A lithium battery SOE estimation method, device and system

Abstract

The present invention provides a lithium battery SOE estimation method, device and system, which relate to the field of lithium battery technology. The lithium battery SOE estimation method described in the present invention includes: estimating the lithium battery SOE based on the parameters of the lithium battery equivalent circuit model using a sliding mode control algorithm; the parameters of the lithium battery equivalent circuit model are obtained by obtaining experimental data through a constant current pulse discharge experiment on the lithium battery, and fitting the relationship curve between the open circuit voltage and the lithium battery SOE in combination with the experimental data to identify the pre-constructed lithium battery equivalent circuit model. The technical solution described in the present invention switches between different control estimation methods based on the internal characteristics of the battery based on a sliding mode control strategy, which helps to improve the anti-interference ability of the control algorithm, refine the control strategy of the lithium battery control system, integrate the control advantages of multiple algorithms, improve the estimation accuracy of the lithium battery SOE, accelerate the convergence speed of the SOE estimation, and optimize the robust performance of the entire management system.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium batteries, and in particular to a method, device and system for estimating SOE of a lithium battery. Background Art

[0002] Lithium battery SOE (State of Energy) is the most important parameter in the battery management system. Currently, the main methods for estimating battery SOE include: open circuit voltage method, Ah measurement method, Kalman filter algorithm, neural network and fuzzy control.

[0003] However, in actual engineering applications, facing the real-time estimation of battery SOE, the above algorithms have the following shortcomings: 1) poor robustness when model parameters are mismatched; 2) more sensitive to system measurement noise and initial state statistical characteristics; 3) when the filter reaches stability, it has poor tracking performance when facing sudden changes; 4) the AH integral method cannot solve the initial error, and there will be cumulative errors during the calculation process; 5) the EKF has the problem of large Taylor series equivalent calculation amount and large error; 6) the filter stability, reliability and convergence are poor due to the influence of many factors. Summary of the Invention

[0004] The problem solved by the present invention is how to improve the estimation accuracy of SOE of lithium batteries.

[0005] To solve the above problems, the present invention provides a lithium battery SOE estimation method, comprising: estimating the lithium battery SOE based on the parameters of the lithium battery equivalent circuit model using a sliding mode control algorithm; the parameters of the lithium battery equivalent circuit model are obtained by obtaining experimental data by performing a constant current pulse discharge experiment on the lithium battery, and fitting a relationship curve between the open circuit voltage and the lithium battery SOE based on the experimental data, and identifying the pre-constructed lithium battery equivalent circuit model.

[0006] The lithium battery SOE estimation method described in the present invention switches between different control estimation methods based on a sliding mode control strategy according to the internal characteristics of the battery, which helps to improve the anti-interference ability of the control algorithm, refine the control strategy of the lithium battery control system, integrate the control advantages of multiple algorithms, improve the estimation accuracy of the lithium battery SOE, accelerate the convergence speed of SOE estimation, and optimize the robust performance of the entire management system.

[0007] Optionally, estimating the lithium battery SOE using a quasi-sliding mode control algorithm based on the parameters of the lithium battery equivalent circuit model includes: determining a state equation of the lithium battery equivalent circuit model according to the identified parameters of the lithium battery equivalent circuit model, establishing a quasi-sliding mode control total relationship according to the state equation and the quasi-sliding mode control algorithm implementation logic, and estimating the lithium battery SOE according to the quasi-sliding mode control total relationship, wherein the quasi-sliding mode control total relationship includes:

[0008] SOE=a*SOE AH +b*SOE EKF +c*SOE SPKF ;

[0009] Where, SOE AH 、SOE EKF and SOE SPKF represent the SOE values ​​calculated by the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively. a, b, and c represent the weighting coefficients corresponding to the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively.

[0010] The lithium battery SOE estimation method described in the present invention estimates the lithium battery SOE based on a sliding mode control-like total relationship, integrates the control advantages of multiple algorithms, improves the estimation accuracy of the lithium battery SOE, accelerates the convergence speed of SOE estimation, and optimizes the robust performance of the entire management system.

[0011] Optionally, the lithium battery SOE estimation method further includes: determining a, b, and c according to the lithium battery SOC and the terminal voltage difference:

[0012] When SOC>90% and the terminal voltage difference is greater than 10mV, a=0,b=1-c,c=(U EKF -U L ) / (U L );

[0013] When 5%<SOC≤90%, a=0,b=(U SPKF -U L ) / (U L ), c = 1-b;

[0014] When SOC≤5%, a=1-bc,b=(U EKF -U L ) / (U L ), c=(U SPKF -U L ) / (U L );

[0015] Among them, U L 、U EKF and U SPKF They represent the load voltage, the voltage calculated by the extended Kalman filter algorithm, and the voltage calculated by the improved Sigma point filter algorithm respectively.

[0016] The lithium battery SOE estimation method described in the present invention determines a, b, and c based on the lithium battery SOC and terminal voltage difference, and then estimates the lithium battery SOE based on a sliding mode control-like total relationship. It integrates the control advantages of multiple algorithms, improves the estimation accuracy of the lithium battery SOE, accelerates the convergence speed of SOE estimation, and optimizes the robust performance of the entire management system.

[0017] Optionally, the improved Sigma point filtering algorithm includes: initializing state variables, using the String interpolation method to calculate the Sigma points and their corresponding weights according to the symmetric sampling strategy, and using the UT transformation method to calculate the Sigma points and their corresponding weights according to the symmetric sampling strategy; calculating the one-step prediction of the Sigma point set, calculating the one-step prediction and covariance matrix of the state variables, substituting the one-step prediction into the observation equation to obtain the predicted observation; determining the observation prediction value of the Sigma point set, and obtaining the system prediction mean and covariance by weighted summation; calculating the Kalman gain matrix, and calculating the state update and covariance update of the system.

[0018] The lithium battery SOE estimation method described in the present invention adopts an improved Sigma point filtering algorithm, which can integrate the control advantages of multiple algorithms, improve the estimation accuracy of lithium battery SOE, accelerate the convergence speed of SOE estimation, and optimize the robust performance of the entire management system.

[0019] Optionally, the improved Sigma point filtering algorithm is based on a hybrid Sigma point filtering strategy, introduces two Sigma point generation methods, and increases the number of Sigma points to 4n+1, where n represents the dimension of the state equation.

[0020] The lithium battery SOE estimation method described in the present invention is based on a hybrid Sigma point filtering strategy, introduces two Sigma point generation methods, and increases the number of Sigma points to 4n+1, thereby increasing the richness of the Sigma point set generated, reducing the excessive weight of individual points in the Sigma point set, avoiding the impoverishment and degradation problems of the Sigma point algorithm, and is conducive to improving the estimation accuracy of the lithium battery SOE.

[0021] Optionally, the method for establishing a lithium battery equivalent circuit model includes:

[0022] Analyze the internal process of the lithium battery and establish a second-order equivalent circuit model, wherein the second-order equivalent circuit model includes a state equation and an observation equation;

[0023] Wherein, the state equation is expressed as:

[0024]

[0025] Wherein, the observation equation is expressed as:

[0026] UL (t) = U OC (t)-i(t)×R0(t)-U1(t)-U2(t)+v(t);

[0027] Among them, SOE0 represents the initial value of the remaining energy of the battery, SOE t Indicates the remaining energy of the battery at time t, Q N represents the rated capacity of the battery, I(t) represents the charge and discharge current at time t, U(t) represents the battery terminal voltage at time t, η represents the charge and discharge efficiency, U L (t) represents the load voltage at time t, U OC (t) represents the open circuit voltage at time t, i(t) represents the discharge current at time t, R0(t) represents the ohmic resistance at time t, U1(t) and U2(t) represent the two polarization voltage values ​​at time t, and v(t) represents the noise at time t;

[0028] Discretizing the second-order equivalent circuit model, and determining a discretized state equation and a discretized observation equation in combination with Kirchhoff's current-voltage law;

[0029] The discretized state equation is expressed as:

[0030]

[0031] The discretized observation equation is expressed as:

[0032] U L (k)=U OC (k)-i(k)×R0(k)-U1(k)-U2(k)+v(k);

[0033] Among them, SOC(k+1) represents the state of charge of the system at time k+1, U1(k+1) and U2(k+1) represent the two polarization voltage values ​​of the system at time k+1, Δt represents the sampling time, R1 and R2 represent the two polarization resistances, R0 represents the ohmic internal resistance, τ1 and τ2 represent the two polarization times, η represents the charge and discharge efficiency, Q N represents the rated capacity of the battery, i(k) represents the discharge current of the system at time k, w(k) and v(k) represent the state and measurement noise of the system at time k, U k (t) represents the load voltage at time k, U OC (k) represents the open circuit voltage at time k, R0(k) represents the ohmic resistance at time k, U1(k) and U2(k) represent the two polarization voltage values ​​at time k.

[0034] The lithium battery SOE estimation method described in the present invention establishes a second-order equivalent circuit model of the lithium battery and discretizes it, completing the basic steps of the algorithm, and then being able to apply a sliding mode control strategy.

[0035] Optionally, the constant current pulse discharge experiment is performed on the lithium battery to obtain experimental data, and the relationship curve between the open circuit voltage and the lithium battery SOE is fitted in combination with the experimental data, and the pre-constructed lithium battery equivalent circuit model is identified, including: performing different constant current pulse discharge experiments on the lithium battery to obtain current and voltage experimental data; fitting the relationship curve between the open circuit voltage and the lithium battery SOE in combination with the current and voltage experimental data; and using the recursive least squares method in combination with the experimental data to identify the parameters of the lithium battery equivalent circuit model.

[0036] The lithium battery SOE estimation method described in the present invention performs different constant current pulse discharge experiments on the lithium battery to obtain experimental data, and then uses the recursive least squares method to identify the parameters of the lithium battery equivalent circuit model, and then can use a sliding mode control algorithm to estimate the lithium battery SOE.

[0037] The present invention also provides a lithium battery SOE estimation device, comprising: an experimental module for obtaining experimental data by performing a constant current pulse discharge experiment on a lithium battery, fitting a curve showing the relationship between open-circuit voltage and lithium battery SOE based on the experimental data, and identifying a pre-constructed lithium battery equivalent circuit model to obtain parameters of the lithium battery equivalent circuit model; and an estimation module for estimating the lithium battery SOE using a sliding mode control algorithm based on the parameters of the lithium battery equivalent circuit model. The lithium battery SOE estimation device of the present invention shares the same advantages as the aforementioned lithium battery SOE estimation method over the prior art, and these advantages are not further elaborated here.

[0038] The present invention also provides a lithium battery SOE estimation system, comprising a computer-readable storage medium storing a computer program and a processor. When the computer program is read and executed by the processor, the system implements the lithium battery SOE estimation method described above. The advantages of the lithium battery SOE estimation system and the lithium battery SOE estimation method described above over the prior art are the same and will not be further described here.

[0039] The present invention also provides a computer-readable storage medium storing a computer program. When read and executed by a processor, the computer program implements the above-described lithium battery SOE estimation method. The computer-readable storage medium and the above-described lithium battery SOE estimation method have the same advantages over the prior art and are not further described here. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 Schematic diagram of a method for estimating SOE of a lithium battery according to an embodiment of the present invention;

[0041] Figure 2 This is a diagram of the Thevenin equivalent circuit model of a lithium battery according to an embodiment of the present invention;

[0042] Figure 3 This is a working condition diagram of a lithium-ion battery according to an embodiment of the present invention;

[0043] Figure 4 A schematic diagram of the distribution of mixed Sigma points generated in an embodiment of the present invention;

[0044] Figure 5 This is a comparison diagram of the effects of the SOE estimation algorithm according to an embodiment of the present invention;

[0045] Figure 6 This is a comparison diagram of the SOE estimation effect errors of an embodiment of the present invention;

[0046] Figure 7 2 is a comparison diagram of the voltages at the terminals of various algorithms according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0048] like Figure 1 As shown, an embodiment of the present invention provides a lithium battery SOE estimation method, comprising: estimating the lithium battery SOE based on the parameters of the lithium battery equivalent circuit model using a sliding mode control algorithm; the parameters of the lithium battery equivalent circuit model are obtained by obtaining experimental data by performing a constant current pulse discharge experiment on the lithium battery, and fitting a relationship curve between the open circuit voltage and the lithium battery SOE in combination with the experimental data, and identifying the pre-constructed lithium battery equivalent circuit model.

[0049] Specifically, in this embodiment, the lithium battery SOE estimation method includes: 1) establishing a mathematical model of the lithium battery system, that is, a lithium battery equivalent circuit model, and deriving its nonlinear system equation expression; 2) designing a working condition experiment of the lithium-ion battery, that is, performing a constant current pulse discharge experiment on the lithium battery to obtain experimental data corresponding to, for example, the open circuit voltage Uoc and the SOC, as well as constant current pulse discharge voltage and discharge current experimental data; 3) identifying the parameters of the lithium battery equivalent circuit model in combination with the constant current pulse discharge experiment; 4) using a quasi-sliding mode control algorithm to estimate the lithium battery SOE. Based on the quasi-sliding mode control strategy, during the entire working stage of the lithium battery, different control estimation methods are switched according to the internal characteristics of the battery, which helps to improve the anti-interference ability of the control algorithm, refine the control strategy of the lithium battery control system, integrate the control advantages of multiple algorithms, improve the estimation accuracy of the lithium battery SOE, accelerate the convergence speed of the SOE estimation, and optimize the robust performance of the entire management system.

[0050] Combine Figure 5 and Figure 6As shown in the figure, it can be seen that the estimation of lithium battery SOE based on the sliding mode control algorithm has high accuracy, good robustness and fast convergence speed. Figure 7 As shown in the figure, it can be seen that the estimated prediction results of the lithium-ion battery SOE based on the sliding mode control algorithm are highly consistent with the experimental measured voltage, and the voltage tracking performance is very good.

[0051] In this embodiment, switching different control estimation methods based on the sliding mode control strategy according to the internal characteristics of the battery helps to improve the anti-interference ability of the control algorithm, refine the control strategy of the lithium battery control system, integrate the control advantages of multiple algorithms, improve the estimation accuracy of the lithium battery SOE, accelerate the convergence speed of SOE estimation, and optimize the robust performance of the entire management system.

[0052] Optionally, the method for establishing a lithium battery equivalent circuit model includes:

[0053] Analyze the internal process of the lithium battery and establish a second-order equivalent circuit model, wherein the second-order equivalent circuit model includes a state equation and an observation equation;

[0054] Wherein, the state equation is expressed as:

[0055]

[0056] Wherein, the observation equation is expressed as:

[0057] U L (t) = U OC (t)-i(t)×R0(t)-U1(t)-U2(t)+v(t);

[0058] Among them, SOE0 represents the initial value of the remaining energy of the battery, SOE t Indicates the remaining energy of the battery at time t, Q N represents the rated capacity of the battery, I(t) represents the charge and discharge current at time t, U(t) represents the battery terminal voltage at time t, η represents the charge and discharge efficiency, U L (t) represents the load voltage at time t, U OC (t) represents the open circuit voltage at time t, i(t) represents the discharge current at time t, R0(t) represents the ohmic resistance at time t, U1(t) and U2(t) represent the two polarization voltage values ​​at time t, and v(t) represents the noise at time t;

[0059] Discretizing the second-order equivalent circuit model, and determining a discretized state equation and a discretized observation equation in combination with Kirchhoff's current-voltage law;

[0060] The discretized state equation is expressed as:

[0061]

[0062] The discretized observation equation is expressed as:

[0063] U L (t) = U OC (t)-i(t)×R0(t)-U1(t)-U2(t)+v(t);

[0064] Among them, SOC(k+1) represents the state of charge of the system at time k+1, U1(k+1) and U2(k+1) represent the two polarization voltage values ​​of the system at time k+1, Δt represents the sampling time, R1 and R2 represent the two polarization resistances, R0 represents the ohmic internal resistance, τ1 and τ2 represent the two polarization times, η represents the charge and discharge efficiency, Q N represents the rated capacity of the battery, i(k) represents the discharge current of the system at time k, w(k) and v(k) represent the state and measurement noise of the system at time k.

[0065] Specifically, in this embodiment, a second-order equivalent circuit model of a lithium battery is established, and the state equation of the battery is:

[0066]

[0067] Observation equation: U L (t) = U OC (t)-i(t)×R0(t)-U1(t)-U2(t)+v(t) (2)

[0068] Among them, SOE0 represents the initial value of the remaining energy of the battery; SOE t Indicates the remaining energy of the battery at time t; Q N is the rated capacity of the battery; I(t) represents the charge and discharge current at time t, which is positive when the battery is discharging and negative when it is charging; U(t) represents the battery terminal voltage at time t; η represents the charge and discharge efficiency, which reflects the influence of the discharge rate on the battery SOE, U L (t) represents the load voltage at time t, U OC (t) represents the open circuit voltage at time t, i(t) represents the discharge current at time t, R0(t) represents the ohmic resistance at time t, U1(t) and U2(t) represent the two polarization voltage values ​​at time t, v(t) represents the noise at time t, and the second-order equivalent circuit model of the lithium battery is shown in the figure below: Figure 2 shown.

[0069] Discretizing Equations (1) and (2) and combining them with Kirchhoff's current-voltage law, we get the matrix representation:

[0070] Equation of state:

[0071]

[0072] Observation equation:

[0073] U L (k)=U OC (k)-i(k)×R0(k)-U1(k)-U2(k)+v(k)

[0074] Among them, SOC(k+1) represents the state of charge of the system at time k+1, U1(k+1) and U2(k+1) represent the two polarization voltage values ​​of the system at time k+1, Δt represents the sampling time, R1 and R2 represent the two polarization resistances, R0 represents the ohmic internal resistance, τ1 and τ2 represent the two polarization times, η represents the charge and discharge efficiency, Q N represents the rated capacity of the battery, i(k) represents the discharge current of the system at time k, w(k) and v(k) represent the state and measurement noise of the system at time k, U k (t) represents the load voltage at time k, U OC (k) represents the open circuit voltage at time k, R0(k) represents the ohmic resistance at time k, U1(k) and U2(k) represent the two polarization voltage values ​​at time k.

[0075] Order respectively:

[0076]

[0077]

[0078] In this embodiment, a second-order equivalent circuit model of the lithium battery is established and discretized to complete the basic steps of the algorithm, thereby enabling the application of a sliding mode control strategy.

[0079] Optionally, the constant current pulse discharge experiment is performed on the lithium battery to obtain experimental data, and the relationship curve between the open circuit voltage and the lithium battery SOE is fitted in combination with the experimental data, and the pre-constructed lithium battery equivalent circuit model is identified, including: performing different constant current pulse discharge experiments on the lithium battery to obtain current and voltage experimental data; fitting the relationship curve between the open circuit voltage and the lithium battery SOE in combination with the current and voltage experimental data; and using the recursive least squares method in combination with the experimental data to identify the parameters of the lithium battery equivalent circuit model.

[0080] Specifically, in this embodiment, different constant current pulse discharge experiments are performed on the lithium battery to obtain experimental data corresponding to the open circuit voltage Uoc and SOC as well as constant current pulse discharge voltage and discharge current experimental data. The current and voltage test waveforms of the experimental process are as follows: Figure 3As shown, the fitting equation of Uoc and SOC is: Uoc=193.7619*SOC^8-706.8653*SOC^7+1.0158e +03 *SOC^6-716.8212*SOC^5+239.871*SOC^4-16.7004*SOC^3-12.572*SOC^2+4.7594*SOC+2.8981; the model parameters were identified using the recursive least squares method combined with the constant current pulse discharge experiment, and the average values ​​were obtained: R0=0.0452, R1=0.0220, R2=0.0392, C1=491.3495, C2=8609.4.

[0081] In this embodiment, after different constant current pulse discharge experiments are performed on the lithium battery to obtain experimental data, the recursive least squares method is used to identify the parameters of the lithium battery equivalent circuit model, and then the sliding mode control algorithm can be used to estimate the SOE of the lithium battery.

[0082] Optionally, estimating the lithium battery SOE using a quasi-sliding mode control algorithm based on the parameters of the lithium battery equivalent circuit model includes: determining a state equation of the lithium battery equivalent circuit model according to the identified parameters of the lithium battery equivalent circuit model, establishing a quasi-sliding mode control total relationship according to the state equation and the quasi-sliding mode control algorithm implementation logic, and estimating the lithium battery SOE according to the quasi-sliding mode control total relationship, wherein the quasi-sliding mode control total relationship includes:

[0083] SOE=a*SOE AH +b*SOE EKF +c*SOE SPKF ;

[0084] Where, SOE AH 、SOE EKF and SOE SPKF represent the SOE values ​​calculated by the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively. a, b, and c represent the weighting coefficients corresponding to the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively.

[0085] Specifically, in this embodiment, SOE=a*SOE AH +b*SOE EKF +c*SOE SPKF Where, SOE AH 、SOE EKF and SOE SPKFThe values ​​a, b, and c represent the SOE values ​​calculated by the AH integral method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively. Introducing a quasi-sliding mode control strategy, which switches between different control estimation methods throughout the lithium battery's operating phase based on the battery's internal characteristics, helps improve the control algorithm's anti-interference capability, refines the control strategy of the lithium battery control system, integrates the control advantages of multiple algorithms, and optimizes the robustness of the entire management system.

[0086] Among them, the implementation logic of the sliding mode control algorithm refers to the logic of implementing the sliding mode control algorithm. Generally speaking, the algorithm represents a strategy mechanism for solving problems in a systematic way, and its implementation logic represents the operating principle of the algorithm from input to output. Therefore, the implementation logic of the sliding mode control algorithm represents the principle, mode and method of using the sliding mode control algorithm to solve the corresponding problem.

[0087] In this embodiment, the SOE of the lithium battery is estimated based on the sliding mode control general relationship, the control advantages of multiple algorithms are integrated, the estimation accuracy of the SOE of the lithium battery is improved, the convergence speed of the SOE estimation is accelerated, and the robust performance of the entire management system is optimized.

[0088] Optionally, the lithium battery SOE estimation method further includes: determining a, b, and c according to the lithium battery SOC and the terminal voltage difference:

[0089] When SOC>90% and the terminal voltage difference is greater than 10mV, a=0,b=1-c,c=(U EKF -U L ) / (U L );

[0090] When 5%<SOC≤90%, a=0,b=(U SPKF -U L ) / (U L ), c = 1-b;

[0091] When SOC≤5%, a=1-bc,b=(U EKF -U L ) / (U L ), c=(U SPKF -U L ) / (U L );

[0092] Among them, U L 、U EKF and U SPKF They represent the load voltage, the voltage calculated by the extended Kalman filter algorithm, and the voltage calculated by the improved Sigma point filter algorithm respectively.

[0093] Specifically, in this embodiment, in the sliding mode control general relationship (U L 、U EKF and U SPKF Represent the load voltage, the voltage calculated by the extended Kalman filter algorithm, and the voltage calculated by the improved Sigma point filter algorithm respectively):

[0094] 1) When SOC>90% and the terminal voltage difference is greater than 10mV, a=0,b=1-c,c=(U EKF -U L ) / (U L );

[0095] 2) When 5%<SOC≤90%, a=0,b=(U SPKF -U L ) / (U L ), c = 1-b;

[0096] 3) When SOC≤5%, the AH integral method is mainly used to calculate SOE, at this time a=1-bc, b=(U EKF -U L ) / (U L ), c=(U SPKF -U L ) / (U L ).

[0097] In this embodiment, a, b, and c are determined based on the lithium battery SOC and the terminal voltage difference, and then the lithium battery SOE is estimated based on the sliding mode control-like total relationship. The control advantages of multiple algorithms are integrated to improve the estimation accuracy of the lithium battery SOE, accelerate the convergence speed of the SOE estimation, and optimize the robust performance of the entire management system.

[0098] Optionally, the improved Sigma point filtering algorithm includes: initializing state variables, using the String interpolation method to calculate the Sigma points and their corresponding weights according to the symmetric sampling strategy, and using the UT transformation method (Unscented Transformation, lossless transformation) to calculate the Sigma points and their corresponding weights according to the symmetric sampling strategy; calculating the one-step prediction of the Sigma point set, calculating the one-step prediction and covariance matrix of the state variables, substituting the one-step prediction into the observation equation to obtain the predicted observation; determining the observation prediction value of the Sigma point set, and obtaining the system prediction mean and covariance by weighted summation; calculating the Kalman gain matrix, and calculating the state update and covariance update of the system.

[0099] Specifically, in this embodiment, combined with Figure 4As shown in the figure, first, the state estimate at time k+1 is obtained by sampling the prior estimate at time k to generate 2n+1 sigma points, where n is the dimension of the state equation, and then by weighted calculation. In order to eliminate the problems of Sigma point depletion and degradation, two Sigma point generation methods are introduced, and the number of Sigma points is increased to 4n+1. The specific recursive formula is as follows:

[0100] (1) State variable initialization:

[0101]

[0102]

[0103] (2-1) The String interpolation method calculates Sigma points and their corresponding weights according to the symmetric sampling strategy:

[0104]

[0105]

[0106]

[0107]

[0108]

[0109] (2-2) The UT transformation method calculates the Sigma points and their corresponding weights according to the symmetric sampling strategy:

[0110]

[0111]

[0112]

[0113] ω m (0) =κ / (n+κ)

[0114] ω c (0) =κ / (n+κ)

[0115] ω m (i) =ω c (i) =1 / {2(n+κ)},i=1~2n

[0116] (3) Calculate the one-step prediction of the Sigma point set;

[0117] γ i,k+1|k =f k(ξ i,k ),i=0,1,…,4n+1

[0118] (4) Calculate the one-step prediction and covariance matrix of the state variables;

[0119]

[0120]

[0121] (5) Substitute the one-step prediction into the observation equation to obtain the predicted observation;

[0122] χ i,k+1|k =h k+1 (ξ i,k+1|k ),i=0,1,…,4n+1

[0123] (6) Obtain the observed predicted value of the Sigma point set, and obtain the system prediction mean and covariance through weighted summation;

[0124]

[0125]

[0126]

[0127] (7) Calculate the Kalman gain matrix;

[0128]

[0129] (8) Finally, the state update and covariance update of the system are calculated;

[0130]

[0131]

[0132] In this embodiment, an improved Sigma point filtering algorithm is used, which can integrate the control advantages of multiple algorithms, improve the estimation accuracy of lithium battery SOE, accelerate the convergence speed of SOE estimation, and optimize the robust performance of the entire management system.

[0133] Optionally, the improved Sigma point filtering algorithm is based on a hybrid Sigma point filtering strategy, introduces two Sigma point generation methods, and increases the number of Sigma points to 4n+1, where n represents the dimension of the state equation.

[0134] Specifically, in this embodiment, based on the hybrid Sigma point filtering strategy, two Sigma point generation methods are introduced, and the number of Sigma points is increased to 4n+1, thereby increasing the richness of the Sigma point set, reducing the excessive weight of individual points in the Sigma point set, and avoiding the impoverishment and degradation problems of the Sigma point algorithm.

[0135] In this embodiment, based on the hybrid Sigma point filtering strategy, two Sigma point generation methods are introduced, and the number of Sigma points is increased to 4n+1, which increases the richness of the Sigma point set, reduces the excessive weight of individual points in the Sigma point set, avoids the depletion and degradation problems of the Sigma point algorithm, and is beneficial to improving the estimation accuracy of the SOE of lithium batteries.

[0136] Another embodiment of the present invention provides a lithium battery SOE estimation device, including: an experimental module, used to obtain experimental data by performing a constant current pulse discharge experiment on the lithium battery, and fit the relationship curve between the open circuit voltage and the lithium battery SOE based on the experimental data, and identify a pre-constructed lithium battery equivalent circuit model to obtain the parameters of the lithium battery equivalent circuit model; an estimation module, used to estimate the lithium battery SOE based on the parameters of the lithium battery equivalent circuit model using a sliding mode control algorithm.

[0137] Another embodiment of the present invention provides a lithium battery SOE estimation system, including a computer-readable storage medium storing a computer program and a processor. When the computer program is read and executed by the processor, the above lithium battery SOE estimation method is implemented.

[0138] Another embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program. When the computer program is read and executed by a processor, the above-mentioned lithium battery SOE estimation method is implemented.

[0139] Although the present invention is disclosed as above, the scope of protection disclosed by the present invention is not limited thereto. Those skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A lithium battery SOE estimation method, characterized in that: include: The SOE of the lithium battery is estimated using a sliding mode control algorithm based on parameters of a lithium battery equivalent circuit model; the parameters of the lithium battery equivalent circuit model are obtained by obtaining experimental data through a constant current pulse discharge experiment on the lithium battery, and fitting a relationship curve between the open circuit voltage and the lithium battery SOE based on the experimental data to identify a pre-constructed lithium battery equivalent circuit model; The method of estimating the SOE of a lithium battery using a quasi-sliding mode control algorithm based on the parameters of a lithium battery equivalent circuit model includes: determining a state equation of the lithium battery equivalent circuit model according to the identified parameters of the lithium battery equivalent circuit model; establishing a quasi-sliding mode control total relational expression according to the state equation and the quasi-sliding mode control algorithm implementation logic; and estimating the SOE of the lithium battery according to the quasi-sliding mode control total relational expression. The quasi-sliding mode control total relational expression includes: SOE=a*SOE AH +b*SOE EKF +c*SOE SPKF ; Where, SOE AH 、SOE EKF and SOE SPKF They represent the SOE values ​​calculated by the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively. a, b, and c represent the weighting coefficients corresponding to the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively. The lithium battery SOE estimation method further includes: determining a, b, and c according to the lithium battery SOC and the terminal voltage difference: When SOC>90% and the terminal voltage difference is greater than 10mV, a=0,b=1-c,c=(U EKF -U L ) / (U L ); When 5%<SOC≤90%, a=0, b=(U SPKF -U L ) / (U L ), c = 1 - b; When this SOC≤5%, a=1-bc, b=(U EKF -U L ) / (U L ), c = (U SPKF -U L ) / (U L ); Among them, U L 、U EKF and U SPKF They represent the load voltage, the voltage calculated by the extended Kalman filter algorithm, and the voltage calculated by the improved Sigma point filter algorithm respectively.

2. The lithium battery SOE estimation method according to claim 1, characterized in that: The improved Sigma point filtering algorithm includes: Initialize the state variables, use the String interpolation method to calculate the Sigma points and their corresponding weights according to the symmetric sampling strategy, and use the UT transformation method to calculate the Sigma points and their corresponding weights according to the symmetric sampling strategy; Calculate the one-step prediction of the Sigma point set, calculate the one-step prediction and covariance matrix of the state variable, substitute the one-step prediction into the observation equation to obtain the predicted observation; Determine the observed prediction value of the Sigma point set, and obtain the system prediction mean and covariance through weighted summation; Calculate the Kalman gain matrix and calculate the state update and covariance update of the system.

3. The lithium battery SOE estimation method according to claim 2, characterized in that: The improved Sigma point filtering algorithm is based on a hybrid Sigma point filtering strategy, introduces two Sigma point generation methods, and increases the number of Sigma points to 4n+1, where n represents the dimension of the state equation.

4. The lithium battery SOE estimation method according to claim 1, characterized in that: The method for constructing the lithium battery equivalent circuit model comprises: Analyze the internal process of the lithium battery and establish a second-order equivalent circuit model, wherein the second-order equivalent circuit model includes a state equation and an observation equation; Wherein, the state equation is expressed as: Wherein, the observation equation is expressed as: U L (t)=U OC (t)-i(t)×R0(t)-U1(t)-U2(t)+v(t); Among them, SOE0 represents the initial value of the remaining energy of the battery, SOE t Indicates the remaining energy of the battery at time t, Q N represents the rated capacity of the battery, I(t) represents the charge and discharge current at time t, U(t) represents the battery terminal voltage at time t, η represents the charge and discharge efficiency, U L (t) represents the load voltage at time t, U OC (t) represents the open circuit voltage at time t, i(t) represents the discharge current at time t, R0(t) represents the ohmic resistance at time t, U1(t) and U2(t) represent the two polarization voltage values ​​at time t, and v(t) represents the noise at time t; Discretizing the second-order equivalent circuit model, and determining a discretized state equation and a discretized observation equation in combination with Kirchhoff's current-voltage law; The discretized state equation is expressed as: The discretized observation equation is expressed as: U L (k)=U OC (k)-i(k)×R0(k)-U1(k)-U2(k)+v(k); Among them, SOC(k+1) represents the state of charge of the system at time k+1, U1(k+1) and U2(k+1) represent the two polarization voltage values ​​of the system at time k+1, Δt represents the sampling time, R1 and R2 represent the two polarization resistances, R0 represents the ohmic internal resistance, τ1 and τ2 represent the two polarization times, η represents the charge and discharge efficiency, Q N represents the rated capacity of the battery, i(k) represents the discharge current of the system at time k, w(k) and v(k) represent the state and measurement noise of the system at time k, U k (t) represents the load voltage at time k, U OC (k) represents the open circuit voltage at time k, R0(k) represents the ohmic resistance at time k, U1(k) and U2(k) represent the two polarization voltage values ​​at time k.

5. The lithium battery SOE estimation method according to claim 4, characterized in that: The constant current pulse discharge experiment on the lithium battery is performed to obtain experimental data, and the relationship curve between the open circuit voltage and the lithium battery SOE is fitted based on the experimental data to identify the pre-constructed lithium battery equivalent circuit model. The method includes: Performing different constant current pulse discharge experiments on the lithium battery to obtain current and voltage experimental data; Fitting the relationship curve between the open circuit voltage and the SOE of the lithium battery based on the current and voltage experimental data; The parameters of the lithium battery equivalent circuit model are identified by using the recursive least squares method combined with the experimental data.

6. A lithium battery SOE estimation device, characterized in that: include: The experimental module is used to obtain experimental data by performing a constant current pulse discharge experiment on the lithium battery, and to fit the relationship curve between the open circuit voltage and the lithium battery SOE based on the experimental data, and to identify the pre-constructed lithium battery equivalent circuit model to obtain the parameters of the lithium battery equivalent circuit model; An estimation module, configured to estimate the SOE of the lithium battery using a sliding mode control algorithm based on parameters of a lithium battery equivalent circuit model; The method of estimating the SOE of a lithium battery using a quasi-sliding mode control algorithm based on the parameters of a lithium battery equivalent circuit model includes: determining a state equation of the lithium battery equivalent circuit model according to the identified parameters of the lithium battery equivalent circuit model; establishing a quasi-sliding mode control total relational expression according to the state equation and the quasi-sliding mode control algorithm implementation logic; and estimating the SOE of the lithium battery according to the quasi-sliding mode control total relational expression. The quasi-sliding mode control total relational expression includes: SOE=a*SOE AH +b*SOE EKF +c*SOE SPKF ; Where, SOE AH 、SOE EKF and SOE SPKF They represent the SOE values ​​calculated by the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively. a, b, and c represent the weighting coefficients corresponding to the AH integration method, the extended Kalman filter algorithm, and the improved Sigma point filter algorithm, respectively. The lithium battery SOE estimation method further includes: determining a, b, and c according to the lithium battery SOC and the terminal voltage difference: When SOC>90% and the terminal voltage difference is greater than 10mV, a=0,b=1-c,c=(U EKF -U L ) / (U L ); When 5%<SOC≤90%, a=0, b=(U SPKF -U L ) / (U L ), c = 1 - b; When this SOC≤5%, a=1-bc, b=(U EKF -U L ) / (U L ), c = (U SPKF -U L ) / (U L ); Among them, U L 、U EKF and U SPKF They represent the load voltage, the voltage calculated by the extended Kalman filter algorithm, and the voltage calculated by the improved Sigma point filter algorithm respectively.

7. A lithium battery SOE estimation system, characterized in that: The invention comprises a computer-readable storage medium storing a computer program and a processor, wherein when the computer program is read and executed by the processor, the method for estimating the SOE of a lithium battery according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is read and executed by a processor, the method for estimating the SOE of a lithium battery according to any one of claims 1 to 5 is implemented.

Images