Extended Kalman filter lithium battery SOC estimation method based on ampere-hour correction

Abstract

The present invention relates to an extended Kalman filter lithium battery SOC estimation method based on ampere-hour correction, which belongs to lithium battery technology and solves the problem that parameter deviation in the extended Kalman filter estimation process in the prior art causes significant SOC estimation error, especially low estimation accuracy in the low SOC segment. It includes: establishing a first-order Thevenin equivalent circuit model, identifying parameters in the circuit model by recursive least squares method with genetic factors, calculating the SOC ampere-hour method estimation value at the current moment based on the final SOC estimation value at the previous moment in the iterative estimation process of the extended Kalman filter algorithm, calculating the ampere-hour correction coefficient based on the optimal SOC estimation value at the current moment, and then weighting the optimal SOC estimation value at the current moment and the SOC ampere-hour method estimation value according to the ampere-hour correction coefficient to calculate the final SOC estimation value at the current moment, and bring it into the estimation at the next moment. It realizes that the correction strength is gradually increased as the SOC decreases, reduces the error caused by the model parameters, and improves the estimation accuracy of the entire SOC segment.

Description

Technical Field

[0001] The present invention relates to the technical field of lithium batteries, and in particular to an extended Kalman filter lithium battery SOC estimation method based on ampere-hour correction. Background Art

[0002] Lithium batteries have become one of the main energy sources for electric vehicles due to their high energy density and high power density. As one of the three core components, they provide energy for the entire vehicle. Therefore, the performance of lithium batteries determines the performance of the entire vehicle. They have become the focus of research in the field of electric vehicles and have huge potential markets and application prospects. In the application technology of lithium batteries, battery management technology is the most critical step in determining the engineering performance of lithium batteries. The estimation of the state of charge (SOC) of lithium batteries is the core of battery management technology.

[0003] At present, equivalent circuit models include Randall model, impedance model, first-order Thevenin equivalent circuit model, and second-order Thevenin equivalent circuit model. The more common lithium battery SOC estimation methods include ampere-hour method and extended Kalman filter (EKF) method.

[0004] The nonlinear characteristics of the Randall model are strong, which limits its application. The impedance model only considers the voltage-current external characteristics of the battery, and the second-order Thevenin equivalent circuit has a large amount of calculation and is generally not practical. The ampere-hour method calculates the current SOC value by integrating the charging and discharging current of the lithium battery. Since this method has no feedback link, the error of the current sampling sensor will gradually accumulate during the estimation process, thereby affecting the accuracy of the SOC estimation. The EKF algorithm corrects the SOC observation results by the error between the predicted value and the actual voltage value of the battery terminal voltage. It has a high dependence on the modeling accuracy of the battery. When there is a certain deviation in the battery parameters, the estimation results will produce a more obvious error. Some studies or estimation methods propose to connect the recursive least squares (RLS) parameter identification in series before the EKF to observe the equivalent circuit model parameters of the battery online. However, the actual working conditions of lithium batteries are mostly DC input and output, which cannot reflect the battery dynamics. Introducing RLS online identification will introduce a large deviation to the model parameters. Summary of the invention

[0005] In view of the above analysis, an embodiment of the present invention aims to provide an extended Kalman filter lithium battery SOC estimation method based on ampere-hour correction, so as to solve the problem of significant SOC estimation error caused by parameter deviation in the EKF estimation process, especially low estimation accuracy in the low SOC segment.

[0006] The embodiment of the present invention provides a method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction, comprising:

[0007] Establish a first-order Thevenin equivalent circuit model;

[0008] The first-order Thevenin equivalent circuit model is Laplace transformed, and the battery ohmic internal resistance, battery polarization resistance and battery polarization capacitance of the circuit model are identified by recursive least squares method with genetic factors.

[0009] Discretize the first-order Thevenin equivalent circuit model, combine it with the ampere-hour integral equation, take the current as input and the battery terminal voltage as output, establish the state space equation with the state of charge SOC and the battery internal resistance terminal voltage as state variables, and establish the extended Kalman filter;

[0010] During the iterative estimation process of the extended Kalman filter algorithm, the current SOC ampere-hour estimation value is calculated based on the final SOC estimation value at the previous moment, the ampere-hour correction coefficient is calculated based on the current SOC optimal estimation value, and then the current SOC optimal estimation value and the SOC ampere-hour estimation value are weighted according to the ampere-hour correction coefficient to calculate the current SOC final estimation value, which is used as the initial value for the next moment to iterate forward.

[0011] Furthermore, the iterative estimation process of the extended Kalman filter algorithm includes:

[0012] Set the initial values ​​of state variables and state mean square errors in the state space equations;

[0013] Time update, calculate the estimated value of the state variables and the estimated value of the state mean square error at the current moment;

[0014] Calculate Kalman gain;

[0015] Calculate the optimal estimated value of the state variable and the optimal estimated value of the state mean square error at the current moment, and extract the optimal estimated value of the SOC at the current moment from the optimal estimated value of the state variable;

[0016] Calculate the SOC ampere-hour method estimated value at the current moment based on the final estimated value of SOC at the previous moment;

[0017] Calculate the current ampere-hour correction coefficient according to the current optimal estimated SOC value;

[0018] The optimal estimated value of SOC at the current moment and the estimated value of SOC by ampere-hour method are weighted to calculate the final estimated value of SOC at the current moment;

[0019] Determine whether the stop condition is reached at the current moment. If not, return to the time update, calculate the estimated value of the state variable and the estimated value of the state mean square error at the current moment, and start the iterative advancement of the next moment.

[0020] Furthermore, the ampere-hour correction coefficient is obtained by dynamically adjusting the optimal estimated value of SOC at the current moment during the estimation process at each moment using the following formula:

[0021]

[0022] Among them, k c ∈(0,1), k c,max k c The maximum value of the parameter, β is the slope coefficient, is the optimal estimated value of SOC at time k, SOC thred k c The SOC threshold, i.e., k c The SOC value corresponding to when the value reaches 1 / 2 of its maximum value.

[0023] Furthermore, the SOC ampere-hour method estimated value is calculated based on the final estimated value of SOC at the previous moment, and the formula is:

[0024]

[0025] Among them, SOC k,ah is the estimated value of SOC by ampere-hour method at time k, SOC k-1,cor is the final estimated value of SOC at time k-1, η is the coulomb efficiency, C is the rated capacity of the battery, Δt is the estimated time interval, i o,k is the measured value of the battery output current at time k.

[0026] Furthermore, the weighted calculation formula of the current SOC optimal estimated value and the SOC ampere-hour method estimated value is:

[0027]

[0028] in, is the optimal estimated value of SOC at time k, SOC k,cor is the final estimated value of SOC at time k, which will be brought into the extended Kalman filter estimation process at time k+1.

[0029] Furthermore, when performing SOC estimation at the first moment, the final estimation value at the current moment is equal to the optimal estimation value at the current moment.

[0030] Furthermore, the expression of the first-order Thevenin equivalent circuit model is:

[0031]

[0032] u o =u ocv -i o R0-u1

[0033] Among them, u ocv is the open circuit voltage, R0 is the battery ohmic internal resistance, R1 and C1 are the battery polarization resistance and capacitance, u o ,i o are the battery terminal voltage and current, u1 is the battery internal resistance terminal voltage, is the derivative of the voltage with respect to time, that is, the rate of change of the voltage.

[0034] Furthermore, the battery is subjected to DC charging and DC discharging experiments, and the average value of the charging voltage changing with SOC and the discharging voltage changing with SOC is taken as the open circuit voltage value u ocv , and then curve fitting is performed to obtain the OCV-SOC curve;

[0035] The open circuit voltage u ocv Perform polynomial fitting with the corresponding SOC value to establish the fitting formula of the battery open circuit voltage and SOC function:

[0036] u ocv (SOC) = a0 + a1SOC + a2SOC 2 +a3SOC 3 +a4SOC 4 +a5SOC 5 +a6SOC 6 +a7SOC 7 +a8SOC 8 +a9SOC 9

[0037] Among them, a i are polynomial fitting coefficients, i=0,1,…,9, and SOC is the state of charge value.

[0038] Furthermore, according to the least squares theory, the first-order Thevenin equivalent circuit model is Laplace transformed to obtain:

[0039]

[0040] Among them, τ1=R1C1 is the time constant of the battery model, and s is the laplace operator;

[0041] The test data of the charge and discharge experiment is obtained, and the forgetting factor is introduced in the process of recursive least squares method parameter identification to identify the battery ohmic internal resistance R0, battery polarization resistance R1, and battery polarization capacitance C1.

[0042] Furthermore, the state space equation used in the extended Kalman filter algorithm is:

[0043] f(x k ,u k )=A k xk +B k u k

[0044] g(x k ,u k )=C k x k +D k u k

[0045] Among them, x k =[SOC k ,u 1,k ] T , represents the state variable at time k, u k Represents the input variable of the system, and the coefficient A in the equation k , B k , C k , D k The expression is:

[0046]

[0047] Where C is the rated capacity of the battery, τ1=R1C1 is the time constant of the battery model, η is the coulomb efficiency, Δt is the estimated time interval, and u ocv (SOC) is the open circuit voltage, and R0 is the battery ohmic internal resistance.

[0048] Compared with the prior art, the present invention can at least achieve the following beneficial effects:

[0049] Taking into account the large error of the extended Kalman filter algorithm in the low SOC segment, the sigma function is introduced to define the ampere-hour correction coefficient, and the ampere-hour correction coefficient is adjusted in real time according to the optimal estimated value of SOC at each moment, so that the correction strength gradually increases with the decrease of SOC; the final SOC estimation result at each moment is obtained by weighting the SOC optimal estimation result and the SOC ampere-hour method estimation value using the ampere-hour correction coefficient, which reduces the error caused by the model parameters and improves the estimation accuracy of the entire SOC segment.

[0050] In the present invention, the above-mentioned technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages can become obvious from the description, or can be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the contents particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings are only for the purpose of illustrating particular embodiments and are not to be considered limiting of the present invention. Like reference symbols denote like components throughout the drawings.

[0052] Figure 1 This is a flow chart of a method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction in an embodiment of the present invention;

[0053] Figure 2 It is an OCV-SOC curve diagram in an embodiment of the present invention;

[0054] Figure 3 This is a polynomial fitting result diagram in an embodiment of the present invention;

[0055] Figure 4 A curve diagram showing the battery terminal voltage and SOC changing with time in the parameter identification experiment in an embodiment of the present invention;

[0056] Figure 5 This is a curve diagram of current and SOC changing with time in the parameter identification experiment in an embodiment of the present invention;

[0057] Figure 6 This is a diagram showing the fitting results of the parameters of the equivalent circuit model in an embodiment of the present invention;

[0058] Figure 7 is an ampere-hour correction factor curve diagram in an embodiment of the present invention;

[0059] Figure 8 The block diagram of the extended Kalman filter algorithm based on ampere-hour correction in an embodiment of the present invention;

[0060] Fig. 9 A battery testing platform in an embodiment of the present invention;

[0061] Fig.10 This is a graph of SOC results estimated by EKF based on ampere-hour correction in an embodiment of the present invention;

[0062] Fig.11 This is a diagram comparing the errors between the SOC value estimated by the EKF based on the ampere-hour correction and the SOC value estimated by the original EKF in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0064] A specific embodiment of the present invention discloses a method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction, such as Figure 1 As shown, the following steps are included:

[0065] S10: Establishing a first-order Thevenin equivalent circuit model of the lithium battery to be estimated;

[0066] Specifically, according to the basic circuit principle, the mathematical expression of the first-order Thevenin equivalent circuit model is:

[0067]

[0068] Among them, u ocv is the open circuit voltage, R0 is the battery ohmic internal resistance, R1 and C1 are the battery polarization resistance and capacitance, u o ,i o are the battery terminal voltage and current, u1 is the battery internal resistance terminal voltage, is the derivative of the voltage with respect to time, that is, the rate of change of the voltage.

[0069] The open circuit voltage u ocv There is a functional relationship with the SOC to be estimated. Preferably, the functional relationship is obtained by the following method:

[0070] First, at the same temperature, the battery is subjected to DC charging and DC discharging experiments, and the average value of the charging voltage changing with SOC and the discharging voltage changing with SOC is taken as the open circuit voltage value u ocv , and then curve fitting is performed to obtain the OCV-SOC curve;

[0071] For example, at an ambient temperature of 25°C, the battery is DC charged and discharged at 0.5C, and the charging voltage u 0.5C,cha (SOC) changes with SOC value and discharge voltage u 0.5C,dis The average value of SOC changes with SOC is taken as the open circuit voltage value u ocv , as shown in the following formula, and then curve fitting is performed to obtain the OCV-SOC curve, as shown in Figure 2 shown.

[0072]

[0073] Then, based on the obtained OCV-SOC curve, the open circuit voltage u ocv Perform polynomial fitting with the corresponding SOC value to establish the fitting formula of the battery open circuit voltage and SOC function:

[0074] u ocv (SOC) = a0 + a1SOC + a2SOC 2 +a3SOC 3 +a4SOC 4 +a5SOC 5 +a6SOC 6 +a7SOC 7 +a8SOC 8 +a9SOC 9 Formula (3)

[0075] Among them, a i , i = 0, 1, ..., 9 are polynomial fitting coefficients, and SOC is the state of charge value. For example, the polynomial fitting coefficients in the following table are used to fit the curve, and the fitting effect is as follows: Figure 3 shown.

[0076] Fitting coefficients Fitted values <![CDATA[a0]]> 1.854673 <![CDATA[a1]]> 0.083164 <![CDATA[a2]]> -0.011283 <![CDATA[a3]]> 0.000807 <![CDATA[a4]]> <![CDATA[-3.288325×10 -5 ]]> <![CDATA[a5]]> <![CDATA[8.062971×10 -7 ]]> <![CDATA[a6]]> <![CDATA[-1.213033×10 -8 ]]> <![CDATA[a7]]> <![CDATA[1.095774×10 -10 ]]> <![CDATA[a8]]> <![CDATA[-5.453389×10 -13 ]]> <![CDATA[a9]]> <![CDATA[1.148920×10 -15 ]]>

[0077] S11: Perform Laplace transform on the first-order Thevenin equivalent circuit model, and identify the battery ohmic internal resistance, battery polarization resistance and battery polarization capacitance of the circuit model by recursive least squares method with genetic factors;

[0078] Specifically, according to the least squares theory, the first-order Thevenin equivalent circuit model is Laplace transformed to obtain:

[0079]

[0080] Among them, τ1=R1C1 is the time constant of the battery model, and s is the laplace operator;

[0081] set up:

[0082]

[0083] The transfer function of y with respect to x can be obtained:

[0084]

[0085] According to the bilinear transformation formula:

[0086]

[0087] Where T is the discrete sampling time interval, the discretization expression of formula (6) can be obtained:

[0088]

[0089] in:

[0090]

[0091] Considering that the recursive least squares method has infinite memory, the influence of old data on the recognition result accumulates continuously, making it impossible to play the correction role of new data, affecting the accuracy of subsequent recognition. In order to eliminate the influence of old data on the recognition result, in a preferred embodiment of the present invention, a forgetting factor λ is introduced in the recognition process, usually taking 0.95≤λ<=1, to accelerate the convergence speed and improve the tracking performance.

[0092] Exemplarily, a charge and discharge experiment is performed on the battery according to the following steps, and parameter values ​​are identified based on data recorded in the experiment:

[0093] a) Charge the battery to SOC=100% according to the standard charging process and leave it for more than 8 hours;

[0094] b) Discharge the battery with a 1C current until the battery SOC drops by 10%, and leave it alone for 15 minutes;

[0095] c) Repeat step b) until the battery SOC = 0%;

[0096] d) Record the battery terminal voltage, current and SOC change curves and data over time during the discharge process, as follows: Figure 4 , Figure 5 shown.

[0097] e) Taking λ=0.99, according to the recursive least square method with forgetting factor, the battery model parameters are identified as battery ohmic internal resistance R0=0.0731, battery polarization resistance R1=0.0917, and battery polarization capacitance C1=356.6362.

[0098] Substituting the parameter values ​​into the equivalent circuit model (1), the battery terminal voltage fitting result is obtained as follows: Figure 6 As shown, it can be seen that the identified parameter values ​​have a good fit to the circuit model.

[0099] S12: Discretize the first-order Thevenin equivalent circuit model, combine it with the ampere-hour integral equation, take the current as input and the battery terminal voltage as output, establish a state space equation with the state of charge SOC and the battery internal resistance terminal voltage as state variables, and establish an extended Kalman filter;

[0100] Specifically, the extended Kalman filter is based on the Kalman filter, which Taylor expands the nonlinear system at the operating point and takes the linear part for state estimation. The state space equation of the nonlinear discrete system is:

[0101]

[0102] Among them, x k With y k are the state variables and output variables of the system at time k, u k represents the input variable of the system, w k represents the process noise, w k ~N(0,R k ), v k represents the measurement noise, v k ~N(0,Q k ), both are uncorrelated zero-mean white noise.

[0103] The state variable x in formula (10) is k With the output variable yk The expression is:

[0104]

[0105] According to the state space equation of the battery equivalent model, equation (10) is linearly transformed into:

[0106]

[0107] Among them, the coefficient A k , B k , C k , D k The expression is:

[0108]

[0109] Where C is the rated capacity of the battery, τ1=R1C1 is the time constant of the battery model, η is the coulomb efficiency, Δt is the estimated time interval, and u ocv (SOC) is the open circuit voltage, and R0 is the battery ohmic internal resistance.

[0110] S13: During the iterative estimation process of the extended Kalman filter algorithm, the current SOC ampere-hour estimation value is calculated based on the final SOC estimation value at the previous moment, the ampere-hour correction coefficient is calculated based on the current SOC optimal estimation value, and then the current SOC optimal estimation value and the SOC ampere-hour estimation value are weighted according to the ampere-hour correction coefficient to calculate the current SOC final estimation value, which is used as the initial value for the next moment to iterate forward.

[0111] Specifically, the extended Kalman filter algorithm includes: state variables State mean square error P k , process noise covariance Q k , measurement noise covariance R k , Kalman gain L k , in the estimation process, represents the prior estimate of the state variable at time k, that is, the optimal estimate of the state variable at time k-1, which is the estimate without Kalman gain correction at time k. represents the optimal estimate of the state variable at time k, represents the mean square error of the prior state at time k, represents the optimal estimate of the mean square error of the state at time k. The block diagram of the extended Kalman filter algorithm based on ampere-hour correction is as follows: Figure 8 The specific estimation process is as follows:

[0112] S131: Set the initial values ​​of the state variables and the state mean square error in the state space equation, that is, k=0. The formula is:

[0113]

[0114] It should be noted that the algorithm in this embodiment is not affected by the initial value of SOC, and will converge to a value close to the true value after iteration. For example, SOC randomly takes a value, which is generally a common value or a value taken in the last test.

[0115] S132: Time update, calculate the estimated value of the state variable and the estimated value of the state mean square error at the current moment, the formula is:

[0116]

[0117]

[0118] S133: Calculate the Kalman gain. The formula is:

[0119]

[0120] S134: Calculate the optimal estimated value of the state variable and the optimal estimated value of the state mean square error at the current moment through equations (18) and (19), and Extract the optimal estimated SOC value at the current moment

[0121]

[0122]

[0123] Where I is the identity matrix.

[0124] S135: Calculate the current SOC ampere-hour estimation value based on the final SOC estimation value at the previous moment, using the formula:

[0125]

[0126] Among them, SOC k,ah is the estimated value of SOC by ampere-hour method at time k, SOC k-1,cor is the final estimated value of SOC at time k-1, η is the coulomb efficiency, C is the rated capacity of the battery, Δt is the estimated time interval, i o,k is the measured current output by the battery at time k.

[0127] S136: Calculate the current ampere-hour correction coefficient according to the current optimal estimated SOC value, and the formula is:

[0128]

[0129] Among them, k c ∈(0,1), k c,max kc The maximum value of the parameter, β is the slope coefficient, is the optimal estimated value of SOC at time k, SOC thred k c The SOC threshold, i.e., k c The SOC value corresponding to when the value reaches 1 / 2 of its maximum value.

[0130] For example, in Figure 7 k is given in c In this case, the horizontal axis is the SOC value and the vertical axis is k c It can be seen that k c The SOC threshold is 50%. When the SOC is lower than 50%, k c The growth rate is faster, especially when SOC < 40%, the correction factor k c It gradually increases with the decrease of SOC and is close to 0.8, which means that the correction force gradually increases with the decrease of SOC.

[0131] S137: The optimal estimated value of SOC at the current moment and the estimated value of SOC by ampere-hour method are weighted to calculate the final estimated value of SOC at the current moment. The formula is:

[0132]

[0133] in, is the optimal estimated value of SOC at time k, SOC k,cor is the final estimated value of SOC at time k, which will be brought into the extended Kalman filter estimation process at time k+1.

[0134] Determine whether the stop condition is met at the current moment. If not, return to step S312 to start the iterative advancement at the next moment.

[0135] For example, a lithium titanate battery (2.5V, 21Ah) was subjected to a DST condition test, and the test platform was as follows: Fig. 9 The tested battery is placed in a thermostat, and the ambient temperature is kept constant at 25°C. Digatron BTS-600 is used to test the battery and record the data, and then the host computer uses the battery data to estimate the SOC.

[0136] The estimated results are as follows Fig.10As shown, the straight line is the actual SOC value change of the battery under DST conditions, calculated by the BTS-600 test platform. The short horizontal line is the EKF estimation result. It can be seen that it has a higher accuracy when SOC ≥ 50%, but there is a more obvious estimation deviation when SOC < 50%, and the deviation gradually increases when SOC < 20%. This is because when the SOC is low, the model parameters of the battery will produce more obvious fluctuations. The dotted line is the EKF estimation result based on ampere-hour correction. It can be seen that it has good accuracy in the entire SOC range.

[0137] Fig.11 The figure shows the comparison of SOC estimation errors between the original EKF and the EKF based on ampere-hour correction under DST conditions. It can be seen that the estimation error of the original EKF in the low SOC segment is relatively large, which can be close to 6%, while the estimation error of the EKF based on ampere-hour correction can be controlled within 2%, which has higher estimation accuracy.

[0138] Compared with the prior art, the extended Kalman filter lithium battery SOC estimation method based on ampere-hour correction provided in this embodiment takes into account that the original extended Kalman filter algorithm has a large error in the low SOC segment, so a sigma function is introduced to define the ampere-hour correction coefficient, and the ampere-hour correction coefficient is adjusted in real time according to the optimal estimated value of SOC at each moment, so that the correction strength is gradually increased as the SOC decreases; the final SOC estimation result at each moment is obtained by weighting the SOC optimal estimation result and the SOC ampere-hour method estimation value using the ampere-hour correction coefficient, thereby reducing the error caused by the model parameters and improving the estimation accuracy of the entire SOC segment.

[0139] Those skilled in the art will appreciate that all or part of the processes of the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the program can be stored in a computer-readable storage medium, wherein the computer-readable storage medium is a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0140] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for estimating SOC of lithium battery using extended Kalman filter based on ampere-hour correction, characterized in that: The steps include: Establish a first-order Thevenin equivalent circuit model; Performing Laplace transformation on the first-order Thevenin equivalent circuit model, and identifying the battery ohmic internal resistance, battery polarization resistance, and battery polarization capacitance of the circuit model by a recursive least squares method with genetic factors; Discretize the first-order Thevenin equivalent circuit model, combine it with the ampere-hour integral equation expression, take the current as input and the battery terminal voltage as output, establish a state space equation with the state of charge SOC and the battery internal resistance terminal voltage as state variables, and establish an extended Kalman filter; In the iterative estimation process of the extended Kalman filter algorithm, the current SOC ampere-hour estimation value is calculated based on the final SOC estimation value at the previous moment, the ampere-hour correction coefficient is calculated based on the current SOC optimal estimation value, and then the current SOC optimal estimation value and the SOC ampere-hour estimation value are weighted according to the ampere-hour correction coefficient to calculate the current SOC final estimation value as the initial value of the next moment for iterative advancement; The extended Kalman filter algorithm iterative estimation process includes: Set the initial values ​​of state variables and state mean square errors in the state space equations; Time update, calculate the estimated value of the state variables and the estimated value of the state mean square error at the current moment; Calculate Kalman gain; Calculating the optimal estimated value of the state variable and the optimal estimated value of the state mean square error at the current moment, and extracting the optimal estimated value of the SOC at the current moment from the optimal estimated value of the state variable; Calculate the SOC ampere-hour method estimated value at the current moment based on the final estimated value of SOC at the previous moment; Calculating the current ampere-hour correction coefficient according to the current optimal estimated SOC value; The optimal estimated value of SOC at the current moment and the estimated value of SOC by ampere-hour method are weighted to calculate the final estimated value of SOC at the current moment; Determine whether the stop condition is met at the current moment, if not, return to the time update, calculate the estimated value of the state variable and the estimated value of the state mean square error at the current moment, and start the iterative advancement at the next moment; The ampere-hour correction coefficient is obtained by dynamically adjusting the optimal estimated value of SOC at the current moment during the estimation process at each moment using the following formula: Among them, k c ∈(0,1),k c,max k c The maximum value of the parameter, β is the slope coefficient, is the optimal estimated value of SOC at time k, SOC thred k c The SOC threshold, i.e., k c The SOC value corresponding to when the value reaches 1 / 2 of its maximum value.

2. The extended Kalman filter lithium battery SOC estimation method based on ampere-hour correction according to claim 1 is characterized in that: The SOC ampere-hour method estimated value is calculated based on the final estimated value of SOC at the previous moment, and the formula is: Among them, SOC k，ah is the estimated value of SOC by ampere-hour method at time k, SOC k-1,cor is the final estimated value of SOC at time k-1, η is the coulomb efficiency, C is the rated capacity of the battery, Δt is the estimated time interval, i o，k is the measured value of the battery output current at time k.

3. The method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction according to claim 2, characterized in that: The weighted calculation formula of the optimal estimated value of SOC at the current moment and the estimated value of SOC ampere-hour method is: in, is the optimal estimated value of SOC at time k, SOC k,cor is the final estimated value of SOC at time k, which will be brought into the extended Kalman filter estimation process at time k+1.

4. The method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction according to claim 3, characterized in that: When performing SOC estimation at the first moment, the final estimation value at the current moment is equal to the optimal estimation value at the current moment.

5. The method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction according to claim 1 or 4, characterized in that: The expression of the first-order Thevenin equivalent circuit model is: in o =in ocv -and o R0-u1 Among them, u ocv is the open circuit voltage, R0 is the battery ohmic internal resistance, R1 and C1 are the battery polarization resistance and capacitance, u o ,i o are the battery terminal voltage and current, u1 is the battery internal resistance terminal voltage, is the derivative of the voltage with respect to time, that is, the rate of change of the voltage.

6. The method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction according to claim 5, characterized in that: Perform DC charging and DC discharging experiments on the battery, and take the average value of the charging voltage changing with SOC and the discharging voltage changing with SOC as the open circuit voltage value u ocv , and then curve fitting is performed to obtain the OCV-SOC curve; The open circuit voltage u ocv Perform polynomial fitting with the corresponding SOC value to establish the fitting formula of the battery open circuit voltage and SOC function: you ocv (SOC)=a0+a1SOC+a2SOC 2 +a3SOC 3 +a4SOC 4 +a5SOC 5 +a6SOC 6 +a7SOC 7 +a8SOC 8 +a9SOC 9 Among them, a i is the polynomial fitting coefficient, i=0, 1, ..., 9, SOC is the state of charge value.

7. The method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction according to claim 6, characterized in that: According to the least squares theory, the first-order Thevenin equivalent circuit model is Laplace transformed to obtain: Among them, τ1=R1C1 is the time constant of the battery model, and s is the laplace operator; The test data of the charge and discharge experiment is obtained, and the forgetting factor is introduced in the process of recursive least squares method parameter identification to identify the battery ohmic internal resistance R0, battery polarization resistance R1, and battery polarization capacitance C1.

8. The method for estimating SOC of a lithium battery using an extended Kalman filter based on ampere-hour correction according to claim 7, characterized in that: The state space equation used in the Extended Kalman Filter algorithm is: f(x k ,u k )=A k x k +B k u k g(x k ,u k )=C k x k +D k u k Among them, x k =[SOC k ,u 1，k ] T , represents the state variable at time k, u k Represents the input variable of the system, and the coefficient A in the equation k , B k , C k , D k The expression is: Where C is the rated capacity of the battery, τ1=R1C1 is the time constant of the battery model, η is the coulomb efficiency, Δt is the estimated time interval, and u ocv (SOC) is the open circuit voltage, and R0 is the battery ohmic internal resistance.

Images