a kind of electrokinetic cell SOC method of estimation based on backward difference discrete model, the charge and discharge process of electrokinetic cell is a process more slowly, namely open-circuit voltage is relatively stable at short notice, this method adopts backward difference, obtains structure simple electrokinetic cell backward difference discrete model; Adopt least square method (the Forgetting Factor Recursive Least Squaresalgorithm containing forgetting factor, FFRLS) parameter identification is carried out to backward difference discrete model, in conjunction with adaptive extended kalman filtering (AEKF) computing method based on maximum likelihood criterion, complete effective estimation of electrokinetic cell state of charge SOC.

Another object of the present invention designs an electrokinetic cell SOC estimating system based on backward difference discrete model, and it can embed in the equipment using electrokinetic cell, realizes real-time SOC On-line Estimation and the display of electrokinetic cell.

The electrokinetic cell SOC method of estimation based on backward difference discrete model of the present invention's design, comprises two steps:

The first step, set up the backward difference discrete model of electrokinetic cell, by the least square method (FFRLS) containing forgetting factor, identification is carried out to the parameter of backward difference discrete model.

Second step, backward difference discrete model based on the electrokinetic cell of first step gained, in conjunction with the nonlinear relationship of open-circuit voltage and SOC, adopt adaptive extended kalman filtering (AEKF) computing method, complete effective estimation of electrokinetic cell SOC.

The first step, electrokinetic cell backward difference discrete model and parameter identification

1.1. electrokinetic cell model

For battery management system, conventional battery model has: thermal model, electrochemical model and equivalent-circuit model etc.Equivalent-circuit model, compared with other battery models, can show the relation between electric current and voltage more intuitively, be easy to the expression of analytic equation, is convenient to analysis and the identification of Model Parameters of battery.

The present invention adopts the most widely used a kind of battery equivalent model at present, and the Thevenin model of battery describes the Static and dynamic performance of battery.The polarization resistance R of battery _pwith the polarization capacity C of battery _pformation single order reinforced concrete structure in parallel, represent the polarization reaction of battery, RC both end voltage is U _p(t); Serial connection Ohmage R ₀with Uoc, Uoc are the open-circuit voltage OCV of battery, sampling obtains battery terminal voltage U (t) and flows through ohmic internal resistance R ₀current i (t).

Battery Thevenin model representation is as follows:

$\frac{d{U}_p\left(t\right)}{\mathrm{dt}}=-\frac{U_p\left(t\right)}{R_p{C}_p}+\frac{i\left(t\right)}{C_p}---\left(1\right)$

U(t)＝U _OC(t)-R ₀i(t)-U _p(t) (2)

1.2 model discretize and parameter identifications

1.2.1 model discretize

By backward difference method to above-mentioned battery model discretize, obtain difference equation, after arrangement

U(k)-U _OC(k)＝a[U(k-1)-U _OC(k-1)]+bI(k)+cI(k-1) (3)

In formula, Uoc (k) represents the open-circuit voltage in k moment; U (k) is k moment battery terminal voltage sampled value; I (k) is k moment loop current sampled value; A, b, c are model parameter.

The charge and discharge process of battery is a process more slowly, and open-circuit voltage Uoc is relatively stable at short notice, namely has

$\frac{d{U}_{\mathrm{OC}}}{\mathrm{dt}}=\frac{U_{\mathrm{OC}}\left(k\right)-U_{\mathrm{OC}}(k-1)}{T}\≈0---\left(4\right)$

△U _OC(k)＝U _OC(k)-U _OC(k-1)≈0 (5)

Then formula (3) and formula (5) combine, and the backward difference discrete model of battery is:

U(k)＝aU(k-1)+bI(k)+cI(k-1)+(1-a)U _OC(k)

After arrangement

U(k)-U _OC(k)＝a[U(k-1)-U _OC(k-1)]+bI(k)+cI(k-1) (6)

A, b and c represent that battery backward difference discrete model parameter is as follows:

$R_0=\frac{c}{a}---\left(7\right)$

$R_p=\frac{-\mathrm{ab}-c}{a(1-a)}---\left(8\right)$

$C_p=\frac{{\mathrm{Ta}}^2}{-\mathrm{ab}-c}---\left(9\right)$

Wherein T is the sampling period.

Compare with conventional bilinear transformation discrete model, battery backward difference walk-off-mode pattern (6) of the present invention, model parameter relation is simple, is convenient to open-circuit voltage Uoc and estimates.

1.2.2 the parameter identification of battery difference discrete model

The present invention adopts least-squares algorithm (FFRLS) the identification battery model parameter containing forgetting factor, and process is as follows:

$\widehat{\mathrm{\θ}}\left(k\right)=\widehat{\mathrm{\θ}}(k-1)+K\left(k\right)e\left(k\right)---\left(10\right)$

Wherein

In formula: φ (k) is data vector, θ (k) is estimated parameter vector.The prediction error that e (k) is U (k). for the estimated value of θ (k), initial value with P (0) rule of thumb assignment, K (k) is gain, and λ is forgetting factor, λ=0.95 ~ 1.

In least square (11) formula of backward difference discrete model (10) correspondence of battery, variable is specially:

φ(k)＝[U(k-1),I(k),I(k-1),1] ^T(11)

θ(k)＝[a,b,c,(1-a)U _OC(k)] ^T(12)

Tried to achieve the value of a, b, c by the least-squares algorithm (FFRLS algorithm) containing forgetting factor, substitute into formula (7) (8) (9), obtain battery model parameter R ₀, R _p, C _p, U _oCvalue.

Second step, to estimate based on the battery state of charge SOC of adaptive extended kalman filtering AEKF

Choose SOC and electric capacity C _pterminal voltage be state variable, corresponding k moment state variable, i.e. X _k=[SOC _ku _p,k] ^t, system state equation and measurement equation as follows:

$\left\{\begin{array}{c}{X}_k=A_{k|k-1}{X}_{k-1}+B_{k-1}{i}_{k-1}+w_{k-1}\\ Y_k=U_{\mathrm{oc}}\left({\mathrm{SOC}}_k\right)-R_0{i}_k-U_{p,k}+v_k\end{array}\right.---\left(13\right)$

Wherein, U _oc(SOC _k) represent battery open circuit voltage U _ocand the k moment nonlinear relationship between SOC is as follows:

U _oc(SOC _k)＝k ₁SOC _k ⁸+k ₂SOC _k ⁷+k ₃SOC _k ⁶+k ₄SOC _k ⁵+

(14)

k ₅SOC _k ⁴+k ₆SOC _k ³+k ₇SOC _k ²+k ₈SOC _k+k ₉

Use least square method, certain the model electrokinetic cell open-circuit voltage U obtained by on-line identification _ocsOC with testing this model electrokinetic cell obtained, tries to achieve this model electrokinetic cell coefficient k ₁~ k ₉.

AEKF algorithm estimates that SOC detailed process is as follows:

2.1 state estimation:

X _k=[SOC _ku _p,k] ^tstate in the estimated value in current time (k moment) is

${\hat{X}}_{k|k}={\hat{X}}_{k|k-1}+K_k(Y_{m|k}-{\hat{Y}}_k)---\left(15\right)$

Wherein K _k, expression formula is

Wherein k represents current time, and k-1 represents previous moment, represent based on current time (k moment) state gained X _k=[SOC _ku _p,k] ^tk moment estimated value, represent based on previous moment ((k-1) moment) state gained k moment X _kestimated value, represent based on previous moment ((k-1) moment) state gained (k-1) moment X _kthe estimated value of state.Y _m|kthe measured value of k moment battery terminal voltage, it is the terminal voltage predicted value after upgrading in the k moment. the discreet value based on previous moment state gained k moment SOC, Q _kk moment systematic procedure noise w _kcovariance.R _kk moment system measurements noise v _kcovariance.

The renewal of parameter and state in 2.2 recurrence calculation processes:

2.2.1 parameter Q _k, R _kupgrade

Wherein, μ _kthe difference of k moment terminal voltage actual value and the terminal voltage predicted value after upgrading, F _kbe the mean value of every L moment corresponding difference, L is self-adapting window.

2.2.2 state upgrade:

Wherein, Q _nfor the rated capacity of electrokinetic cell, η represents efficiency for charge-discharge, and T represents the sampling period.

The present invention is based on the electrokinetic cell SOC estimating system of backward difference discrete model, comprise the display of microcontroller and connection thereof, described microcontroller is embedded microcontroller, and electrokinetic cell output terminal is connected to voltage sensor and current sensor.Voltage sensor and current sensor are connected through analog-to-digital conversion module and embed microcontroller, embed microcontroller and contain low-pass filtering pretreatment module, the on-line parameter identification module of backward difference discrete electrical pool model and AEKF algorithm SOC state estimation module.Embed microcontroller and connect display, be also connected to CAN (controller local area network Controller Area Network) bus interface and/or RS232 interface.Native system is together with electrokinetic cell, be embedded in the equipment using electrokinetic cell, within a sampling period, complete voltage, the collection of electric current, battery model parameter identification and modification and SOC On-line Estimation, gained SOC result shows over the display or is directly sent to the controller local area network of this equipment.

Compared with prior art, the advantage of the method and system that the electrokinetic cell SOC that the present invention is based on backward difference discrete model estimates is: 1, adopt in backward difference discrete electrical pool model, have nothing to do with the history value of battery model inner parameter, structure is simple, is conducive to the identification speed and the precision that improve parameter; 2, with the least square method of recursion containing forgetting factor, identification is carried out to the parameter of backward difference discrete model, focus on new data to the renewal of weights, reduce historical data parameter identification influence degree, efficiently avoid the continuous iteration of least square method of recursion along with slow change battery parameter data of routine, upgrade " data the are saturated " problem that will occur; 3, consider the nonlinear relationship of SOC and open-circuit voltage Uoc, adopt adaptive extended kalman filtering (AEKF) algorithm of maximum likelihood criterion to carry out On-line Estimation to SOC, features simple structure, convenience of calculation, SOC estimated accuracy is high.

Accompanying drawing explanation

The Thevenin model circuit diagram of battery in Fig. 1 embodiment of the method that to be this estimate based on the electrokinetic cell SOC of backward difference discrete model.

The SOC estimated value of Fig. 2 embodiment of the method that to be this estimate based on the electrokinetic cell SOC of backward difference discrete model and SOC test the comparative graph of acquired value.

Fig. 3 is the system embodiment structural representation that this is estimated based on the electrokinetic cell SOC of backward difference discrete model.

Embodiment

Based on the embodiment of the method that the electrokinetic cell SOC of backward difference discrete model estimates

This embodiment of the method estimated based on the electrokinetic cell SOC of backward difference discrete model, concrete steps are as follows:

The first step, electrokinetic cell backward difference discrete model and parameter identification

1.1. electrokinetic cell model

Adopt the Thevenin model of battery, as shown in Figure 1, the polarization resistance R of battery _pwith the polarization capacity C of battery _pformation single order reinforced concrete structure in parallel, represent the polarization reaction of battery, RC both end voltage is U _p(t); Serial connection Ohmage R ₀with Uoc, Uoc are the open-circuit voltage OCV of battery, sampling obtains battery terminal voltage U (t) and flows through ohmic internal resistance R ₀current i (t).

Battery Thevenin model representation is as follows:

$\frac{d{U}_p\left(t\right)}{\mathrm{dt}}=-\frac{U_p\left(t\right)}{R_p{C}_p}+\frac{i\left(t\right)}{C_p}---\left(1\right)$

U(t)＝U _OC(t)-R ₀i(t)-U _p(t) (2)

1.2 model discretize and parameter identifications

1.2.1 model discretize

By backward difference method to above-mentioned battery model discretize, obtain difference equation, after arrangement

U(k)-U _OC(k)＝a[U(k-1)-U _OC(k-1)]+bI(k)+cI(k-1) (6)

In formula, Uoc (k) represents the open-circuit voltage in k moment; U (k) is the battery terminal voltage sampled value in current k moment; I (k) is the loop current sampled value in current k moment; A, b, c are model parameter.

The relation of a, b and c and battery backward difference discrete model parameter is as follows:

$R_0=\frac{c}{a}---\left(7\right)$

$R_p=\frac{-\mathrm{ab}-c}{a(1-a)}---\left(8\right)$

$C_p=\frac{{\mathrm{Ta}}^2}{-\mathrm{ab}-c}---\left(9\right)$

Wherein T is the sampling period.

1.2.2 the parameter identification of battery difference discrete model

Containing the estimated value of least-squares algorithm (FFRLS) identification model parameter θ (k) of forgetting factor process as follows:

$\widehat{\mathrm{\θ}}\left(k\right)=\widehat{\mathrm{\θ}}(k-1)+K\left(k\right)e\left(k\right)---\left(10\right)$

Wherein:

In formula: φ (k) is data vector, θ (k) is estimated parameter vector.The prediction error that e (k) is U (k).Initial value with P (0) rule of thumb assignment. for the estimated value of θ (k).λ is forgetting factor, and this example gets λ=0.99.

Tried to achieve the value of a, b, c by FFRLS algorithm, thus obtain model parameter R ₀, R _p, C _p, U _oCvalue.

Second step, to estimate based on the battery state of charge SOC of adaptive extended kalman filtering AEKF

Choose SOC and electric capacity C _pterminal voltage be the state representation in state variable X, k moment be X _k=[SOC _ku _p,k] ^t, system state equation and measurement equation as follows:

$\left\{\begin{array}{c}{X}_k=A_{k|k-1}{X}_{k-1}+B_{k-1}{i}_{k-1}+w_{k-1}\\ Y_k=U_{\mathrm{oc}}\left({\mathrm{SOC}}_k\right)-R_0{i}_k-U_{p,k}+v_k\end{array}\right.---\left(13\right)$

Wherein, ν _kmeasurement noise, U _oc(SOC _k) represent battery open circuit voltage U _ocand the nonlinear relationship between SOC is as follows:

U _oc(SOC _k)＝k ₁SOC _k ⁸+k ₂SOC _k ⁷+k ₃SOC _k ⁶+k ₄SOC _k ⁵+

k ₅SOC _k ⁴+k ₆SOC _k ³+k ₇SOC _k ²+k ₈SOC _k+k ₉(14)

The open-circuit voltage U of certain model electrokinetic cell is obtained by on-line identification _oc, and this model electrokinetic cell SOC obtained by traditional means of experiment, use least square method to try to achieve this model electrokinetic cell coefficient k ₁~ k ₉.

AEKF algorithm estimates that SOC process is as follows:

2.1 state estimation:

X _k=[SOC _ku _p,k] ^tthe k estimated value at current time of state

${\hat{X}}_{k|k}={\hat{X}}_{k|k-1}+K_k(Y_{m|k}-{\hat{Y}}_k)---\left(15\right)$

Wherein K _k, expression formula is

Wherein represent state X respectively _k=[SOC _ku _p,k] ^tthe estimated value of moment k of state is being seen based on current time, the estimated value of moment k under based on previous moment state, the estimated value of moment k-1 under based on previous moment state, k is current time, and k-1 is previous moment.Y _m|kthe measured value of k moment battery terminal voltage, the terminal voltage predicted value after upgrading in the k moment,

the discreet value based on k moment SOC under previous moment state, Q _ksystematic procedure noise w _kcovariance.R _ksystem measurements noise v _kcovariance.

Parameter in 2.2 recurrence calculation processes and the renewal of state:

2.2.1 parameter Q _k, R _kupgrade

Wherein, μ _kthe difference of k moment terminal voltage actual value and the terminal voltage predicted value after upgrading, F _kbe the mean value of every L moment corresponding difference, L is self-adapting window.

2.2.2 state upgrade:

Wherein, Q _nfor the rated capacity of electrokinetic cell, η represents efficiency for charge-discharge, and T represents the sampling period.Based on the electrokinetic cell SOC estimating system embodiment of backward difference discrete model

As shown in Figure 2, described microcontroller is embedded microcontroller to this electrokinetic cell SOC estimating system embodiment based on backward difference discrete model, and electrokinetic cell output terminal is connected to voltage sensor and current sensor.This example adopts voltage sensor and the Hall current sensor with isolation features.Voltage sensor and current sensor are connected through analog-to-digital conversion module and embed microcontroller, embed microcontroller and contain low-pass filtering pretreatment module, the on-line parameter identification module of backward difference discrete electrical pool model and AEKF algorithm SOC state estimation module.Embed microcontroller and connect display, be also connected to CAN interface and RS232 interface.Native system is together with electrokinetic cell, be embedded in the equipment using electrokinetic cell, within a sampling period, complete voltage, the collection of electric current, battery model parameter identification and modification and SOC On-line Estimation, gained SOC result shows over the display or is directly sent to the controller local area network of this equipment.

The system embodiment estimated with above-mentioned the electrokinetic cell SOC based on backward difference discrete model, by this method, the gamut of the state of charge SOC of certain model electrokinetic cell to be changed, getting SOC initial value is respectively SOC (0)=1, SOC (0)=0.8, SOC (0)=0.6, SOC (0)=0.4, carry out SOC estimation, the estimated result of gained SOC as shown in Figure 3, wherein, ordinate is SOC value, horizontal ordinate is the time, and great Tu chronomere is second (s), and the little Tu chronomere of upper right is 10 ^-4second.The curve of SOC (0)=1 is solid line, the curve of SOC (0)=0.8 is dash dotted line, the curve of SOC (0)=0.6 is short stroke of dotted line, the curve of SOC (0)=0.4 is dot-and-dash line, and the curve of the SOC value of this model electrokinetic cell of traditional experiment gained is pecked line.Can be seen by large figure, this law gained SOC estimated value conforms to the SOC curve of experiment gained all substantially.Can see in the little figure of upper right, different initial values is in the difference estimating to start to have maximum 0.6 in very short time with SOC experiment value, and the SOC estimated result of initial value that can be different after 0.004 second just reaches unanimity with SOC traditional experiment value.Visible this method initial value is more weak on the impact of SOC estimated accuracy, and this SOC method of estimation has higher estimated accuracy, has practicality.

The gamut change native system of the state of charge SOC of certain model electrokinetic cell is estimated by this method, the experiment experiencing 10 hours obtains SOC estimated value data, the SOC estimated value that this method obtains compares with the SOC of this model electrokinetic cell of traditional experiment gained, corresponding error statistics data result in table 1, the inventive method SOC estimated accuracy reaches 0.45%.

Table 1 SOC error statistics result

Above-described embodiment, be only the specific case further described object of the present invention, technical scheme and beneficial effect, the present invention is not defined in this.All make within scope of disclosure of the present invention any amendment, equivalent replacement, improvement etc., be all included within protection scope of the present invention.

Claims (5)

1., based on the method that the electrokinetic cell SOC of backward difference discrete model estimates, concrete steps are as follows:

The first step, electrokinetic cell backward difference discrete model and parameter identification

1.1. electrokinetic cell model

Adopt the Thevenin model of battery, the polarization resistance R of battery _pwith the polarization capacity C of battery _pformation single order reinforced concrete structure in parallel, represent the polarization reaction of battery, RC both end voltage is U _p(t); Serial connection Ohmage R ₀with Uoc, Uoc are the open-circuit voltage OCV of battery, sampling obtains battery terminal voltage U (t) and flows through ohmic internal resistance R ₀current i (t);

Battery Thevenin model representation is as follows:

$\frac{{\mathrm{dU}}_p\left(t\right)}{\mathrm{dt}}=\frac{U_p\left(t\right)}{R_p{C}_p}+\frac{i\left(t\right)}{C_p}$

U(t)＝U _OC(t)-R ₀i(t)-U _p(t)；

1.2 model discretize and parameter identifications

1.2.1 model discretize

By backward difference method to above-mentioned battery model discretize, obtain difference equation, after arrangement

U(k)-U _OC(k)＝a[U(k-1)-U _OC(k-1)]+bI(k)+cI(k-1)

In formula, Uoc (k) represents the open-circuit voltage in k moment; U (k) is the battery terminal voltage in current k moment; I (k) is the loop current in current k moment; A, b, c are model parameter;

The relation of a, b and c and battery backward difference discrete model parameter is as follows:

$R_0=\frac{c}{a}$

$R_p=\frac{-\mathrm{ab}-c}{a(1-a)}$

$C_p=\frac{{\mathrm{Ta}}^2}{-\mathrm{ab}-c}$

Wherein T is the sampling period;

1.2.2 the parameter identification of battery difference discrete model

Containing the estimated value of least-squares algorithm identification model parameter θ (k) of forgetting factor process as follows:

$\widehat{\mathrm{\θ}}\left(k\right)=\widehat{\mathrm{\θ}}(k-1)+K\left(k\right)e\left(k\right)$

Wherein:

In formula: φ (k) is data vector, θ (k) is estimated parameter vector, the prediction error that e (k) is U (k), initial value with P (0) rule of thumb assignment, for the estimated value of θ (k), λ is forgetting factor, λ=0.95 ~ 1;

Tried to achieve the value of a, b, c by FFRLS algorithm, thus obtain model parameter R ₀, R _p, C _p, U _oCvalue;

Second step, to estimate based on the battery state of charge SOC of adaptive extended kalman filtering AEKF

Choose SOC and electric capacity C _pterminal voltage be state variable X, namely the state X in k moment, is expressed as X _k=[SOC _ku _p,k] ^t, system state equation and measurement equation as follows:

$\left\{\begin{array}{c}{X}_k=A_{k|k-1}{X}_{k-1}{B}_{k-1}{i}_{k-1}+w_{k-1}\\ Y_k=U_{\mathrm{oc}}\left({\mathrm{SOC}}_k\right)-R_0{i}_k-U_{p,k}+v_k\end{array}\right.$

Wherein, ν _kmeasurement noise, U _oc(SOC _k) represent battery open circuit voltage U _ocand the nonlinear relationship between SOC is as follows:

U _oc(SOC _k)＝k ₁SOC _k ⁸+k ₂SOC _k ⁷+k ₃SOC _k ⁶+k ₄SOC _k ⁵+

k ₅SOC _k ⁴+k ₆SOC _k ³+k ₇SOC _k ²+k ₈SOC _k+k ₉

Certain the model electrokinetic cell open-circuit voltage U obtained by on-line identification _octhis model electrokinetic cell SOC obtained with experiment, uses least square method to try to achieve this model electrokinetic cell coefficient k ₁~ k ₉;

AEKF algorithm estimates that SOC process is as follows:

2.1 state estimation:

X _k=[SOC _ku _p,k] ^tthe k estimated value at current time of state

${\hat{X}}_{k|k}={\hat{X}}_{k|k-1}+K_k(Y_{m|k}-{\hat{Y}}_k)$

Wherein K _k, expression formula is

Wherein represent state X respectively _k=[SOC _ku _p,k] ^tthe estimated value of moment k of state is being seen based on current time, the estimated value of moment k under based on previous moment state, the estimated value of moment k-1 under based on previous moment state, k is current time, and k-1 is previous moment; Y _m|kthe measured value of k moment battery terminal voltage, the terminal voltage predicted value after upgrading in the k moment, the discreet value based on k moment SOC under previous moment state, Q _ksystematic procedure noise w _kcovariance, R _ksystem measurements noise v _kcovariance;

Parameter in 2.2 recurrence calculation processes and the renewal of state:

2.2.1 parameter Q _k, R _kupgrade

Wherein, μ _kthe difference of terminal voltage predicted value after upgrading in the k moment and terminal voltage actual value, F _kbe the mean value of every L moment corresponding difference, L is self-adapting window;

2.2.2 state upgrade:

Wherein, Q _nfor the rated capacity of electrokinetic cell, η represents efficiency for charge-discharge.

2. the electrokinetic cell SOC estimating system based on backward difference discrete model of method design that the electrokinetic cell SOC based on backward difference discrete model according to claim 1 estimates comprises the display of microcontroller and connection thereof, electrokinetic cell output terminal is connected to voltage sensor and current sensor, voltage sensor and current sensor are connected embedding microcontroller through analog-to-digital conversion module, it is characterized in that:

Described microcontroller is embedded microcontroller, embed microcontroller and contain low-pass filtering pretreatment module, the on-line parameter identification module of backward difference discrete electrical pool model and AEKF algorithm SOC state estimation module, gained battery state of charge estimated value result shows over the display or is directly sent to the controller local area network of this equipment.

3. the electrokinetic cell SOC estimating system based on backward difference discrete model according to claim 2, is characterized in that:

Native system, together with electrokinetic cell, is embedded in the equipment using electrokinetic cell.

4. the electrokinetic cell SOC estimating system based on backward difference discrete model according to claim 2, is characterized in that:

Described microcontroller is connected to CAN interface and/or RS232 interface.

5. the electrokinetic cell SOC estimating system based on backward difference discrete model according to claim 2, is characterized in that:

Described voltage sensor is the voltage sensor with isolation features; Described current sensor is Hall current sensor.