CN117074980A - Method for estimating SOC of lithium iron phosphate battery based on ampere-hour integral reference compensation - Google Patents

Abstract

An SOC estimation method of a lithium iron phosphate battery based on ampere-hour integral reference compensation belongs to the technical field of battery state of charge estimation. The invention aims at the problem that the result of estimating the SOC of the lithium iron phosphate battery by adopting a Kalman filtering algorithm under the condition of larger voltage sampling error is unreliable. The method comprises the steps of calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an ampere-hour integration method and an EKF-AUKF algorithm respectively; taking the SOC value calculation result of the ampere-hour integration method as a reference value, and selecting the SOC value of the two algorithms at the moment of 100s to make a difference to obtain a constant deviation; taking the difference value of the SOC values of the two algorithms after 100s as a preliminary difference value, and obtaining an approximate deviation value by combining with a constant deviation; when the approximate deviation value is larger than the set deviation threshold value, the current EKF-AUKF algorithm SOC value is compensated by adopting different weight coefficients according to the comparison result of the EKF-AUKF algorithm SOC value and the preset demarcation value, so that a compensation result is obtained. The method is used for estimating the SOC of the battery.

Description

Method for estimating SOC of lithium iron phosphate battery based on ampere-hour integral reference compensation

Technical Field

The invention relates to an SOC estimation method of a lithium iron phosphate battery based on ampere-hour integral reference compensation, and belongs to the technical field of battery state of charge estimation.

Background

The lithium battery is widely applied to the fields of electric automobiles, energy storage power stations, electromagnetic ejection and the like due to high safety, long cycle life and high energy density.

The State of Charge (SOC) is one of the indicators of the battery management system, and reflects the remaining battery usage capacity. The method is applied to the electric vehicle, and the lithium battery SOC is directly related to the driving mileage of the vehicle, so that accurate estimation of the lithium battery SOC has important significance. However, the SOC of the lithium battery cannot be directly measured, and online estimation of the SOC needs to be realized through a corresponding algorithm.

Currently, the lithium battery state of charge estimation methods mainly include an open loop method, a data driving method and a model-based state estimation method. The open loop method is simple to realize and wide in application, but the open loop method does not have a feedback link, cannot eliminate the influence caused by the initial error of the SOC and the accumulated error of the current, and has low estimation precision; the data driving method has the characteristics of self-learning and self-adaption, takes parameters such as rated capacity, voltage, current, temperature and the like of the battery as input, takes the SOC as output, and can describe the characteristics of a complex nonlinear system; however, the accuracy of this method depends on the input parameters, the amount and quality of training data; the state estimation method based on the model is mainly a kalman filter and a derivative filter thereof, and considering the nonlinear characteristics of the lithium battery, a number of state estimation algorithms suitable for nonlinear systems, such as extended kalman filter (extended Kalman filter, EKF), unscented kalman filter (unscented Kalman filter, UKF) and Particle Filter (PF), are proposed to improve the estimation accuracy of the SOC of the lithium battery. These algorithms assume that the process noise and the measurement noise are gaussian white noise independent of each other, but in reality the noise tends to be irregular noise. Thus, a number of noise adaptation algorithms, such as adaptive EKF, adaptive UKF, adaptive PF, etc., have been proposed accordingly to mitigate the effects of irregular noise. The parameters of the lithium battery are affected by the changes of factors such as the battery temperature, the discharge multiplying power, the aging and the like, so that the parameters obtained by adopting a fixed parameter or off-line mode may not be practical. Thus, a double extended kalman filter, a double unscented kalman filter, etc. are proposed to improve the SOC estimation accuracy.

The model-based state estimation algorithm mainly focuses on model accuracy and algorithm accuracy, and influences of voltage measurement noise on estimation accuracy are ignored. The open circuit voltage of the lithium iron phosphate battery is quite flat, and a small voltage error can generate a large SOC estimation error. When a large voltage sampling error exists, a Kalman filtering algorithm is adopted to estimate the SOC of the lithium iron phosphate battery, so that a large error exists.

Disclosure of Invention

Aiming at the problem that under a larger voltage sampling error, the result of estimating the SOC of the lithium iron phosphate battery by adopting a Kalman filtering algorithm is unreliable, the invention provides the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation.

The invention relates to a lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, which comprises the following steps,

step one: constructing a discrete state space equation set of the lithium iron phosphate battery according to an equivalent circuit model of the lithium iron phosphate battery;

step two: calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an ampere-hour integration method;

meanwhile, calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an EKF-AUKF algorithm, wherein the method comprises the steps of identifying real-time equivalent circuit parameters of an equivalent circuit model by adopting the EKF algorithm, and calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an AUKF algorithm based on a discrete state space equation set updated by the real-time equivalent circuit parameters; then feeding back the obtained SOC value and Kalman gain to an EKF algorithm to serve as a basis for identifying real-time equivalent circuit parameters of the equivalent circuit model at the next sampling moment;

step three: taking the SOC value calculation result of the ampere-hour integration method as a reference value, and selecting the difference between the SOC value of the EKF-AUKF algorithm at the moment of 100s and the SOC value of the ampere-hour integration method to obtain a result as constant deviation;

step four: taking the 100s moment as a starting point, and taking the difference between the SOC value of the EKF-AUKF algorithm and the SOC value of the ampere-hour integration method at each sampling moment to obtain a preliminary difference value; then taking the absolute value of the result obtained by the difference between the preliminary difference value and the constant deviation as the approximate deviation value of the SOC value and the SOC true value of the EKF-AUKF algorithm;

step five: setting a deviation threshold;

if the current approximate deviation value is smaller than the deviation threshold value, taking the current EKF-AUKF algorithm SOC value as the current SOC estimation result of the lithium iron phosphate battery;

if the current approximate deviation value is larger than or equal to the deviation threshold value, then different weight coefficients are adopted according to the comparison result of the SOC value of the EKF-AUKF algorithm and the preset demarcation value, the current SOC value of the EKF-AUKF algorithm is compensated based on the current time-of-day integration method SOC value, and the compensation result is used as the SOC estimation result of the current lithium iron phosphate battery.

According to the method for estimating the SOC of the lithium iron phosphate battery based on ampere-hour integral reference compensation, in the first step, the input-output relationship of an equivalent circuit model of the lithium iron phosphate battery is as follows:

u in ₁ Polarization internal resistance terminal voltage for battery reaction, R ₁ For the battery reaction polarization internal resistance, C ₁ Polarization capacitor for battery reactionI is battery current, U ₂ For the voltage of the internal resistance terminal of the concentration polarization of the battery, R ₂ For the concentration polarization internal resistance of the battery, C ₂ U is the concentration polarization capacitance of the battery _t For battery terminal voltage, U _oc R is the open circuit voltage of the battery ₀ Is the ohmic internal resistance of the battery.

According to the method for estimating the SOC of the lithium iron phosphate battery based on ampere-hour integral reference compensation, in the first step, a discrete state space equation set of the constructed lithium iron phosphate battery is as follows:

z in _k+1 For the SOC value at k+1 moment obtained by adopting the EKF-AUKF algorithm, z _k In order to obtain a k-moment SOC value by adopting an EKF-AUKF algorithm, eta is coulomb efficiency, deltat is sampling interval time, and C _n Maximum available capacity for the battery; u (U) _1,k+1 U at time k+1 ₁ ，α ₁ For the battery reaction polarization time index, alpha ₁ ＝exp(-Δt/R ₁ C ₁ )，U _1,k U at time k ₁ ，U _2,k+1 U at time k+1 ₂ ，α ₂ For the cell concentration polarization time index, alpha ₂ ＝exp(-Δt/R ₂ C ₂ )，U _2,k U at time k ₂ ，I _k Is I, w at time k _k U is the process noise at time k _t,k For time k U _t ，U _oc,k For time k U _oc ，v _k Noise is measured for time k.

According to the method for estimating the SOC of the lithium iron phosphate battery based on the ampere-hour integral reference compensation, in the second step, the method for calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting the ampere-hour integral method comprises the following steps:

z _Ah,k+1 ＝z _Ah,k -ηI _k Δt/C _n ， (3)

z in _Ah,k+1 For the SOC value at k+1 moment obtained by adopting an ampere-hour integration method, z _Ah,k The SOC value at k moment is obtained by adopting an ampere-hour integration method.

According to the method for estimating the SOC of the lithium iron phosphate battery based on the ampere-hour integral reference compensation, in the second step, the method for calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an EKF-AUKF algorithm comprises the following steps:

first, a state space equation describing the variation characteristics of the equivalent circuit parameters is constructed:

representing the equivalent circuit parameters in the k-time equivalent circuit model as a circuit parameter vector theta _k ：

θ _k ＝[R ₀ R ₁ C ₁ R ₂ C ₂ ] ^T ， (4)

Regarding the equivalent circuit parameters as constants with disturbance, and obtaining a state space equation:

in the middle ofEstimated value of circuit parameter vector for k+1 time, ">For the estimated value of the circuit parameter vector at the moment k, r _k White noise r is the white noise at time k _k Fitting Gaussian distribution-> Is r _k Error covariance matrix of (a);

d in _k Represents the battery terminal voltage U at time k _t,k ，g(x _k ,u _k ,θ _k ) Representing measurement equation U containing parameters _oc,k -U _1,k -U _2,k -I _k R ₀ ，x _k As a state vector, the k-time SOC value z obtained by adopting the EKF-AUKF algorithm is represented _k+1 、U _1,k+1 And U _2,k ；u _k Representing the battery current I at time k _k ，λ _k Represents the measurement noise at time k, the measurement noise lambda _k Conforming to gaussian distribution Is lambda _k Error covariance matrix of (a) is obtained.

According to the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, in the third step, the constant deviation is expressed as d:

d＝SOC _{EKF-AUKF,100s} -SOC _Ah,100s ， (33)

SOC in _{EKF-AUKF,100s} SOC value and SOC of 100s time EKF-AUKF algorithm _Ah,100s The SOC value is obtained by an ampere-hour integration method at the moment of 100 s.

According to the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, in the fourth step, the preliminary difference value is expressed as D _k ：

D _k ＝SOC _EKF-AUKF,k -SOC _Ah,k ， (34)

SOC in _EKF-AUKF,k After 100s, the SOC value and the SOC of the EKF-AUKF algorithm at k moment _Ah,k After 100s, the SOC value of the ampere-hour integration method at the k moment;

the approximate deviation value is expressed as Δd:

Δd＝|D _k -d|。 (35)

according to the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, in the fifth step, a deviation threshold value is set as delta;

if the current approximate deviation value delta d is greater than or equal to the deviation threshold delta, compensating the current EKF-AUKF algorithm SOC value to obtain a k-moment compensation result SOC _k The method comprises the following steps:

SOC _k ＝SOC _EKF-AUKF,k -pΔd， (36)

wherein p is a weight coefficient of the formula,

combining formulas (34) and (35) to obtain k-moment compensation result SOC _k Is the final expression of (2):

SOC _k ＝(1-p)SOC _EKF-AUKF,k +p(SOC _Ah,k +d)。 (37)

according to the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, in the fifth step, if the current approximate deviation value is greater than or equal to the deviation threshold value, the current EKF-AUKF algorithm SOC value is compensated:

setting the demarcation value to be 0.3;

if the SOC value of the current EKF-AUKF algorithm is smaller than or equal to 0.3, the weight coefficient p is selected to be 0.15; if the SOC value of the current EKF-AUKF algorithm is larger than 0.3, selecting a weight coefficient p to be 0.85; deviation threshold δ=0.03.

The invention has the beneficial effects that: the method fully considers the influence of the accuracy of the model and the accuracy of the algorithm, and adopts the EKF-AUKF algorithm to realize the accurate estimation of the SOC of the lithium iron phosphate battery. Under the condition that the SOC estimation result of the EKF-AUKF algorithm is influenced by a larger voltage sampling error, taking the SOC estimation result of the ampere-hour integration method as a reference, utilizing the characteristic of stable SOC change to calculate the approximate difference value between the SOC estimation result of the EKF-AUKF algorithm and the true value, further utilizing the compensation coefficient to compensate the SOC estimation result of the EKF-AUKF algorithm, improving the accuracy and the robustness of the SOC estimation, eliminating the influence of the initial error and the accumulated error of the ampere-hour integration method SOC to a certain extent, and not obviously increasing the calculated amount.

Drawings

FIG. 1 is a flow chart of a lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation according to the invention; in the figure, t is time;

fig. 2 is an equivalent circuit model of a lithium iron phosphate battery.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other.

The invention is further described below with reference to the drawings and specific examples, which are not intended to be limiting.

The invention provides a lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, which is shown in the accompanying drawings in the specification 1 and 2,

step one: constructing a discrete state space equation set of the lithium iron phosphate battery according to an equivalent circuit model of the lithium iron phosphate battery;

step two: calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an ampere-hour integration method;

meanwhile, calculating the SOC value of the lithium iron phosphate battery at each sampling time by adopting an EKF-AUKF algorithm, wherein the method comprises the steps of identifying real-time equivalent circuit parameters of an equivalent circuit model by adopting the EKF algorithm, and calculating the SOC value of the lithium iron phosphate battery at each sampling time by adopting an AUKF (adaptive unscented Kalman filtering) algorithm based on a discrete state space equation set updated by the real-time equivalent circuit parameters; then feeding back the obtained SOC value and Kalman gain to an EKF algorithm to serve as a basis for identifying real-time equivalent circuit parameters of the equivalent circuit model at the next sampling moment;

step three: taking the SOC value calculation result of the ampere-hour integration method as a reference value, and selecting the difference between the SOC value of the EKF-AUKF algorithm at the moment of 100s and the SOC value of the ampere-hour integration method to obtain a result as constant deviation;

step four: taking the 100s moment as a starting point, and taking the difference between the SOC value of the EKF-AUKF algorithm and the SOC value of the ampere-hour integration method at each sampling moment to obtain a preliminary difference value; then taking the absolute value of the result obtained by the difference between the preliminary difference value and the constant deviation as the approximate deviation value of the SOC value and the SOC true value of the EKF-AUKF algorithm;

step five: setting a deviation threshold;

if the current approximate deviation value is smaller than the deviation threshold value, taking the current EKF-AUKF algorithm SOC value as the current SOC estimation result of the lithium iron phosphate battery; no compensation is required at this time;

if the current approximate deviation value is greater than or equal to the deviation threshold value, compensating; and then, according to the comparison result of the SOC value of the EKF-AUKF algorithm and the preset demarcation value, adopting different weight coefficients, compensating the current SOC value of the EKF-AUKF algorithm based on the current time-of-day integration method SOC value, and taking the compensation result as the SOC estimation result of the current lithium iron phosphate battery. And compensating the approximate deviation value to the SOC estimation result of the EKF-AUKF algorithm as the final estimation value of the current SOC of the lithium battery to be measured under the condition that the approximate deviation value is not smaller than the deviation threshold value.

Further, in the first step, according to fig. 2, the relationship between input and output of the equivalent circuit model of the lithium iron phosphate battery can be obtained according to kirchhoff's law:

u in ₁ Polarization internal resistance terminal voltage for battery reaction, R ₁ For the battery reaction polarization internal resistance, C ₁ The battery reaction polarization capacitor, I is battery current, U ₂ For the voltage of the internal resistance terminal of the concentration polarization of the battery, R ₂ For the concentration polarization internal resistance of the battery, C ₂ U is the concentration polarization capacitance of the battery _t For battery terminal voltage, U _oc R is the open circuit voltage of the battery ₀ Is the ohmic internal resistance of the battery.

Referring to fig. 2, the equivalent circuit model of the lithium iron phosphate battery includes a second-order resistance-capacitance equivalent circuit composed of a voltage source, ohmic internal resistance, reactive polarized capacitance, concentration polarized internal resistance and concentration polarized capacitance. The reaction polarization internal resistance and the reaction polarization capacitance are used for describing the internal reaction polarization phenomenon of the battery, and the concentration polarization internal resistance and the concentration polarization capacitance are used for describing the internal concentration polarization phenomenon of the battery.

In the first step of this embodiment, the discrete state space equation set of the constructed lithium iron phosphate battery is:

z in _k+1 For the SOC value at k+1 moment obtained by adopting the EKF-AUKF algorithm, z _k In order to obtain a k-moment SOC value by adopting an EKF-AUKF algorithm, eta is coulomb efficiency, deltat is sampling interval time, and C _n Maximum available capacity for the battery; u (U) _1,k+1 U at time k+1 ₁ ，α ₁ For the battery reaction polarization time index, alpha ₁ ＝exp(-Δt/R ₁ C ₁ )，U _1,k U at time k ₁ ，U _2,k+1 U at time k+1 ₂ ，α ₂ For the cell concentration polarization time index, alpha ₂ ＝exp(-Δt/R ₂ C ₂ )，U _2,k U at time k ₂ ，I _k Is I, w at time k _k The noise is process noise at the moment k and is used for representing current measurement errors and state equation errors; u (U) _t,k For time k U _t ，U _oc,k For time k U _oc ，v _k The noise is measured at the moment k and used for representing the voltage measurement error and the output equation error.

In the second step, the method for calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting the discrete form ampere-hour integration method comprises the following steps:

z _Ah,k+1 ＝z _Ah,k -ηI _k Δt/C _n ， (3)

z in _Ah,k+1 For the SOC value at k+1 moment obtained by adopting an ampere-hour integration method, z _Ah,k The SOC value at k moment is obtained by adopting an ampere-hour integration method.

Meanwhile, in the second step, the method for calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting the EKF-AUKF algorithm comprises the following steps:

firstly, on-line identification of equivalent circuit parameters of a lithium iron phosphate battery by adopting an EKF algorithm, and the change characteristics of the equivalent circuit parameters need to be described by constructing a state space equation:

representing the equivalent circuit parameters in the k-time equivalent circuit model as a circuit parameter vector theta _k ：

θ _k ＝[R ₀ R ₁ C ₁ R ₂ C ₂ ] ^T ， (4)

The parameters of the lithium battery change slowly and can change obviously after a plurality of discharge cycles, so that the parameters of the equivalent circuit can be regarded as a constant with disturbance, and therefore, the state space equation can be written as follows:

in the middle ofEstimated value of circuit parameter vector for k+1 time, ">For the estimated value of the circuit parameter vector at the moment k, r _k White noise r is the white noise at time k _k Fitting Gaussian distribution-> Is r _k Error covariance matrix of (a);

d in _k Represents the battery terminal voltage U at time k _t,k ，g(x _k ,u _k ,θ _k ) Representing measurement equation U containing parameters _oc,k -U _1,k -U _2,k -I _k R ₀ ，x _k As a state vector, the k-time SOC value z obtained by adopting the EKF-AUKF algorithm is represented _k+1 、U _1,k+1 And U _2,k ；u _k Representing the battery current I at time k _k ，λ _k Represents the measurement noise at time k, the measurement noise lambda _k Conforming to gaussian distribution Is lambda _k Error covariance matrix of (a) is obtained.

In the second step, after the state space equation (5) is constructed, the SOC value at each sampling time is calculated:

and (3) a step of: initializing:

1. initializing a parameter vector and an error covariance of the parameter vector:

in the middle ofFor initial estimation of circuit parameter vector, θ ₀ E represents an average value for the initial value of the circuit parameter vector;

initial error covariance for the parameter vector;

2. initializing a state vector and an error covariance of the state vector:

in the middle ofFor initial estimation of state vector, x ₀ Is the initial value of the state vector; />Is the first of the state vectorInitial error covariance;

and II: time update of parameter estimation:

in the middle ofFor the estimated value of the circuit parameter vector at time k, < >>For the circuit parameter vector a priori estimate at time k +1,for the k moment parameter vector error covariance, +.>A priori estimated value of the parameter vector error covariance at the time of k+1;

thirdly,: time update of state estimation:

1) Weights of the unscented transform sampling points are calculated:

in the middle ofAs initial values of state mean weight factors, lambda is a scale factor for reducing global prediction errors, and n is a battery system state dimension;

for the initial value of the error covariance weighting factor, beta is a higher-order term error adjustment coefficient, alpha is a control coefficient of sigma point distributed around a state mean point,W _i ^m for the ith state mean weight factor, W _i ^c For the ith error covariance weighting factor, κ is a second order scale parameter used to ensure +.>Setting; />The error covariance of the state vector at the moment k;

as an example: n=3, α=0.01, β=2, and κ=0.

2) 2n+1 sigma points at k time are calculated:

in the middle ofFor the 0 th state vector value at time k, < ->For the state vector estimate at time k, +.>The i state vector value at the k moment;

3) Updating a priori state of sigma points:

in the middle ofFor the i-th state vector estimate at time k +1,

representation->

Then:

y _k ＝U _t,k ，g(x _k ,u _k ,θ _k )＝U _oc,k -U _1,k -U _2,k -I _k R ₀ ；

in which y _k As an output value of the measurement equation;

4) System state and system state error covariance a priori estimates:

in the middle ofA state vector priori estimated value at the moment k+1; />An error covariance priori estimated value of the state vector at the moment k+1;

fourth, the method comprises the following steps: measurement update:

1) Output sigma point calculation:

in the middle ofOutputting an estimated value for a measurement equation corresponding to the ith sigma point at the k+1 moment;

2) Cell terminal voltage average and terminal voltage average error covariance calculation:

in the middle ofOutputting a priori estimated value for the k+1 moment measurement equation; p (P) _yy,k+1 For the mean value error covariance of the voltage at the output end of the measurement equation at the moment k+1, V _k Is the output error covariance;

3) UpdatingAnd->Cross covariance between:

p in the formula _xy,k+1 Time k+1And->Cross covariance between;

4) Calculating Kalman gain:

in the middle ofKalman gain at time k+1;

5) State and error covariance update:

6) Noise covariance update:

in e _k+1 Is the noise covariance at time k+1, m _k+1 E is a sliding average value of terminal voltage information _i Terminal voltage innovation; l (L) _w For moving window size;

process noise covariance for state vector at time k+1,/->Measurement noise covariance for the state vector at time k+1;

fifth step: parameter measurement update:

1) Calculating a parameter gain matrix:

in the middle ofEstimating a gain matrix for the k+1 time instant parameter, < >>Estimating an output matrix for the k+1 time parameter;

2) Parameter and error covariance update:

wherein the method comprises the steps ofThe iterative initial value is 0, which can be obtained by the following three iterative partial differential equations:

in the third step of the embodiment, the SOC calculation value of the ampere-hour integration method is used as a reference, and the difference between the EKF-AUKF algorithm at 100s and the SOC estimation result of the ampere-hour integration method is used as a constant deviation;

in the initial stage of estimation, according to the convergence rate of the algorithm, the correction logic delay is acted for 100 seconds, so that the SOC estimation value of the adaptive algorithm can be ensured to approach to a true value. The self-adaptive algorithm is not corrected in the time delay period, when the time delay time point is reached, the difference value between the self-adaptive algorithm and the ampere-hour integration method is calculated, the difference value is used as the constant deviation of the self-adaptive algorithm and the ampere-hour integration method, and the constant deviation is expressed as d:

d＝SOC _{EKF-AUKF,100s} -SOC _Ah,100s ， (33)

in the middle ofSOC _{EKF-AUKF,100s} SOC value and SOC of 100s time EKF-AUKF algorithm _Ah,100s The SOC value is obtained by an ampere-hour integration method at the moment of 100 s.

Before 100s, the compensation logic was disabled and the ampere-hour integration method and the EKF-AUKF algorithm independently estimated the lithium battery SOC.

In step four, the preliminary difference is denoted as D _k ：

D _k ＝SOC _EKF-AUKF,k -SOC _Ah,k ， (34)

SOC in _EKF-AUKF,k After 100s, the SOC value and the SOC of the EKF-AUKF algorithm at k moment _Ah,k After 100s, the SOC value of the ampere-hour integration method at the k moment;

the approximate deviation value is expressed as Δd:

Δd＝|D _k -d|。 (35)

finally, in the fifth step, setting the deviation threshold value as delta;

when the delta d is smaller than delta, the estimation error of the self-adaptive algorithm is within a limited range, and compensation is not needed; if the current approximate deviation value delta d is greater than or equal to the deviation threshold delta, compensating the current EKF-AUKF algorithm SOC value to obtain a k-moment compensation result SOC _k The method comprises the following steps:

SOC _k ＝SOC _EKF-AUKF,k -pΔd， (36)

wherein p is a weight coefficient of the formula,

the estimation error of the ampere-hour integration method is gradually accumulated along with the progress of the estimation process, and the self-adaptive algorithm estimates very accurately at the end stage of battery discharge, so that if the difference value of the estimation results of the two algorithms is used for compensation, larger error is introduced due to the inaccuracy of the estimation results of the ampere-hour integration method. The compensation value is thus changed by adjusting the weighting coefficients, which may be referred to as compensation coefficients. In the area with small ampere-hour integral SOC estimation error, the compensation coefficient is large; and in the area with large ampere-hour integral SOC estimation error, the compensation coefficient is small.

Combining formulas (34) and (35) to obtain k-moment compensation result SOC _k Is the final expression of (2):

SOC _k ＝(1-p)SOC _EKF-AUKF,k +p(SOC _Ah,k +d)。 (37)

wherein p is more than or equal to 0 and less than or equal to 1.

As an example, in the fifth step, if the current approximate deviation value is greater than or equal to the deviation threshold value, the current EKF-AUKF algorithm SOC value is compensated:

setting the demarcation value to be 0.3; and taking P as a demarcation point by taking an SOC estimated value of 0.3 as a demarcation point according to the SOC estimated characteristics of the self-adaptive algorithm and the ampere-hour integration method.

If the SOC value of the current EKF-AUKF algorithm is smaller than or equal to 0.3, the weight coefficient p is selected to be 0.15; if the SOC value of the current EKF-AUKF algorithm is larger than 0.3, selecting a weight coefficient p to be 0.85; deviation threshold δ=0.03.

Considering that the SOC estimation error of the ampere-hour integration method gradually increases along with the current sampling error, the SOC estimation result of the EKF-AUKF algorithm approaches to a true value when discharging is finished, the 0.3 of the SOC estimation value of the EKF-AUKF algorithm is taken as a boundary, and different compensation weight coefficients are selected on two sides, so that the stability of the ampere-hour integration method estimation is fully utilized, the influence caused by the ampere-hour integration accumulated error is eliminated, and the compensated SOC can be regarded as the current accurate SOC.

The invention also provides electronic equipment, which comprises a processor and a memory, wherein the memory stores a computer control program, and the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation can be realized when the processor processes the program stored in the memory.

The invention also provides a computer readable storage medium for storing a computer control program, and the computer control program can realize the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation.

Although the invention herein has been described with reference to particular embodiments, it is to be understood that these embodiments are merely illustrative of the principles and applications of the present invention. It is therefore to be understood that numerous modifications may be made to the illustrative embodiments and that other arrangements may be devised without departing from the spirit and scope of the present invention as defined by the appended claims. It should be understood that the different dependent claims and the features described herein may be combined in ways other than as described in the original claims. It is also to be understood that features described in connection with separate embodiments may be used in other described embodiments.

Claims (10)

1. An ampere-hour integral reference compensation-based lithium iron phosphate battery SOC estimation method is characterized by comprising the steps of,

step one: constructing a discrete state space equation set of the lithium iron phosphate battery according to an equivalent circuit model of the lithium iron phosphate battery;

step two: calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an ampere-hour integration method;

meanwhile, calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an EKF-AUKF algorithm, wherein the method comprises the steps of identifying real-time equivalent circuit parameters of an equivalent circuit model by adopting the EKF algorithm, and calculating the SOC value of the lithium iron phosphate battery at each sampling moment by adopting an AUKF algorithm based on a discrete state space equation set updated by the real-time equivalent circuit parameters; then feeding back the obtained SOC value and Kalman gain to an EKF algorithm to serve as a basis for identifying real-time equivalent circuit parameters of the equivalent circuit model at the next sampling moment;

step three: taking the SOC value calculation result of the ampere-hour integration method as a reference value, and selecting the difference between the SOC value of the EKF-AUKF algorithm at the moment of 100s and the SOC value of the ampere-hour integration method to obtain a result as constant deviation;

step four: taking the 100s moment as a starting point, and taking the difference between the SOC value of the EKF-AUKF algorithm and the SOC value of the ampere-hour integration method at each sampling moment to obtain a preliminary difference value; then taking the absolute value of the result obtained by the difference between the preliminary difference value and the constant deviation as the approximate deviation value of the SOC value and the SOC true value of the EKF-AUKF algorithm;

step five: setting a deviation threshold;

if the current approximate deviation value is smaller than the deviation threshold value, taking the current EKF-AUKF algorithm SOC value as the current SOC estimation result of the lithium iron phosphate battery;

if the current approximate deviation value is larger than or equal to the deviation threshold value, then different weight coefficients are adopted according to the comparison result of the SOC value of the EKF-AUKF algorithm and the preset demarcation value, the current SOC value of the EKF-AUKF algorithm is compensated based on the current time-of-day integration method SOC value, and the compensation result is used as the SOC estimation result of the current lithium iron phosphate battery.

2. The method for estimating SOC of a lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 1, wherein in the first step, an input-output relationship of an equivalent circuit model of the lithium iron phosphate battery is:

u in ₁ Polarization internal resistance terminal voltage for battery reaction, R ₁ For the battery reaction polarization internal resistance, C ₁ The battery reaction polarization capacitor, I is battery current, U ₂ For the voltage of the internal resistance terminal of the concentration polarization of the battery, R ₂ For the concentration polarization internal resistance of the battery, C ₂ U is the concentration polarization capacitance of the battery _t For battery terminal voltage, U _oc R is the open circuit voltage of the battery ₀ Is the ohmic internal resistance of the battery.

3. The method for estimating SOC of a lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 2, wherein in the first step, a discrete state space equation set of the constructed lithium iron phosphate battery is:

z in _k+1 For the SOC value at k+1 moment obtained by adopting the EKF-AUKF algorithm, z _k In order to obtain a k-moment SOC value by adopting an EKF-AUKF algorithm, eta is coulomb efficiency, deltat is sampling interval time, and C _n Maximum available capacity for the battery; u (U) _1,k+1 U at time k+1 ₁ ，α ₁ For the battery reaction polarization time index, alpha ₁ ＝exp(-Δt/R ₁ C ₁ )，U _1,k U at time k ₁ ，U _2,k+1 Time k+1U of (2) ₂ ，α ₂ For the cell concentration polarization time index, alpha ₂ ＝exp(-Δt/R ₂ C ₂ )，U _2,k U at time k ₂ ，I _k Is I, w at time k _k U is the process noise at time k _t,k For time k U _t ，U _oc,k For time k U _oc ，v _k Noise is measured for time k.

4. The method for estimating the SOC of the lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 3, wherein in the second step, the method for calculating the SOC value of the lithium iron phosphate battery at each sampling time by using the ampere-hour integral method comprises the following steps:

z _Ah,k+1 ＝z _Ah,k -ηI _k Δt/C _n ， (3)

z in _Ah,k+1 For the SOC value at k+1 moment obtained by adopting an ampere-hour integration method, z _Ah,k The SOC value at k moment is obtained by adopting an ampere-hour integration method.

5. The SOC estimation method of the lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 3, wherein in the second step, the method for calculating the SOC value of the lithium iron phosphate battery at each sampling time by using the EKF-AUKF algorithm comprises the following steps:

first, a state space equation describing the variation characteristics of the equivalent circuit parameters is constructed:

representing the equivalent circuit parameters in the k-time equivalent circuit model as a circuit parameter vector theta _k ：

θ _k ＝[R ₀ R ₁ C ₁ R ₂ C ₂ ] ^T ， (4)

Regarding the equivalent circuit parameters as constants with disturbance, and obtaining a state space equation:

in the middle ofEstimated value of circuit parameter vector for k+1 time, ">For the estimated value of the circuit parameter vector at the moment k, r _k White noise r is the white noise at time k _k Fitting Gaussian distribution-> Is r _k Error covariance matrix of (a);

d in _k Represents the battery terminal voltage U at time k _t,k ，g(x _k ,u _k ,θ _k ) Representing measurement equation U containing parameters _oc,k -U _1,k -U _2,k -I _k R ₀ ，x _k As a state vector, the k-time SOC value z obtained by adopting the EKF-AUKF algorithm is represented _k+1 、U _1,k+1 And U _2,k ；u _k Representing the battery current I at time k _k ，λ _k Represents the measurement noise at time k, the measurement noise lambda _k Conforming to gaussian distribution Is lambda _k Error covariance matrix of (a) is obtained.

6. The method for estimating SOC of lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 5, wherein in the second step, after the state space equation (5) is constructed, the calculation of SOC value at each sampling time is performed:

and (3) a step of: initializing:

1. initializing a parameter vector and an error covariance of the parameter vector:

in the middle ofFor initial estimation of circuit parameter vector, θ ₀ E represents an average value for the initial value of the circuit parameter vector;

initial error covariance for the parameter vector;

2. initializing a state vector and an error covariance of the state vector:

in the middle ofFor initial estimation of state vector, x ₀ Is the initial value of the state vector; />Initial error covariance for the state vector;

and II: time update of parameter estimation:

in the middle ofFor the estimated value of the circuit parameter vector at time k, < >>A priori estimate of the circuit parameter vector for time k+1,/for time k+1>For the k moment parameter vector error covariance, +.>A priori estimated value of the parameter vector error covariance at the time of k+1;

thirdly,: time update of state estimation:

1) Weights of the unscented transform sampling points are calculated:

w in the formula ₀ ^m As initial values of state mean weight factors, lambda is a scale factor for reducing global prediction errors, and n is a battery system state dimension;

W ₀ ^c for the initial value of the error covariance weight factor, beta is a higher-order term error adjustment coefficient, alpha is a control coefficient of sigma point distributed around a state mean point, and W _i ^m For the ith state mean weight factor, W _i ^c For the ith error covariance weighting factor, κ is the second order scale parameter, usingTo ensure thatSetting; />The error covariance of the state vector at the moment k;

2) 2n+1 sigma points at k time are calculated:

in the middle ofFor the 0 th state vector value at time k, < ->For the state vector estimate at time k, +.>The i state vector value at the k moment;

3) Updating a priori state of sigma points:

in the middle ofFor the i-th state vector estimate at time k +1,

representation->

Then:

x _k+1 ＝[z _k+1 U _1,k+1 U _2,k+1 ] ^T ，

y _k ＝U _t,k ，g(x _k ,u _k ,θ _k )＝U _oc,k -U _1,k -U _2,k -I _k R ₀ ；

in which y _k As an output value of the measurement equation;

4) System state and system state error covariance a priori estimates:

in the middle ofA state vector priori estimated value at the moment k+1; />An error covariance priori estimated value of the state vector at the moment k+1;

fourth, the method comprises the following steps: measurement update:

1) Output sigma point calculation:

in the middle ofOutputting an estimated value for a measurement equation corresponding to the ith sigma point at the k+1 moment;

2) Cell terminal voltage average and terminal voltage average error covariance calculation:

in the middle ofOutputting a priori estimated value for the k+1 moment measurement equation; p (P) _yy,k+1 For the mean value error covariance of the voltage at the output end of the measurement equation at the moment k+1, V _k Is the output error covariance;

3) UpdatingAnd->Cross covariance between:

p in the formula _xy,k+1 Time k+1And->Cross covariance between;

4) Calculating Kalman gain:

in the middle ofKalman gain at time k+1;

5) State and error covariance update:

6) Noise covariance update:

in e _k+1 Is the noise covariance at time k+1, m _k+1 E is a sliding average value of terminal voltage information _i Terminal voltage innovation; l (L) _w For moving window size;

process noise covariance for state vector at time k+1,/->Measurement noise covariance for the state vector at time k+1;

fifth step: parameter measurement update:

1) Calculating a parameter gain matrix:

in the middle ofEstimating a gain matrix for the k+1 time instant parameter, < >>Estimating an output matrix for the k+1 time parameter;

2) Parameter and error covariance update:

wherein the method comprises the steps ofThe iterative initial value is 0, which is obtained by the following three iterative partial differential equations:

7. the method for estimating SOC of a lithium iron phosphate battery based on ampere-hour integral reference compensation of claim 3, 4, 5 or 6, wherein in step three, the constant deviation is expressed as d:

d＝SOC _{EKF-AUKF,100s} -SOC _Ah,100s ， (33)

SOC in _{EKF-AUKF,100s} SOC value and SOC of 100s time EKF-AUKF algorithm _Ah,100s The SOC value is obtained by an ampere-hour integration method at the moment of 100 s.

8. The method for estimating SOC of lithium iron phosphate battery based on ampere-hour integral reference compensation as claimed in claim 7, wherein in the fourth step, the preliminary difference is denoted as D _k ：

D _k ＝SOC _EKF-AUKF,k -SOC _Ah,k ， (34)

SOC in _EKF-AUKF,k After 100s, the SOC value and the SOC of the EKF-AUKF algorithm at k moment _Ah,k After 100s, the SOC value of the ampere-hour integration method at the k moment;

the approximate deviation value is expressed as Δd:

Δd＝|D _k -d|。 (35)

9. the method for estimating SOC of a lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 8, wherein in the fifth step, a deviation threshold is set to be δ;

if the current approximate deviation value delta d is greater than or equal to the deviation threshold delta, compensating the current EKF-AUKF algorithm SOC value to obtain a k-moment compensation result SOC _k The method comprises the following steps:

SOC _k ＝SOC _EKF-AUKF,k -pΔd， (36)

wherein p is a weight coefficient of the formula,

combining formulas (34) and (35) to obtain k-moment compensation result SOC _k Is the final expression of (2):

SOC _k ＝(1-p)SOC _EKF-AUKF,k +p(SOC _Ah,k +d)。 (37)

10. the method for estimating SOC of lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 9, wherein in step five, if the current approximate deviation value is greater than or equal to the deviation threshold value, the current EKF-AUKF algorithm SOC value is compensated:

setting the demarcation value to be 0.3;

if the SOC value of the current EKF-AUKF algorithm is smaller than or equal to 0.3, the weight coefficient p is selected to be 0.15; if the SOC value of the current EKF-AUKF algorithm is larger than 0.3, selecting a weight coefficient p to be 0.85; deviation threshold δ=0.03.