An accurate estimation method of state of charge of aviation lithium batteries based on improved Kalman filter algorithm

Abstract

The present invention relates to a method for accurately estimating the state of charge of a lithium battery for aviation based on an improved Kalman filtering algorithm, which obtains the current, open circuit voltage and terminal voltage of the battery; establishes an ARX model of the lithium battery, uses a recursive least squares method to solve the model coefficients, obtains a coefficient matrix, discretizes the lithium battery system to obtain a state space equation; uses a window estimation method to estimate the process noise covariance matrix Q k and the observation noise covariance matrix R k Dynamic adjustment is performed; real-time estimation of lithium battery SOC is performed according to the classic KF recursive process; this method can be used in nonlinear systems, has the advantages of high algorithm accuracy, strong robustness, and automatic correction of initial value errors, which is beneficial to improving the accuracy and speed of lithium battery SOC estimation.

Description

Technical Field

[0001] The invention belongs to the technical field of battery charge state estimation, and in particular relates to a method for accurately estimating the charge state of an aviation lithium battery based on an improved Kalman filter algorithm. Background Art

[0002] Lithium batteries are often used for energy storage in more-electric aircraft due to their outstanding advantages such as high cell voltage, high energy density, long cycle life, and fast charge and discharge capabilities. In actual applications, lithium batteries are usually combined with power sources such as supercapacitors, fuel cells or engines to form a hybrid energy system. In order to enable lithium batteries to better perform their performance in the energy system, improve their safety, and extend their service life, effective energy control and management are particularly important.

[0003] The state of charge (SOC) of lithium batteries is an important constraint in aircraft energy management, but the SOC of lithium batteries cannot be measured directly. It can only be obtained indirectly by obtaining the external characteristic parameters of the battery (such as current, terminal voltage, ambient temperature, etc.) and combining them with related functions. A good SOC estimation method needs to have the characteristics of high accuracy and fast response speed. The Kalman filter (KF) algorithm is a linear filtering algorithm that only needs to store the best estimate obtained at the previous sampling time, without recording a large amount of historical data. When the initial value is difficult to estimate accurately, the Kalman filter can gradually approach the true value through continuous filtering updates. The improved KF algorithm can be used for nonlinear systems. It has the advantages of high algorithm accuracy, strong robustness, and automatic correction of initial value errors, which is conducive to improving the accuracy and rapidity of lithium battery SOC estimation. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a lithium battery SOC accurate estimation method based on an improved Kalman filter algorithm. The method obtains a coefficient matrix based on the established autoregressive ergodic (ARX) model of the lithium battery, and uses a window estimation method to calculate the observed noise covariance R according to the residual between the actual value and the estimated value of the battery terminal voltage. k and process noise covariance Q k Iterative updates are performed, and then the state is estimated through the recursive process of the KF algorithm, thereby realizing online real-time estimation of SOC.

[0005] Lithium battery is a nonlinear system, and it is easily affected by factors such as ambient temperature and different working conditions during operation. Therefore, in order to ensure the accuracy of the model, the equivalent circuit model of lithium battery is established according to the ARX model, and the coefficient matrix of the lithium battery model is obtained by recursive least squares method.

[0006] In the traditional KF algorithm, the observation noise covariance R k and process noise covariance Qk It is fixed during the filtering process, and its value is mainly given by experience. There is no specific calculation formula, which can easily lead to large errors, thus affecting the accuracy of the KF algorithm. The improved adaptive Kalman filter (AKF) algorithm has a better effect on R k and Q k Dynamic adjustment is performed, and then the SOC is estimated according to the recursive process of the KF algorithm, thereby realizing online real-time and accurate estimation of the SOC.

[0007] Specifically, the present invention proposes a method for accurately estimating the SOC of a lithium battery based on an improved Kalman filter algorithm. The method includes the following steps.

[0008] Step S1. Obtain the current, open circuit voltage and terminal voltage of the battery; and establish an ARX model of the lithium battery, using the recursive least squares method to solve the model coefficients to obtain a coefficient matrix;

[0009] The observed data I(k) is the battery current, and U(k) is the battery open circuit voltage U oc With terminal voltage U b The difference between oc -U b ;

[0010] Solution process:

[0011] First, determine the initial value P(0) of the covariance matrix and the initial value θ(0) of the coefficient to be identified, and then sample to obtain the observed data U(k) and I(k), U(k) = U oc -U b , U oc Indicates the battery open circuit voltage, U b Represents the battery terminal voltage; I(k)=I b , I b Represents the battery current, together forming the observation vector

[0012]

[0013] Where k represents the kth sampling, m and n are the model orders;

[0014] Finally, the parameter estimates are calculated according to the recursive formula

[0015]

[0016] In the formula, θ(k) is the coefficient to be identified, θ=[a 1 a 2 …a m b 0 b 1 …b n ], a1 a 2 …a m b 0 b 1 …b n represents the voltage and current coefficients of the ARX model, and m and n represent the order of the model; is the observation vector, K(k) is the gain vector, and P(k) is the covariance matrix; the k-1 in the brackets represents the value of the variable at the previous moment.

[0017] Step S2. Discretize the lithium battery system to obtain the state space equation

[0018]

[0019] In the formula, x(k-1) represents the state variable of the system at time k-1, y(k-1) represents the observed variable of the system at time k-1, u(k-1) represents the controlled variable of the system at time k-1, and w k-1 represents the process noise at time k-1, v k-1 Representing the observation noise at time k-1, the specific form of the A, B, C, and D coefficient matrices is as follows:

[0020]

[0021] Among them U OC (k-1) represents the open circuit voltage at time k-1, which has a corresponding functional relationship with SOC, s(k-1) represents the state of charge of the lithium battery at time k-1, η represents the coulomb efficiency, T s represents the system sampling time, C N Indicates rated capacity, a 1 、a 2 Represents the two voltage coefficients of the ARX model, b 0 、b 1 Represents the two current coefficients of the ARX model.

[0022] Step S3: Use the window estimation method to estimate R k and Q k Make dynamic adjustments.

[0023] The specific form of the window method is as follows:

[0024]

[0025] Among them, l represents the window length, k represents the current sampling time, and e k Indicates the actual value of the battery terminal voltage U L,k Determine the residual U′ with the estimated value L,k , e k =U L,k -U′L,k .

[0026] According to the obtained covariance matrix G k , for Q k and R k Perform iterative updates:

[0027]

[0028] Among them, K k is the gain matrix, P k is the error covariance matrix, usually in the first d (d = l) sampling moments, Q k , R k The value remains unchanged.

[0029] Step S4: Real-time estimation of the lithium battery SOC is performed according to the classic KF recursive process.

[0030] The specific recursive process is as follows: Step 1: Set the initial value

[0031]

[0032]

[0033] x 0 represents the initial value of the state variable, E() represents the mathematical expectation of the content in the brackets, represents the estimated value of the initial value of the state variable, P 0 Represents the initial value of the covariance matrix.

[0034] The second step is state estimation.

[0035]

[0036] represents the estimated value of the state variable at time k-1, u(k-1) represents the control variable at time k-1, A represents the value of the state variable at time k estimated at time k-1, k-1 and B k-1 They represent the transfer matrix and input matrix at time k-1 respectively.

[0037] The third step is to estimate the error covariance matrix.

[0038]

[0039] A k-1 represents the transfer matrix at time k-1, represents the transpose of the state variable coefficient matrix at time k-1, P k-1 represents the covariance matrix at time k-1, Q k-1represents the process noise covariance matrix at time k-1, P k|k-1 represents the k-time covariance matrix calculated at time k-1.

[0040] Step 4: Calculate the Kalman gain

[0041]

[0042] K k represents the gain vector at time k, P k|k-1 represents the k-time covariance matrix calculated at time k-1. k represents the observation noise covariance at time k, C k represents the measurement matrix at time k, Represents the transpose of the measurement matrix at time k.

[0043] Step 5: Status Update

[0044]

[0045] y k represents the observed variable at time k, represents the value of the observed variable at time k estimated at time k-1, represents the estimated value of the state variable at time k-1, represents the estimated value of the state variable at time k, represents the value of the state variable at time k estimated at time k-1, u(k-1) represents the control variable at time k-1, K k represents the gain vector at time k, C k-1 represents the measurement matrix at time k-1, D k-1 Represents the output matrix at time k-1

[0046] Step 6: Update the error covariance matrix

[0047] P k =P k|k-1 -K k C k P k|k-1

[0048] P k represents the covariance matrix at time k, P k|k-1 represents the k-time covariance matrix calculated at time k-1, K k represents the gain vector at time k, C k represents the measurement matrix at k moments.

[0049] The above prediction and correction steps are continuously repeated to optimize the estimated SOC value of the lithium battery.

[0050] The advantages of the present invention are:

[0051] The present invention realizes a lithium battery SOC accurate estimation method based on an adaptive Kalman filter algorithm, realizes online real-time estimation of the lithium battery SOC, and contributes to efficient energy management of an airborne hybrid energy system.

[0052] The present invention adopts the moving window method to calculate the observation noise covariance R k and process noise covariance Q k Dynamic adjustment is performed to eliminate the influence of large errors caused by fixed covariance values ​​in the traditional Kalman algorithm, thereby improving the accuracy of SOC estimation.

[0053] Compared with the traditional KF algorithm and the extended KF algorithm, the AKF algorithm proposed in the present invention not only has a lower computational cost, but also ensures the sensitivity of the algorithm, and is more suitable for practical engineering.

[0054] The present invention can also be extended to the fields where lithium batteries are commonly used, such as electric vehicles, special equipment, satellites, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present invention. Moreover, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0056] Figure 1 Implementation mechanism of lithium-ion battery SOC estimation based on adaptive Kalman filter

[0057] Figure 2 ARX ​​model equivalent circuit diagram of lithium battery

[0058] Figure 3 This is the working principle diagram of the Kalman filter algorithm DETAILED DESCRIPTION

[0059] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Based on the embodiments provided in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0060] Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. In addition, it can also be understood that although the efforts made in this development process may be complicated and lengthy, for ordinary technicians in this field related to the content disclosed in this application, some changes in design, manufacturing or production based on the technical content disclosed in this application are just conventional technical means, and should not be understood as insufficient content disclosed in this application.

[0061] Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those of ordinary skill in the art that the embodiments described in this application may be combined with other embodiments without conflict.

[0062] Unless otherwise defined, the technical terms or scientific terms involved in this application should be understood by people with ordinary skills in the technical field to which this application belongs. The words "one", "a", "a kind of", "the" and the like involved in this application do not indicate a quantitative limitation and may represent the singular or plural. The terms "include", "comprise", "have" and any of their variations involved in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device comprising a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units inherent to these processes, methods, products or devices.

[0063] The implementation mechanism of the lithium-ion battery SOC estimation based on adaptive Kalman filtering proposed in the present invention is as follows: Figure 1 As shown. First, the ARX model of lithium battery is established. The model order is obtained by the genetic algorithm transformed from the AIC criterion, and then the model coefficient is obtained by the recursive least squares method. Then, according to the established battery model, the coefficient matrix is ​​obtained, and the window estimation method is used to calculate Q according to the residual between the actual value and the estimated value of the battery terminal voltage. k and R k Iterative updates are performed, and finally the state is estimated through the recursive process of the KF algorithm, thereby realizing online real-time estimation of SOC. The following is a detailed description:

[0064] The ARX model equivalent circuit of lithium battery is as follows Figure 2As shown in the figure, the model uses the input and output of the past moment to obtain the linear model of the current moment, and no longer needs to rely on the specific parameter values ​​of the equivalent circuit model. Based on the lithium battery ARX model, the AKF algorithm can obtain accurate model coefficients for SOC estimation.

[0065] Specifically, a method for accurately estimating the SOC of an aviation lithium battery is implemented in the following steps:

[0066] Step 1: Establish the ARX model of lithium battery and solve the coefficient matrix.

[0067] The ARX model expression is:

[0068] U(k)=-a 1 U(k-1)-a 2 U(k-2)+b 0 I(k)+b 1 I(k-1) (1)

[0069] Where a 1 、a 2 Represents the two voltage coefficients of the ARX model, b 0 、b 1 Represents the two current coefficients of the ARX model

[0070] In the formula, the observed data U(k) is the battery open circuit voltage U oc With terminal voltage U b The difference between oc -U b , the observed data I(k) is the battery current. The recursive least squares method is used to solve the model coefficients and obtain the coefficient matrix. Solution process:

[0071] First, determine the initial value of the covariance matrix P(0) and the initial value of the coefficient to be identified θ(0), then sample to obtain the observation data U(k) and I(k), which together form the observation vector

[0072]

[0073] In the formula, k represents the kth sampling, m and n are the model orders, and finally the parameter estimation value is calculated according to the recursive formula;

[0074]

[0075] In the formula, θ(k) is the coefficient to be identified, θ=[a 1 a 2 …a m b 0 b 1 …b n ], a 1a 2 …a m b 0 b 1 …b n Represents multiple voltage and current coefficients of the ARX model, and m and n represent the order of the model; is the observation vector, K(k) is the gain vector, and P(k) is the covariance matrix; the k-1 in the brackets represents the value of the variable at the previous moment.

[0076] Step 2: Discretize the lithium battery system to obtain the state space equation.

[0077] According to formula (1), the state and observation equations can be obtained as follows:

[0078]

[0079]

[0080] Where U L (k-1) represents the estimated value of the battery terminal voltage at time k-1, U OC (k-1) represents the open circuit voltage at time k-1, which has a corresponding functional relationship with SOC, U(k), U(k-1), and U(k-2) represent the output voltage of the ARX model at time k, time k-1, and time k-2, respectively, i(k), i(k-1), and i(k-2) represent the battery current of the ARX model at time k, time k-1, and time k-2, respectively, s(k) and s(k-1) represent the state of charge of the lithium battery at different sampling times, η represents the coulomb efficiency, T s represents the system sampling time, C N Indicates rated capacity, a 1 、a 2 represents the voltage coefficient of the ARX model, b 0 、b 1 Represents the current coefficient of the ARX model.

[0081] The discrete state equation is

[0082]

[0083] In formula (6), x(k-1) represents the state variable of the system at time k-1, y(k-1) represents the observed variable of the system at time k-1, u(k-1) represents the controlled variable of the system at time k-1, and w k-1 represents the process noise at time k-1, v k-1 represents the observation noise at time k-1, and its characteristics are as follows:

[0084]

[0085]

[0086] In formula (8) and (9), Q k is the process noise covariance matrix, R k is the observed noise covariance matrix, E[] represents the expectation of the contents in brackets, w z represents the process noise at time z, w k T represents the transpose of the process noise at time k, v z represents the observation noise at time z, represents the transpose of the observation noise at time k, z represents any sampling time, and k represents the current sampling time. The specific form of the A, B, C, and D coefficient matrices is as follows:

[0087]

[0088] Among them U OC (k-1) represents the open circuit voltage at time k-1, which has a corresponding functional relationship with SOC, s(k-1) represents the state of charge of the lithium battery at time k-1, η represents the coulomb efficiency, T s represents the system sampling time, C N Indicates rated capacity, a 1 、a 2 represents the voltage coefficient of the ARX model, b 0 、b 1 Represents the current coefficient of the ARX model.

[0089] Step 3: Use the window estimation method to estimate R k and Q k Make dynamic adjustments.

[0090] The windowing estimation method uses the residuals between multiple actual observations and estimated values ​​before the next moment to estimate the covariance matrix in real time, thereby reducing the model error. The specific form of the windowing method is as follows:

[0091]

[0092] Among them, l represents the window length, k represents the current sampling time, and e k Indicates the actual value of the battery terminal voltage U L,k Determine the residual U′ with the estimated value L,k , e k =U L,k -U′ L,k .

[0093] According to the obtained covariance matrix G k , for the process noise covariance matrix Q k and the observation noise covariance matrix Rk Perform iterative updates:

[0094]

[0095] Among them, K k is the gain matrix, P k is the error covariance matrix, usually in the first d (d = l) sampling moments, Q k , R k The value remains unchanged.

[0096] Step 4: Figure 3 The classic KF recursive process shown is used to estimate the SOC of lithium batteries in real time.

[0097] The specific recursive process is as follows: Step 1: Set the initial value

[0098]

[0099]

[0100] x 0 represents the initial value of the state variable, E() represents the mathematical expectation of the content in the brackets, represents the estimated value of the initial value of the state variable, P 0 Represents the initial value of the covariance matrix.

[0101] The second step is state estimation.

[0102]

[0103] represents the estimated value of the state variable at time k-1, u(k-1) represents the control variable at time k-1, A represents the value of the state variable at time k estimated at time k-1, k-1 and B k-1 They represent the transfer matrix and input matrix at time k-1 respectively.

[0104] The third step is to estimate the error covariance matrix.

[0105]

[0106] A k-1 represents the transfer matrix at time k-1, represents the transpose of the state variable coefficient matrix at time k-1, P k-1 represents the covariance matrix at time k-1, Q k -1 represents the process noise covariance matrix at time k-1, P k|k-1 represents the k-time covariance matrix calculated at time k-1.

[0107] Step 4: Calculate the Kalman gain

[0108]

[0109] K k represents the gain vector at time k, P k|k-1 represents the k-time covariance matrix calculated at time k-1. k represents the observation noise covariance at time k, C k represents the measurement matrix at time k, Represents the transpose of the measurement matrix at time k.

[0110] Step 5: Status Update

[0111]

[0112] y k represents the observed variable at time k, represents the value of the observed variable at time k estimated at time k-1, represents the estimated value of the state variable at time k-1, represents the estimated value of the state variable at time k,

[0113] represents the value of the state variable at time k estimated at time k-1, u(k-1) represents the control variable at time k-1, K k represents the gain vector at time k, C k-1 represents the measurement matrix at time k-1, D k-1 Represents the output matrix at time k-1

[0114] Step 6: Update the error covariance matrix

[0115] P k =P k|k-1 -K k C k P k|k-1 (16)

[0116] P k represents the covariance matrix at time k, P k|k-1 represents the k-time covariance matrix calculated at time k-1, K k represents the gain vector at time k, C k represents the measurement matrix at k moments.

[0117] The above prediction and correction steps are continuously repeated to optimize the estimated SOC value of the lithium battery.

[0118] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for accurately estimating the state of charge of an aviation lithium battery based on an improved Kalman filter algorithm, characterized in that: The above method comprises the following steps: Step S1. Obtain the current, open circuit voltage and terminal voltage of the battery; and establish an ARX model of the lithium battery, using the recursive least squares method to solve the model coefficients to obtain a coefficient matrix; The observed data I(k) is the battery current, and U(k) is the battery open circuit voltage U oc With terminal voltage U b The difference between oc -U b ; In step S1, the specific solution process of obtaining the coefficient matrix is: First, determine the initial value P(0) of the covariance matrix and the initial value θ(0) of the coefficient to be identified, and then sample to obtain the observed data U(k) and I(k), U(k) = U oc -U b , U oc Indicates the battery open circuit voltage, U b Represents the battery terminal voltage; I(k)=I b , I b Represents the battery current, together forming the observation vector In the formula, k represents the kth sampling, m and n are the model orders, and finally the parameter estimation value is calculated according to the recursive formula; In the formula, θ(k) is the coefficient to be identified, θ=[a1 a2…a m b0 b1…b n ], a1 a2…a m b0 b1…b n Represents multiple voltage and current coefficients of the ARX model, and m and n represent the order of the model; is the observation vector, K(k) is the gain vector, and P(k) is the covariance matrix; the k-1 in the brackets represents the value of the variable at the previous moment; Step S2. Discretize the lithium battery system to obtain a state space equation; The state space equation of the lithium battery system is: In the formula, x(k-1) represents the state variable of the system at time k-1, y(k-1) represents the observed variable of the system at time k-1, u(k-1) represents the controlled variable of the system at time k-1, and w k-1 represents the process noise at time k-1, v k-1 Representing the observation noise at time k-1, the specific form of the A, B, C, and D coefficient matrices is as follows: Among them U OC (k-1) represents the open circuit voltage at time k-1, which has a corresponding functional relationship with SOC, s(k-1) represents the state of charge of the lithium battery at time k-1, η represents the coulomb efficiency, T s represents the system sampling time, C N represents the rated capacity, a1 and a2 represent the two voltage coefficients of the ARX model, and b0 and b1 represent the two current coefficients of the ARX model; Step S3: Use the window estimation method to estimate the process noise covariance matrix Q k and the observation noise covariance matrix R k Make dynamic adjustments; The specific form of the window method is as follows: Among them, l represents the window length, k represents the current sampling time, and e k Indicates the actual value of the battery terminal voltage U L,k Determine the residual U′ with the estimated value L,k , e k =U L,k -U′ L,k ; According to the obtained covariance matrix G k , for the process noise covariance matrix Q k and the observation noise covariance matrix R k Perform iterative updates: Among them, K k is the gain matrix, P k is the error covariance matrix, usually in the first d (d = l) sampling moments, Q k , R k The value remains unchanged; Step S4. Real-time estimation of lithium battery SOC according to the classic KF recursive process; The specific recursive process of real-time estimation of lithium battery SOC is as follows: Step 1: Set the initial value x0 represents the initial value of the state variable, E() represents the mathematical expectation of the content in the brackets, represents the estimated value of the initial value of the state variable, and P0 represents the initial value of the covariance matrix; The second step is state estimation. represents the estimated value of the state variable at time k-1, u(k-1) represents the control variable at time k-1, A represents the value of the state variable at time k estimated at time k-1, k-1 and B k-1 They represent the transfer matrix and input matrix at time k-1 respectively; The third step is to estimate the error covariance matrix. A k-1 represents the transfer matrix at time k-1, represents the transpose of the state variable coefficient matrix at time k-1, P k-1 represents the covariance matrix at time k-1, Q k-1 represents the process noise covariance matrix at time k-1, P k|k-1 represents the k-time covariance matrix calculated at time k-1; Step 4: Calculate the Kalman gain K k represents the gain vector at time k, P k|k-1 represents the k-time covariance matrix calculated at time k-1; R k represents the observation noise covariance at time k, C k represents the measurement matrix at time k, represents the transpose of the measurement matrix at time k; Step 5: Status Update y k represents the observed variable at time k, represents the value of the observed variable at time k estimated at time k-1, represents the estimated value of the state variable at time k-1, represents the estimated value of the state variable at time k, represents the value of the state variable at time k estimated at time k-1, u(k-1) represents the control variable at time k-1, K k represents the gain vector at time k, C k-1 represents the measurement matrix at time k-1, D k-1 Represents the output matrix at time k-1; Step 6: Update the error covariance matrix P k =P k|k-1 -K k C k P k|k-1 P k represents the covariance matrix at time k, P k|k-1 represents the k-time covariance matrix calculated at time k-1, K k represents the gain vector at time k, C k represents the measurement matrix at k moments; The above steps of setting the initial value to updating the error covariance matrix are continuously cycled, thereby optimizing the estimated value of the lithium battery SOC.

2. The method for accurately estimating the state of charge of an aviation lithium battery based on an improved Kalman filter algorithm according to claim 1, characterized in that: Since the observation noise covariance R in the traditional Kalman algorithm k and process noise covariance Q k After filtering, it remains constant.

3. The method for accurately estimating the state of charge of an aviation lithium battery based on an improved Kalman filter algorithm according to claim 1, characterized in that: An ARX model of lithium battery is established, and the model coefficients are solved using the recursive least squares coefficient method. The recursive least squares algorithm only needs one set of observation data to estimate the parameters once, and there is no need to store all the data. The more observations, the more corrections are made, and the SOC of the lithium battery can be estimated in real time.

Images