CN117074980A - SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation - Google Patents

Abstract

The SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation belongs to the technical field of battery state-of-charge estimation. The present invention aims at the problem of unreliable results of estimating SOC of lithium iron phosphate batteries using the Kalman filter algorithm under large voltage sampling errors. This includes using the ampere-hour integration method and the EKF-AUKF algorithm to calculate the SOC value of the lithium iron phosphate battery at each sampling time; using the SOC value calculation result of the ampere-hour integration method as a reference value, selecting the SOC value of the two algorithms at 100s to make a difference, and we get Constant deviation; the difference between the SOC values of the two algorithms after 100s is used as the preliminary difference, and then combined with the constant deviation to obtain the approximate deviation value; when the approximate deviation value is greater than the set deviation threshold, the SOC value is demarcated from the preset according to the EKF‑AUKF algorithm Based on the comparison results of values, different weight coefficients are used to compensate the current EKF‑AUKF algorithm SOC value and the compensation result is obtained. The present invention is used for battery SOC estimation.

Description

SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation

Technical field

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

Background technique

Lithium batteries have been widely used in electric vehicles, energy storage power stations, electromagnetic catapults and other fields due to their high safety, long cycle life and high energy density.

State of Charge (SOC) is one of the indicators of the battery management system, which can reflect the remaining usage capacity of the battery. When applied to electric vehicles, lithium battery SOC is directly related to the vehicle's mileage, so it is of great significance to accurately estimate lithium battery SOC. However, lithium battery SOC cannot be measured directly, and its online estimation needs to be achieved through corresponding algorithms.

Currently, lithium battery state of charge estimation methods mainly include open-loop methods, data-driven methods and model-based state estimation methods. The open-loop method is simple to implement and widely used. However, the open-loop method does not have a feedback link and cannot eliminate the impact of the SOC initial error and current accumulation error, and the estimation accuracy is not high. The data-driven method has the characteristics of self-learning and self-adaptation. Parameters such as battery rated capacity, voltage, current and temperature are used as input and SOC is used as output, which can describe the characteristics of complex nonlinear systems; however, the accuracy of this method depends on the input parameters, the quantity and quality of training data; model-based The state estimation method is mainly the Kalman filter and its derivative filters. Considering the nonlinear characteristics of lithium batteries, many state estimation algorithms are suitable for nonlinear systems, such as extended Kalman filter (EKF), unscented Kalman filter (unscented Kalman filter, UKF) and particle filter (particle filter, PF) are proposed to improve the estimation accuracy of lithium battery SOC. These algorithms assume that the process noise and measurement noise are independent Gaussian white noise, but in actual situations the noise is often irregular noise. Therefore, many noise adaptive algorithms, such as adaptive EKF, adaptive UKF, adaptive PF, etc., have been proposed accordingly to reduce the impact of irregular noise. Lithium battery parameters will be affected by changes in battery temperature, discharge rate, aging and other factors, so parameters obtained using fixed parameters or offline methods may not be realistic. Therefore, double extended Kalman filtering, double unscented Kalman filtering, etc. were proposed to improve the SOC estimation accuracy.

Model-based state estimation algorithms mainly focus on model accuracy and algorithm accuracy, ignoring the impact of voltage measurement noise on estimation accuracy. The open circuit voltage of lithium iron phosphate batteries is very flat, and a small voltage error will produce a large SOC estimation error. When there is a large voltage sampling error, there will be a large error in estimating the SOC of the lithium iron phosphate battery using the Kalman filter algorithm.

Contents of the invention

Aiming at the problem of unreliable SOC estimation results of lithium iron phosphate batteries using the Kalman filter algorithm under large voltage sampling errors, the present invention provides a SOC estimation method for lithium iron phosphate batteries based on ampere-hour integral reference compensation.

A method for estimating SOC of lithium iron phosphate batteries based on ampere-hour integral reference compensation of the present invention includes:

Step 1: Based on the equivalent circuit model of the lithium iron phosphate battery, construct the discrete state space equations of the lithium iron phosphate battery;

Step 2: Calculate the SOC value of the lithium iron phosphate battery at each sampling time using the ampere-hour integration method;

At the same time, the EKF-AUKF algorithm is used to calculate the SOC value of the lithium iron phosphate battery at each sampling moment, including using the EKF algorithm to identify the real-time equivalent circuit parameters of the equivalent circuit model, and then using the AUKF algorithm based on the updated discrete equivalent circuit parameters of the real-time equivalent circuit. The state space equations calculate the SOC value of the lithium iron phosphate battery at each sampling moment; then the obtained SOC value and Kalman gain are fed back to the EKF algorithm as the basis for identifying the real-time equivalent circuit parameters of the equivalent circuit model at the next sampling moment;

Step 3: Use the SOC value calculation result of the ampere-hour integration method as the reference value, select the EKF-AUKF algorithm SOC value at 100s and make the difference between the SOC value of the ampere-hour integration method, and the result is used as a constant deviation;

Step 4: Using the 100s time as the starting point, compare the SOC value of the EKF-AUKF algorithm and the SOC value of the ampere-hour integration method at each subsequent sampling time to obtain the preliminary difference; then compare the preliminary difference with the constant deviation to obtain the result. The absolute value is used as the approximate deviation value between the SOC value of the EKF-AUKF algorithm and the true SOC value;

Step 5: Set the deviation threshold;

If the current approximate deviation value is less than the deviation threshold, the current EKF-AUKF algorithm SOC value is used as the SOC estimation result of the current lithium iron phosphate battery;

If the current approximate deviation value is greater than or equal to the deviation threshold, then based on the comparison result of the EKF-AUKF algorithm SOC value and the preset boundary value, different weight coefficients are used to compare the current EKF-AUKF algorithm SOC value based on the current ampere-hour integration method SOC value. Compensate and use the compensation result as the SOC estimation result of the current lithium iron phosphate battery.

According to the lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation of the present invention, in step one, the input-output relationship of the equivalent circuit model of the lithium iron phosphate battery is:

In the formula, U ₁ is the terminal voltage of the battery's reactive polarization internal resistance, R ₁ is the battery's reactive polarization internal resistance, C ₁ is the battery's reactive polarization capacitance, I is the battery current, and U ₂ is the terminal voltage of the battery's concentration polarization internal resistance. , R ₂ is the battery concentration polarization internal resistance, C ₂ is the battery concentration polarization capacitance, U _t is the battery terminal voltage, U _oc is the battery open circuit voltage, and R ₀ is the battery ohmic internal resistance.

According to the SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation of the present invention, in step one, the discrete state space equation set of the lithium iron phosphate battery is constructed as:

In the formula, z _k+1 is the SOC value at time k +1 obtained by using the EKF-AUKF algorithm, z _k is the SOC value at time k obtained by using the EKF-AUKF algorithm, eta is the Coulomb efficiency, Δt is the sampling interval time, and C _n is The maximum available capacity of the battery; U _{1 ,k+1} is U 1 at k+1 moment, α ₁ is the battery reaction polarization time index, α ₁ =exp(-Δt/R ₁ C ₁ ), U _1,k is k U ₁ , U _{2 , k+1} at time are U ₂ at time k + 1 , α ₂ is the battery concentration polarization time index, α ₂ =exp(-Δt/R ₂ C ₂ ), U _2,k is U ₂ and I _k at time k are I at time k, w _k is the process noise at time k, U _t,k is U t at time k, U _{oc ,k} is U _oc at time k, and v _k is the measurement noise at time k.

According to the SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation of the present invention, in step 2, the method of calculating the SOC value of the lithium iron phosphate battery at each sampling time using the ampere-hour integral method includes:

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

In the formula, z _Ah,k+1 is the SOC value at k+1 obtained by the ampere-hour integration method, and z _Ah,k is the SOC value at k obtained by the ampere-hour integration method.

According to the SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation of the present invention, in step 2, the method of using the EKF-AUKF algorithm to calculate the SOC value of the lithium iron phosphate battery at each sampling moment includes:

First, construct a state space equation to describe the changing characteristics of equivalent circuit parameters:

Express the equivalent circuit parameters in the equivalent circuit model at time k as the circuit parameter vector θ _k :

θ _k =[R ₀ R ₁ C ₁ R ₂ C ₂ ] ^T , (4)

Treating the equivalent circuit parameters as constants with disturbances, the state space equation is obtained:

in the formula is the estimated value of the circuit parameter vector at time k+1,/> is the estimated value of the circuit parameter vector at time k, r _k is the white noise at time k, and the white noise r _k conforms to Gaussian distribution/> is the error covariance matrix of r _k ;

In the formula, d _k represents the battery terminal voltage U _t,k at time k, and g(x _k , _uk ,θ _k ) represents the measurement equation including parameters U _oc,k -U _1,k -U _2,k -I _k R ₀ , x _k is used as a state vector, indicating the SOC values z _k+1 , U _1,k+1 and U _2,k at time k obtained by using the EKF-AUKF algorithm; u _k represents the battery current I _k at time k, and λ _k represents Measure the noise at time k, and the measurement noise λ _k conforms to Gaussian distribution. is the error covariance matrix of λ _k .

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

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

In the formula, SOC _{EKF-AUKF,100s} is the SOC value of the EKF-AUKF algorithm at 100s, and SOC _Ah,100s is the SOC value of the ampere-hour integration method at 100s.

According to the SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation of the present invention, in step 4, the preliminary difference is expressed as D _k :

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

In the formula, SOC _EKF-AUKF,k is the SOC value of the EKF-AUKF algorithm at time k after 100s, and SOC _Ah,k is the SOC value of the ampere-hour integration method at time k after 100s;

Then the approximate deviation value is expressed as Δd:

Δd=|D _k -d|. (35)

According to the SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation of the present invention, in step five, the deviation threshold is set to δ;

If the current approximate deviation value Δd is greater than or equal to the deviation threshold δ, the current EKF-AUKF algorithm SOC value is compensated, and the compensation result SOC _k at time k is:

SOC _k =SOC _EKF-AUKF,k -pΔd, (36)

where p is the weight coefficient,

Combining formulas (34) and (35), the final expression of the compensation result SOC _k at time k is obtained:

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 of the present invention, in step five, if the current approximate deviation value is greater than or equal to the deviation threshold, the current EKF-AUKF algorithm SOC value is compensated:

Set the cutoff value to 0.3;

If the SOC value of the current EKF-AUKF algorithm is less 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 greater than 0.3, the weight coefficient p is selected to be 0.85; the deviation threshold δ=0.03.

Beneficial effects of the present invention: The method of the present invention fully considers the influence of model accuracy and algorithm accuracy, and uses the EKF-AUKF algorithm to achieve accurate estimation of SOC of lithium iron phosphate batteries. When the SOC estimation result of the EKF-AUKF algorithm is affected by a large voltage sampling error, the SOC estimation result of the ampere-hour integration method is used as a reference, and its SOC change characteristics are used to calculate the SOC estimation result and true value of the EKF-AUKF algorithm. The approximate difference is then used to compensate the EKF-AUKF algorithm SOC estimation result using the compensation coefficient. While improving the SOC estimation accuracy and robustness, it also eliminates the influence of the ampere-hour integration method SOC initial error and cumulative error to a certain extent, and does not Significantly increases the amount of computation.

Description of the drawings

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

Figure 2 is the equivalent circuit model of a lithium iron phosphate battery.

Detailed ways

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

It should be noted that, as long as there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, but shall not be used as a limitation of the present invention.

Specific Embodiment 1. As shown in Figure 1 and Figure 2, the present invention provides a lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, including:

Step 1: Based on the equivalent circuit model of the lithium iron phosphate battery, construct the discrete state space equations of the lithium iron phosphate battery;

Step 2: Calculate the SOC value of the lithium iron phosphate battery at each sampling time using the ampere-hour integration method;

At the same time, the EKF-AUKF algorithm is used to calculate the SOC value of the lithium iron phosphate battery at each sampling moment, including using the EKF algorithm to identify the real-time equivalent circuit parameters of the equivalent circuit model, and then using the AUKF (Adaptive Unscented Kalman Filtering) algorithm Calculate the SOC value of the lithium iron phosphate battery at each sampling time based on the discrete state space equations after the real-time equivalent circuit parameters are updated; then the obtained SOC value and Kalman gain are fed back to the EKF algorithm to identify the equivalent circuit at the next sampling time The basis for the real-time equivalent circuit parameters of the model;

Step 3: Use the SOC value calculation result of the ampere-hour integration method as the reference value, select the EKF-AUKF algorithm SOC value at 100s and make the difference between the SOC value of the ampere-hour integration method, and the result is used as a constant deviation;

Step 4: Using the 100s time as the starting point, compare the SOC value of the EKF-AUKF algorithm and the SOC value of the ampere-hour integration method at each subsequent sampling time to obtain the preliminary difference; then compare the preliminary difference with the constant deviation to obtain the result. The absolute value is used as the approximate deviation value between the SOC value of the EKF-AUKF algorithm and the true SOC value;

Step 5: Set the deviation threshold;

If the current approximate deviation value is less than the deviation threshold, the current EKF-AUKF algorithm SOC value is used 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, compensation is performed; then based on the comparison result of the EKF-AUKF algorithm SOC value and the preset boundary value, different weight coefficients are used to calculate the current EKF-AUKF- The AUKF algorithm SOC value is compensated, and the compensation result is used as the SOC estimation result of the current lithium iron phosphate battery. That is, when the approximate deviation value is not less than the deviation threshold, the approximate deviation value is compensated to the SOC estimation result of the EKF-AUKF algorithm as the final estimated value of the current SOC of the lithium battery to be tested.

Furthermore, in step one, according to Figure 2, the input-output relationship of the equivalent circuit model of the lithium iron phosphate battery can be obtained from Kirchhoff's law:

In the formula, U ₁ is the terminal voltage of the battery's reactive polarization internal resistance, R ₁ is the battery's reactive polarization internal resistance, C ₁ is the battery's reactive polarization capacitance, I is the battery current, and U ₂ is the terminal voltage of the battery's concentration polarization internal resistance. , R ₂ is the battery concentration polarization internal resistance, C ₂ is the battery concentration polarization capacitance, U _t is the battery terminal voltage, U _oc is the battery open circuit voltage, and R ₀ is the battery ohmic internal resistance.

As shown in Figure 2, the equivalent circuit model of the lithium iron phosphate battery includes a voltage source, ohmic internal resistance, reactive polarization internal resistance, reactive polarization capacitance, concentration polarization internal resistance and concentration polarization capacitance. Second-order resistor-capacitor equivalent circuit. The reactive polarization internal resistance and reactive polarization capacitance are used to describe the reactive polarization phenomenon inside the battery, and the concentration polarization internal resistance and concentration polarization capacitance are used to describe the concentration polarization phenomenon inside the battery.

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

In the formula, z _k+1 is the SOC value at time k +1 obtained by using the EKF-AUKF algorithm, z _k is the SOC value at time k obtained by using the EKF-AUKF algorithm, eta is the Coulomb efficiency, Δt is the sampling interval time, and C _n is The maximum available capacity of the battery; U _{1 ,k+1} is U 1 at k+1 moment, α ₁ is the battery reaction polarization time index, α ₁ =exp(-Δt/R ₁ C ₁ ), U _1,k is k U ₁ , U _{2 , k+1} at time are U ₂ at time k + 1 , α ₂ is the battery concentration polarization time index, α ₂ =exp(-Δt/R ₂ C ₂ ), U _2,k is U ₂ and I _k at time k are I at time k, w _k is the process noise at time k, which is used to represent the current measurement error and state equation error; U _t,k is U _t at time k, and U _oc,k is time k U _oc and v _k are the measurement noise at time k, which are used to characterize the voltage measurement error and output equation error.

Furthermore, in step two, the method of calculating the SOC value of the lithium iron phosphate battery at each sampling time using the discrete ampere-hour integration method includes:

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

In the formula, z _Ah,k+1 is the SOC value at k+1 obtained by the ampere-hour integration method, and z _Ah,k is the SOC value at k obtained by the ampere-hour integration method.

At the same time, in step two, the method of using the EKF-AUKF algorithm to calculate the SOC value of the lithium iron phosphate battery at each sampling moment includes:

First, to use the EKF algorithm to identify the equivalent circuit parameters of lithium iron phosphate batteries online, it is necessary to construct a state space equation to describe the changing characteristics of the equivalent circuit parameters:

Express the equivalent circuit parameters in the equivalent circuit model at time k as the circuit parameter vector θ _k :

θ _k =[R ₀ R ₁ C ₁ R ₂ C ₂ ] ^T , (4)

Lithium battery parameters change slowly and can change significantly after many discharge cycles. Therefore, the equivalent circuit parameters can be regarded as constants with disturbances. Therefore, its state space equation can be written as follows:

in the formula is the estimated value of the circuit parameter vector at time k+1,/> is the estimated value of the circuit parameter vector at time k, r _k is the white noise at time k, and the white noise r _k conforms to Gaussian distribution/> is the error covariance matrix of r _k ;

In the formula, d _k represents the battery terminal voltage U _t,k at time k, and g(x _k , _uk ,θ _k ) represents the measurement equation including parameters U _oc,k -U _1,k -U _2,k -I _k R ₀ , x _k is used as a state vector, indicating the SOC values z _k+1 , U _1,k+1 and U _2,k at time k obtained by using the EKF-AUKF algorithm; u _k represents the battery current I _k at time k, and λ _k represents Measure the noise at time k, and the measurement noise λ _k conforms to Gaussian distribution. is the error covariance matrix of λ _k .

Furthermore, in step 2, after constructing the state space equation (5), the SOC value at each sampling moment is calculated:

1: Initialization:

1. Initialization parameter vector and error covariance of parameter vector:

in the formula is the initial estimated value of the circuit parameter vector, θ ₀ is the initial value of the circuit parameter vector, and E represents averaging;

is the initial error covariance of the parameter vector;

2. Initialize the state vector and the error covariance of the state vector:

in the formula is the initial estimated value of the state vector, x ₀ is the initial value of the state vector;/> is the initial error covariance of the state vector;

2: Time update of parameter estimation:

in the formula is the estimated value of the circuit parameter vector at time k,/> is the prior estimate of the circuit parameter vector at time k+1, is the parameter vector error covariance at time k,/> is the prior estimate of the error covariance of the parameter vector at time k+1;

Three: Time update of status estimate:

1) Calculate the weight of the unscented transformation sampling point:

in the formula is the initial value of the state mean weight factor, λ is the scale factor used to reduce the global prediction error, and n is the battery system state dimension;

is the initial value of the error covariance weight factor, β is the error adjustment coefficient of the high-order term, α is the control coefficient of the distribution of sigma points around the state mean point, W _i ^m is the i-th state mean weight factor, and W _i ^c is the i-th state mean weight factor. i error covariance weight factors, κ is a second-order scale parameter, used to ensure/> Zhengding;/> is the error covariance of the state vector at time k;

As an example: n=3, α=0.01, β=2, κ=0.

2) Calculate 2n+1 sigma points at time k:

in the formula is the 0th state vector value at time k,/> is the estimated value of the state vector at time k,/> is the i-th state vector value at time k;

3) Update the prior state of the sigma point:

in the formula is the estimated value of the i-th state vector at time k+1,

Express/>

but:

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

In the formula, y _k is used as the output value of the measurement equation;

4) Priori estimation of system state and system state error covariance:

in the formula is the prior estimate of the state vector at time k+1;/> is the prior estimate of the error covariance of the state vector at time k+1;

Four: Measurement update:

1) Output sigma point calculation:

in the formula Output the estimated value for the measurement equation corresponding to the i-th sigma point at time k+1;

2) Calculation of the error covariance of the average battery terminal voltage and the average terminal voltage:

in the formula Output a priori estimate value for the measurement equation at time k+1; P _yy,k+1 is the error covariance of the average voltage at the output terminal of the measurement equation at time k+1, V _k is the output error covariance;

3)Update with/> Cross covariance between:

In the formula, P _xy,k+1 is the k+1 time with/> cross covariance between;

4) Calculate the Kalman gain:

in the formula is the Kalman gain at time k+1;

5) Status and error covariance update:

6) Noise covariance update:

In the formula, e _k+1 is the noise covariance at time k+1, m _k+1 is the sliding average of the terminal voltage information, e _i is the terminal voltage information; L _w is the moving window size;

is the process noise covariance of the state vector at time k+1,/> is the measurement noise covariance of the state vector at time k+1;

Five: Parameter measurement update:

1) Calculate parameter gain matrix:

in the formula Estimating the gain matrix for the parameter at time k+1,/> The output matrix is the parameter estimate at time k+1;

2) Parameters and their error covariance update:

in It can be obtained through the following three iterable partial differential equations, with the initial iteration value being 0:

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

In the initial stage of estimation, based on the convergence speed of the algorithm, the logic delay effect of 100s is corrected to ensure that the SOC estimated value of the adaptive algorithm can approach the true value. The adaptive algorithm is not modified during the delay period. When the delay time point is reached, the difference between the adaptive algorithm and the ampere-hour integration method is calculated, which is used as the constant deviation between the adaptive algorithm and the ampere-hour integration method. The constant deviation is expressed as d :

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

In the formula, SOC _{EKF-AUKF,100s} is the SOC value of the EKF-AUKF algorithm at 100s, and SOC _Ah,100s is the SOC value of the ampere-hour integration method at 100s.

Before 100s, the compensation logic does not work, and the ampere-hour integration method and the EKF-AUKF algorithm independently estimate the lithium battery SOC.

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

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

In the formula, SOC _EKF-AUKF,k is the SOC value of the EKF-AUKF algorithm at time k after 100s, and SOC _Ah,k is the SOC value of the ampere-hour integration method at time k after 100s;

Then the approximate deviation value is expressed as Δd:

Δd=|D _k -d|. (35)

Finally, in step five, set the deviation threshold to δ;

When Δd is less than δ, the adaptive algorithm estimation error is within a limited range and no compensation is needed; if the current approximate deviation value Δd is greater than or equal to the deviation threshold δ, the current EKF-AUKF algorithm SOC value is compensated, and the compensation result SOC at time k is obtained _k is:

SOC _k =SOC _EKF-AUKF,k -pΔd, (36)

where p is the weight coefficient,

The estimation error of the ampere-hour integral method gradually accumulates as the estimation process proceeds, while the adaptive algorithm estimates very accurately at the end of battery discharge. At this time, if the difference between the estimation results of the two algorithms is used to compensate, there will be The inaccurate estimation result of the ampere-hour integration method introduces a large error. Therefore, the compensation value is changed by adjusting the weight coefficient, which can be called the compensation coefficient. In the area where the ampere-hour integrated SOC estimation error is small, the compensation coefficient is large; in the area where the ampere-hour integrated SOC estimation error is large, the compensation coefficient is small.

Combining formulas (34) and (35), the final expression of the compensation result SOC _k at time k is obtained:

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

where 0≤p≤1.

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

The cut-off value is set to 0.3; according to the SOC estimation characteristics of the adaptive algorithm and the ampere-hour integration method, P takes the SOC estimated value of 0.3 as the cut-off point and takes values respectively.

If the SOC value of the current EKF-AUKF algorithm is less 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 greater than 0.3, the weight coefficient p is selected to be 0.85; the deviation threshold δ=0.03.

Considering that the SOC estimation error of the ampere-hour integration method will gradually increase with the current sampling error, the SOC estimation result of the EKF-AUKF algorithm will approach the true value at the end of discharge, with the EKF-AUKF algorithm SOC estimation value of 0.3 as the dividing line. , select different compensation weight coefficients on both sides to make full use of the stability of the ampere-hour integral method estimation and eliminate the impact of the ampere-hour integral cumulative error. The compensated SOC can be considered the current accurate SOC.

The present invention also provides an electronic device, including a processor and a memory. A computer control program is stored in the memory. When the processor processes the program stored in the memory, it can realize the iron phosphate based on ampere-hour integral reference compensation according to the present invention. Lithium battery SOC estimation method.

The present invention also provides a computer-readable storage medium for storing a computer control program. The computer control program can implement the SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation according to the present invention.

Although the present invention is described herein with reference to specific embodiments, it is to be understood that these embodiments are merely exemplary of the principles and applications of the invention. It is therefore to be understood that many modifications may be made to the exemplary embodiments and other arrangements may be devised without departing from the spirit and scope of the invention as defined by the appended claims. It is to be understood that the features described in the different dependent claims may be combined in a different manner than that described in the original claims. It will also be understood that features described in connection with individual embodiments can be used in other described embodiments.

Claims (10)

1. A lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation, which is characterized by including: Step 1: Based on the equivalent circuit model of the lithium iron phosphate battery, construct the discrete state space equations of the lithium iron phosphate battery; Step 2: Calculate the SOC value of the lithium iron phosphate battery at each sampling time using the ampere-hour integration method; At the same time, the EKF-AUKF algorithm is used to calculate the SOC value of the lithium iron phosphate battery at each sampling moment, including using the EKF algorithm to identify the real-time equivalent circuit parameters of the equivalent circuit model, and then using the AUKF algorithm based on the updated discrete equivalent circuit parameters of the real-time equivalent circuit. The state space equations calculate the SOC value of the lithium iron phosphate battery at each sampling moment; then the obtained SOC value and Kalman gain are fed back to the EKF algorithm as the basis for identifying the real-time equivalent circuit parameters of the equivalent circuit model at the next sampling moment; Step 3: Use the SOC value calculation result of the ampere-hour integration method as the reference value, select the EKF-AUKF algorithm SOC value at 100s and make the difference between the SOC value of the ampere-hour integration method, and the result is used as a constant deviation; Step 4: Using the 100s time as the starting point, compare the SOC value of the EKF-AUKF algorithm and the SOC value of the ampere-hour integration method at each subsequent sampling time to obtain the preliminary difference; then compare the preliminary difference with the constant deviation to obtain the result. The absolute value is used as the approximate deviation value between the SOC value of the EKF-AUKF algorithm and the true SOC value; Step 5: Set the deviation threshold; If the current approximate deviation value is less than the deviation threshold, the current EKF-AUKF algorithm SOC value is used as the SOC estimation result of the current lithium iron phosphate battery; If the current approximate deviation value is greater than or equal to the deviation threshold, then based on the comparison result of the EKF-AUKF algorithm SOC value and the preset boundary value, different weight coefficients are used to compare the current EKF-AUKF algorithm SOC value based on the current ampere-hour integration method SOC value. Compensate and use the compensation result as the SOC estimation result of the current lithium iron phosphate battery. 2. The SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 1, characterized in that in step one, the input-output relationship of the equivalent circuit model of the lithium iron phosphate battery is: In the formula, U ₁ is the terminal voltage of the battery's reactive polarization internal resistance, R ₁ is the battery's reactive polarization internal resistance, C ₁ is the battery's reactive polarization capacitance, I is the battery current, and U ₂ is the terminal voltage of the battery's concentration polarization internal resistance. , R ₂ is the battery concentration polarization internal resistance, C ₂ is the battery concentration polarization capacitance, U _t is the battery terminal voltage, U _oc is the battery open circuit voltage, and R ₀ is the battery ohmic internal resistance. 3. The SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 2, characterized in that in step one, the discrete state space equation set of the lithium iron phosphate battery constructed is: In the formula, z _k+1 is the SOC value at time k +1 obtained by using the EKF-AUKF algorithm, z _k is the SOC value at time k obtained by using the EKF-AUKF algorithm, eta is the Coulomb efficiency, Δt is the sampling interval time, and C _n is The maximum available capacity of the battery; U _{1 ,k+1} is U 1 at k+1 moment, α ₁ is the battery reaction polarization time index, α ₁ =exp(-Δt/R ₁ C ₁ ), U _1,k is k U ₁ , U _{2 , k+1} at time are U ₂ at time k + 1 , α ₂ is the battery concentration polarization time index, α ₂ =exp(-Δt/R ₂ C ₂ ), U _2,k is U ₂ and I _k at time k are I at time k, w _k is the process noise at time k, U _t,k is U t at time k, U _{oc ,k} is U _oc at time k, and v _k is the measurement noise at time k. 4. The lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation according to claim 3, characterized in that in step two, the ampere-hour integral method is used to calculate the SOC value of the lithium iron phosphate battery at each sampling moment. Methods include: z _Ah,k+1 =z _Ah,k -ηI _k Δt/C _n , (3) In the formula, z _Ah,k+1 is the SOC value at k+1 obtained by the ampere-hour integration method, and z _Ah,k is the SOC value at k obtained by the ampere-hour integration method. 5. The SOC estimation method of lithium iron phosphate battery based on ampere-hour integral reference compensation according to claim 3, characterized in that in step two, the EKF-AUKF algorithm is used to calculate the SOC value of the lithium iron phosphate battery at each sampling moment. Methods include: First, construct a state space equation to describe the changing characteristics of equivalent circuit parameters: Express the equivalent circuit parameters in the equivalent circuit model at time k as the circuit parameter vector θ _k : θ _k =[R ₀ R ₁ C ₁ R ₂ C ₂ ] ^T , (4) Treating the equivalent circuit parameters as constants with disturbances, the state space equation is obtained: in the formula is the estimated value of the circuit parameter vector at time k+1,/> is the estimated value of the circuit parameter vector at time k, r _k is the white noise at time k, and the white noise r _k conforms to Gaussian distribution/> is the error covariance matrix of r _k ; In the formula, d _k represents the battery terminal voltage U _t,k at time k, and g(x _k , _uk ,θ _k ) represents the measurement equation including parameters U _oc,k -U _1,k -U _2,k -I _k R ₀ , x _k is used as a state vector, indicating the SOC values z _k+1 , U _1,k+1 and U _2,k at time k obtained by using the EKF-AUKF algorithm; u _k represents the battery current I _k at time k, and λ _k represents Measure the noise at time k, and the measurement noise λ _k conforms to Gaussian distribution is the error covariance matrix of λ _k . 6. The lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation according to claim 5, characterized in that in step 2, after constructing the state space equation (5), the SOC value at each sampling moment is calculated: 1: Initialization: 1. Initialization parameter vector and error covariance of parameter vector: in the formula is the initial estimated value of the circuit parameter vector, θ ₀ is the initial value of the circuit parameter vector, and E represents averaging; is the initial error covariance of the parameter vector; 2. Initialize the state vector and the error covariance of the state vector: in the formula is the initial estimated value of the state vector, x ₀ is the initial value of the state vector;/> is the initial error covariance of the state vector; 2: Time update of parameter estimation: in the formula is the estimated value of the circuit parameter vector at time k,/> is the a priori estimate of the circuit parameter vector at time k+1,/> is the parameter vector error covariance at time k,/> is the prior estimate of the error covariance of the parameter vector at time k+1; Three: Time update of status estimate: 1) Calculate the weight of the unscented transformation sampling point: In the formula, W ₀ ^m is the initial value of the state mean weight factor, λ is the scale factor used to reduce the global prediction error, and n is the battery system state dimension; W ₀ ^c is the initial value of the error covariance weight factor, β is the error adjustment coefficient of the high-order term, α is the control coefficient of the distribution of sigma points around the state mean point, W _i ^m is the i-th state mean weight factor, W _i ^c is the i-th error covariance weight factor, κ is the second-order scale parameter, used to ensure Zhengding;/> is the error covariance of the state vector at time k; 2) Calculate 2n+1 sigma points at time k: in the formula is the 0th state vector value at time k,/> is the estimated value of the state vector at time k,/> is the i-th state vector value at time k; 3) Update the prior state of the sigma point: in the formula is the estimated value of the i-th state vector at time k+1, Express/> but: x _k+1 ＝[z _k+1 U _1,k+1 U _2,k+1 ] ^T , y _k =U _t,k , g(x _k , _uk ,θ _k )=U _oc,k -U _1,k -U _2,k -I _k R ₀ ; In the formula, y _k is used as the output value of the measurement equation; 4) Priori estimation of system state and system state error covariance: in the formula is the prior estimate of the state vector at time k+1;/> is the prior estimate of the error covariance of the state vector at time k+1; Four: Measurement update: 1) Output sigma point calculation: in the formula Output the estimated value for the measurement equation corresponding to the i-th sigma point at time k+1; 2) Calculation of the error covariance between the average battery terminal voltage and the average terminal voltage: in the formula Output a priori estimate value for the measurement equation at time k+1; P _yy,k+1 is the error covariance of the average voltage at the output terminal of the measurement equation at time k+1, V _k is the output error covariance; 3)Update with/> Cross covariance between: In the formula, P _xy,k+1 is the k+1 time with/> cross covariance between; 4) Calculate the Kalman gain: in the formula is the Kalman gain at time k+1; 5) Status and error covariance update: 6) Noise covariance update: In the formula, e _k+1 is the noise covariance at time k+1, m _k+1 is the sliding average of the terminal voltage information, e _i is the terminal voltage information; L _w is the moving window size; is the process noise covariance of the state vector at time k+1,/> is the measurement noise covariance of the state vector at time k+1; Five: Parameter measurement update: 1) Calculate parameter gain matrix: in the formula Estimating the gain matrix for the parameter at time k+1,/> The output matrix is the parameter estimate at time k+1; 2) Parameters and their error covariance update: in Obtained through the following three iterable partial differential equations, the initial value of the iteration is 0: 7. The lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation according to claim 3, 4, 5 or 6, characterized in that, in step three, the constant deviation is expressed as d: d＝SOC _{EKF-AUKF,100s} -SOC _Ah,100s , (33) In the formula, SOC _{EKF-AUKF,100s} is the SOC value of the EKF-AUKF algorithm at 100s, and SOC _Ah,100s is the SOC value of the ampere-hour integration method at 100s. 8. The lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation according to claim 7, characterized in that in step four, the preliminary difference is expressed as D _k : D _k =SOC _EKF-AUKF,k -SOC _Ah,k , (34) In the formula, SOC _EKF-AUKF,k is the SOC value of the EKF-AUKF algorithm at time k after 100s, and SOC _Ah,k is the SOC value of the ampere-hour integration method at time k after 100s; Then the approximate deviation value is expressed as Δd: Δd=|D _k -d|. (35) 9. The lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation according to claim 8, characterized in that, in step five, the deviation threshold is set to δ; If the current approximate deviation value Δd is greater than or equal to the deviation threshold δ, the current EKF-AUKF algorithm SOC value is compensated, and the compensation result SOC _k at time k is: SOC _k =SOC _EKF-AUKF,k -pΔd, (36) where p is the weight coefficient, Combining formulas (34) and (35), the final expression of the compensation result SOC _k at time k is obtained: SOC _k =(1-p)SOC _EKF-AUKF,k +p(SOC _Ah,k +d). (37) 10. The lithium iron phosphate battery SOC estimation method based on ampere-hour integral reference compensation according to claim 9, characterized in that in step five, if the current approximate deviation value is greater than or equal to the deviation threshold, the current EKF-AUKF algorithm SOC value To compensate: Set the cutoff value to 0.3; If the SOC value of the current EKF-AUKF algorithm is less 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 greater than 0.3, the weight coefficient p is selected to be 0.85; the deviation threshold δ=0.03.