CN113791360A

CN113791360A - Power lithium battery SOC estimation model under variable temperature condition based on EKF algorithm

Info

Publication number: CN113791360A
Application number: CN202111031095.XA
Authority: CN
Inventors: 程桂石; 叶芯榕; 程琳瑞; 赵莹
Original assignee: North China Electric Power University
Current assignee: North China Electric Power University
Priority date: 2021-09-03
Filing date: 2021-09-03
Publication date: 2021-12-14

Abstract

The embodiment of the present invention proposes a method for estimating the SOC of a lithium battery under variable temperature based on an improved EKF algorithm. The modified RC Thevenin battery simulation model was established, the parameters of the lithium power battery were identified through the hybrid power pulse capability characteristic (HPPC) test, and the relationship between the battery SOC, open circuit voltage and internal resistance was explored. Considering factors that affect battery performance, such as the EKF algorithm and the charging and discharging rate, an estimation model of the SOC of the power lithium battery under variable temperature conditions based on the EKF algorithm is proposed. , the SOC value of the actual two ends of the battery is replaced by the SOC value obtained by the last calculation in the battery management system, which reduces the cumulative error of the EKF calculation caused by ignoring the high-order term of the Taylor expansion of the nonlinear function, and also takes into account. Due to the influence of the battery shutdown factor, the estimation of SOC is closer to the actual value each time. After establishing a simulation model and conducting experiments, the error value of the model is verified.

Description

Power lithium battery SOC estimation model under variable temperature condition based on EKF algorithm

Technical Field

The invention belongs to the technical field of lithium power battery management systems, and particularly relates to SOC algorithm estimation of a lithium battery, which is a basis for battery charge and discharge management and balance control management.

Background

In a battery management system, accurate estimation of a battery state of charge (SOC) is important not only in that a remaining capacity of a battery can be presented to a user, but also in that it is a basis for battery charge and discharge management and balance control management. SOC is affected by many factors, such as temperature, and the magnitude and direction of current, and its accurate prediction is difficult. Improving the estimation accuracy of the SOC plays a certain role in prolonging the service life of the battery and improving the use feeling of a user.

The current classical battery models are many, and parameters of ideal equivalent models in the models are all invariable, so that the accuracy of the models is low. Therefore, it is necessary to improve a battery model capable of improving the estimation accuracy to study the battery characteristics, so that the existing SOC algorithm model can be improved and optimized by using the battery model, and an algorithm model for SOC estimation with high accuracy is provided.

Disclosure of Invention

Aiming at various problems of the existing battery model, the modified RC Thevenin equivalent battery circuit model which is adaptive to temperature change and can reflect the relation between each parameter and the state of charge (SOC) is provided, the influence of each performance parameter of the battery on the SOC at different temperatures can be effectively simulated and tested, and the battery model has the characteristics of high accuracy, wide application range and the like.

The estimation model of the SOC of the power lithium battery under the condition of variable temperature based on the EKF algorithm comprises a modified RC Thevenin equivalent circuit model for researching the battery characteristics and an SOC algorithm estimation model of the power lithium battery under the condition of comprehensively considering the influences of factors such as temperature, battery shutdown time and the like,

wherein:

the modified RC Thevenin equivalent circuit model for the research on the battery characteristics increases the influence of battery polarization, can reflect the relation between each parameter and the state of charge, can reflect the influence of the internal resistance and the current of the battery on the SOC, and has good dynamic and static characteristics. A controlled voltage source is added to simulate the influence of the charging and discharging current characteristics of the lithium battery on external voltage, and the simulation working condition of the model can be improved. The RC equivalent circuit module meets the precision requirement of the dynamic change process in the battery, effectively describes the corresponding relation between the electromotive force and the terminal voltage of the battery, obtains a simulation equation of the power lithium battery, and estimates the characteristics of the battery more accurately.

For the SOC estimation model of the lithium power battery under the condition of variable temperature, the conventional EKF algorithm is modified, the influence of temperature and battery shutdown time on the SOC of the battery is taken into consideration, the influence of the current on the estimation precision is optimized, and the SOC estimation algorithm model of the lithium power battery adapting to the temperature change is obtained.

The technical scheme of the invention discloses an estimation model of the SOC of a power lithium battery under the condition of variable temperature based on an EKF algorithm, which comprises the traditional EKF algorithm, wherein the SOC obtained by calculating the input quantity from the last algorithm is replaced by the corresponding actual SOC value of the lithium battery at the corresponding temperature, so that the accumulated error of the algorithm is reduced; the battery shutdown factor is included in the influence on the SOC values of the two ends of the lithium battery. The technical scheme of the invention has the following beneficial effects:

1. the characteristic of the battery is studied more thoroughly, and the influence factor is calculated more accurately during SOC estimation;

2. the estimation precision of the SOC of the battery is improved, the cycle service life of the lithium battery is prolonged, and the use experience of a user is improved;

3. the influence of the battery shutdown time and the battery temperature on the battery charging and discharging characteristics is brought into the algorithm, and the accuracy and the practicability of SOC algorithm estimation are improved.

Drawings

FIG. 1 is a flow chart of an algorithm model proposed in this patent

FIG. 2 is a modified RC Thevenin equivalent cell circuit model used in this patent

[ description of symbols ]

Rp is a polarization resistance Cp and is a lithium battery electrode polarization capacitance; rl is the ohmic internal resistance of the battery; uoc is the open circuit voltage of the battery; f (I) is a function of the current I

Detailed Description

In order to make the technical problems and the innovative points to be solved by the present invention clearer, the following detailed description is made with reference to the accompanying drawings.

Aiming at the problem of the existing SOC estimation precision, the algorithm shown in FIG. 1 provides a lithium battery SOC estimation method based on an improved EKF algorithm at variable temperature, firstly, an SOC value is read according to an SOC curve of a lithium battery at different temperatures, then, whether the open-circuit voltage of the battery is in a stable state or not is judged according to the size relation between the shutdown time and the judgment duration, and the SOC value under the comprehensive consideration of two factors is obtained.

For the problem of SOC curve fitting at different temperatures, the battery circuit model adopted by the method is a modified RC Thevenin equivalent model, and is shown in figure 2. The RC equivalent circuit module adopted here is first-order, and along with the increase of series RC module, the equivalent precision can be more accurate, but in practice, the requirement of experimental precision in engineering practice can be satisfied just to first-order model, so this patent adopts first-order RC module. In addition, a controlled source is added to the circuit to simulate the influence of the charge-discharge current characteristics of the lithium battery on an external circuit, and the simulation working condition of the model can be improved. The external voltage UI of the battery can be obtained according to the circuit principle as follows:

U_L(t)＝U_{oc_soc}(t)-R_LI(t)-U_p(t)

F[I(t)]＝0.00067[I(t)-3]²+0.0057I(t)-3]-0.0015

wherein, S (t) and S (t +1) are the real-time values of the state of charge of the lithium battery at the time t and t +1, respectively: c_NThe rated capacity of the lithium battery; eta_cIs coulombic efficiency; i (t) is the instantaneous charge-discharge current at time t, positive in the discharged state, otherwise the opposite.

The extended Kalman system space equation is:

wherein a ═ C ═ 1;

d is 0; w (k) is system noise; v (k) is measurement noise

The EKF algorithm filtering comprises the following specific steps:

(1) setting initial value X of covariance matrix of state quantity and state error₀And P₀Recording the covariance matrix of state quantity and state error at the time k as X_kAnd P_k；

(2) The one-step prediction value of the state quantity and the error covariance is as follows:

(3) the correction matrix K is:

K＝P_k,k-1C^T(CP_k,k-1C^T+R)

(4) and correcting the one-step predicted value by using the measured value to obtain an estimated value of the previous time, wherein the estimated value is as follows:

and (3) repeating the steps (2) and (3), continuously predicting and correcting the SOC estimated value by the system, continuously updating the SOC estimated value, considering noise and errors, reducing system accumulated errors and inhibiting the influence of the noise to a great extent.

Obtaining an SOC value at the k +1 moment after the integration, obtaining an SOC estimation value at the k +1 moment by adopting an EKF algorithm, compensating the SOC estimation value for correcting the deviation at the moment, obtaining an open-circuit voltage value at the k +1 moment according to the SOC estimation value at the k +1 moment and a filter input value, comparing the open-circuit voltage value with an open-circuit voltage value read out from an initial OCV curve, obtaining an estimation error of the SOC value at the k +1 moment, adding Kalman gain to compensate the open-circuit voltage value, obtaining a corrected SOC value, and outputting the corrected SOC value to obtain a result.

Claims

1. An estimation model of the SOC of a power lithium battery under the condition of variable temperature based on an EKF algorithm comprises a modified RC Thevenin equivalent circuit model for researching the battery characteristics and an SOC algorithm estimation model of the power lithium battery under the condition of comprehensively considering the influences of factors such as temperature, charge and discharge multiplying power, downtime and the like.

2. The algorithm model of claim 1, wherein the EKF algorithm incorporates the effect of temperature factors on the SOC of the battery, and the input values comprise OCV-SOC curves at different temperatures; the judgment time length x and the reading shutdown time y are obtained according to the relation curve between the time required by the open-circuit voltage of the battery to reach the stable value and the SOC of the battery, the shutdown factors of the battery are considered in the algorithm, and the output quantity of the algorithm is closer to the actual value.

3. The algorithm model of claim 1, wherein a controlled voltage source is added to the RC thevenin equivalent circuit model for studying and correcting the battery characteristics to simulate the influence of the charging and discharging current characteristics of the lithium battery on the external voltage, so that the simulation working condition of the model can be improved. The RC equivalent circuit model meets the precision requirement of the dynamic change process in the battery, effectively describes the corresponding relation between the electromotive force and the terminal voltage of the battery, obtains a simulation equation of the power lithium battery, and estimates the characteristics of the battery more accurately.

4. The algorithmic model of claim 1, wherein the temperature factor is also included in the research influence of the SOC characteristic of the battery, and the effect results of various influencing factors are combined. The influence of the temperature is used as one variable in the algorithm, and an algorithm model suitable for estimating the SOC of the battery at different temperatures can be obtained.

5. The algorithmic model of claim 1, wherein the equivalent circuit model is capable of adapting to changes in temperature, taking into account the effect of temperature on battery operation. In the study of the battery characteristics, a more accurate result is obtained.

6. The algorithm model as claimed in claim 1, wherein the algorithm proposed in this patent replaces the SOC value obtained from the last calculation in the battery management system with the SOC value at the actual two ends of the battery, so as to reduce the accumulated error of the EKF caused by neglecting the high-order term of taylor expansion of the nonlinear function.

7. The algorithmic model of claim 1, wherein the battery SOC estimation value is more accurate to improve the user experience by taking into account the effect of battery shutdown factors on the battery SOC.

8. The algorithm model of claim 1, wherein whether the open-circuit voltage reaches the steady state under the condition that the stop time of the battery is different is obtained according to a relation graph between the time required for the open-circuit voltage of the automotive power lithium battery to reach the steady value and the SOC value of the battery. Therefore, the SOC value of the battery at the time k is taken as the system input quantity to be introduced into the algorithm, or the SOC value obtained by recalculation and introduced into the equivalent circuit model is taken as the system input quantity according to the OCV value of the battery at the time.

9. The algorithm model of claim 1, wherein the SOC values of the battery at the actual operating temperature are read from the SOC curves at different temperatures obtained from the RC Thevenin equivalent circuit model and are introduced into the algorithm.

10. The algorithmic model according to any of claims 1 to 9, characterized in that the equivalent circuit model and the algorithmic model for SOC estimation have consistent impact factors and consistent impact process analysis.