Method for lithium-ion battery state-of-charge and state-of-health joint estimation

Abstract

The application discloses a lithium ion battery charge state and health state combined estimation method, and is specifically implemented according to the following steps: step 1: establishing a fractional order equivalent circuit model of a lithium ion battery, and then establishing a mathematical model of model parameters; step 2: adopting a double extended Kalman filtering algorithm to estimate the fractional order equivalent circuit model parameters and the SOC and the maximum available capacity of the battery under a current state on line; and step 3: adopting a time sequence weighting method to calculate the average maximum available capacity and SOH of the battery, and combining the estimated ohmic internal resistance with a multiple relationship of the ohmic internal resistance of a new battery to comprehensively judge the health state of the battery. The method can accurately estimate the health state of the lithium ion battery.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of energy storage, and particularly relates to a lithium ion battery state of charge and state of health joint estimation method. BACKGROUND

[0002] Lithium ion batteries have the advantages of high energy density, long cycle life and small self-discharge, and are increasingly widely used in the field of energy storage. Unlike other forms of energy, the residual capacity of a large number of single batteries in an energy storage system cannot be directly measured, so overcharging and overdischarging of the batteries occur from time to time, and the resulting safety accidents of the energy storage system have attracted enough attention in the industry.

[0003] Generally, state of charge (SOC) and state of health (SOH) estimation is the core of the battery management system, which is the basis for ensuring battery safety and improving reliability. SOC represents the ratio of the remaining capacity of the battery to its maximum actual available capacity, and SOH represents the ratio of the maximum actual available capacity of the battery to its rated capacity, which is a measure of the degree of battery aging. Many documents use various forms of least squares, particle filters, intelligent algorithms, data-driven algorithms, and Kalman filters, extended Kalman filters, etc. to separately estimate or jointly estimate the equivalent circuit parameters, SOC and SOH of lithium ion batteries. However, for SOH, due to its large time scale and slow battery aging speed, it is difficult to accurately estimate SOH using a single estimate value in a short period of time, resulting in the problem of large SOH estimation error in existing estimation methods. Therefore, the present application proposes a lithium ion battery state of health estimation method based on double extended Kalman filtering and time series weighting, which can accurately estimate SOH in a long time scale, consistent with the evolution process of lithium ion battery health decline, and lay a good foundation for accurately judging the health status of the battery. SUMMARY

[0004] The purpose of the present application is to provide a lithium ion battery state of charge and state of health joint estimation method that can accurately estimate the state of health of a lithium ion battery.

[0005] The technical solution adopted by the present application is a lithium ion battery state of charge and state of health joint estimation method, which is implemented according to the following steps:

[0006] Step 1: Establish a fractional order equivalent circuit model of the lithium ion battery, and then establish a mathematical model of the model parameters;

[0007] Step 2: Use a double extended Kalman filter algorithm to estimate the fractional order equivalent circuit model parameters and the SOC and maximum available capacity of the battery under the current state online;

[0008] Step 3: Calculate the average battery maximum available capacity and SOH using the time series weighting method, and combine the estimated ohmic resistance with the multiple relationship of the new battery ohmic resistance to comprehensively judge the health status of the battery.

[0009] The application is also characterized in that,

[0010] Step 1 is specifically:

[0011] Step 1.1, establish a fractional order equivalent circuit model of the lithium ion battery, including a power supply (1), an ohmic resistance R0, an electrochemical polarization resistance R1, a concentration polarization resistance R2 and a Warburg element C W , which are connected in sequence, further comprising a polarization capacitance CPE1 connected in parallel with the electrochemical polarization resistance R1 and a concentration polarization capacitance CPE2 connected in parallel with the concentration polarization resistance R2.

[0012] Among them, CPE1 and CPE2 are constant phase elements, and their impedance is shown in formula (1):

[0013]

[0014] In the formula, α and β are the fractional orders of the constant phase elements CPE1 and CPE2 respectively, and s is the Laplace operator.

[0015] Warburg element C W represents the positive and negative electrode solid phase diffusion reaction, and its impedance is shown in formula (2):

[0016]

[0017] In the formula, γ represents the fractional order of the Warburg element C W ;

[0018] Step 1.2, establish a mathematical model of the parameters of the fractional order equivalent circuit model of the lithium ion battery, which is specifically as follows:

[0019]

[0020] In the formula, η is the charge and discharge efficiency; Q n is the maximum available capacity of the battery; T s is the sampling time; α and β are the fractional orders of the constant phase elements CPE1 and CPE2 respectively; γ is the fractional order of the element C W ; I L (t) represents the battery discharge current; U1(t) represents the voltage of the constant phase element CPE1; U2(t) represents the voltage of the constant phase element CPE2; U3(t) represents the voltage of the element C W ; SOC(t) represents the state of charge of the battery; U o (t) represents the battery terminal voltage at the kth moment.

[0021] Step 1.3: After discretizing the fractional-order calculus equations of Kirchhoff's voltage law and the fractional-order equivalent circuit model of the lithium-ion battery, we obtain equation (4), as shown below:

[0022]

[0023] In the formula, I L (k) represents the battery discharge current sampled at time k; U1(k) and U1(k+1) represent the voltages of the constant-phase element CPE1 sampled at times k and k+1, respectively; U2(k) and U2(k+1) represent the voltages of the constant-phase element CPE2 sampled at times k and k+1, respectively; U3(k) and U3(k+1) represent the voltages of the constant-phase element CPE2 sampled at times k and k+1, respectively. W The voltage; SOC(k) and SOC(k+1) represent the battery state of charge at times k and k+1, respectively; U o (k) represents the battery terminal voltage at time k;

[0024] Step 1.4, Define x k =[SOC(k), U1(k), U2(k), U3(k), R0(k), 1 / Q n (k)] T θ k =[1 / R 1,k 1 / CPE 1,k α k , 1 / R 2,k 1 / CPE 2,k ,β k , 1 / C W,k γ k ], u k = I L (k), y k = U o (k); Rearrange formula (4) into formula (5):

[0025]

[0026] in:

[0027]

[0028]

[0029] In the formula, x i,k x represents k The i-th element, θ i,k Represents θ k The i-th element.

[0030] Step 2 is as follows:

[0031] Step 2.1, the first extended Kalman filter EKF1 is used to estimate x k parameters, and the second extended Kalman filter EKF2 is used to estimate θ k parameters; EKF2 estimates θ k parameters, and θ k estimated by EKF2 is brought into EKF1 to estimate SOC(k), ohmic internal resistance R0(k) and maximum available capacity Q n (k) of the battery k parameters;

[0032] Step 2.2, according to x k parameters estimated by EKF1, and in combination with the measured battery discharge current I L (k) and terminal voltage U o (k), SOC(k), ohmic internal resistance R0(k) and maximum available capacity Q n (k) of the battery in the current state are estimated n (k), and the discharge current and discharge rate of the battery are recorded, and the amount of time taken by the battery under the same discharge current and discharge rate is measured; during the discharge process, if the discharge current and discharge rate change, the time corresponding to the different discharge current and discharge rate is recorded.

[0033] In step 2.2, EKF2 and EKF1 need to be repeated once for each estimation, and it is necessary to judge whether the discharge current and discharge rate of the battery change during the period, if the discharge current and discharge rate do not change, the corresponding discharge time is continuously recorded; if the discharge current and discharge rate change, the time recording of the last time is ended, and the time recording corresponding to the new discharge current and discharge rate is started; the estimation process and time recording process of EKF2 and EKF1 are cycled in this way until the battery power is completely consumed or the discharge process is ended; after EKF1 is repeated N times, a series of ohmic internal resistances are obtained, which are respectively represented as R0(1), R0(2),..., R0(N), in addition, a batch of data are obtained, which are respectively (Q n (1), t1, SOC(1)), (Q n (2), t2, SOC(2)), (Q n (3), t3, SOC(3)),..., (Q n (N), t N , SOC(N)).

[0034] Step 3 is specifically:

[0035] Step 3.1, using the N sets of data obtained by EKF1 in Step 2.2 (Q n (1), t1, SOC(1)), (Q n (2), t2, SOC(2)), (Q n (3), t3, SOC(3)),..., (Q n (N), t N , SOC(N)), the average battery maximum available capacity Q0 is calculated using the time series weighting method represented by formula (7), Q0 is as follows:

[0036]

[0037] In the formula, Q n (1), Q n (2),..., Q n (N) are the maximum available capacities of the battery at each discharge current and discharge rate estimated by EKF1, t n represents the time experienced by the battery at each corresponding discharge current and discharge rate, and SOC(N) is the SOC at the end of the discharge process.

[0038] Step 3.2, SOH is calculated using formula (8), as follows:

[0039]

[0040] In the formula, Q rate is the rated capacity of the battery.

[0041] Step 3.3, the SOH of Step 3.2 represents the health state of a discharge depth, and the SOH is analyzed: if there are w health states of the same or different discharge depths in a long time range, the average of the w health states is taken as the global health state SOH whole in the long time range, as shown in formula (9):

[0042]

[0043] In the formula, SOH1, SOH2,..., SOH w represent the health states of the same or different discharge depths in a long time range, respectively.

[0044] At the same time, a series of ohmic internal resistances estimated by EKF1 in step 2.2 are represented as R0(1), R0(2),..., R0(N) respectively for auxiliary evaluation. Since it is a discharge process, the aging degree of the battery gradually deepens over time, so the ohmic internal resistance gradually increases, and therefore R0(N) is usually the largest. R0(N) is compared with the ohmic internal resistance of a new battery. If R0(N) is not less than m times (for example, m = 2) of the ohmic internal resistance of the new battery, it is considered that the battery has aged and can be retired; if R0(N) is less than m times (for example, m = 2) of the ohmic internal resistance of the new battery, the degree of aging of the battery needs to be combined with the global state of health SOH whole to determine whether SOH whole ≥ 50%, the battery can be used in echelon although it has aged; if SOH whole < 50%, it is considered that the battery has aged.

[0045] The beneficial effects of the present application are: the method of the present application uses double extended Kalman filtering method to online estimate the fractional order equivalent circuit model parameters, SOC and the maximum available capacity of the lithium ion battery under the current state. For the maximum available capacity of the battery under the current state and the corresponding constant current discharge time estimated by the double extended Kalman filtering method, the time series weighting method is used to calculate the average maximum available capacity of the battery in a long time scale regardless of different discharge depths and different discharge rates, so as to more accurately estimate SOH, and the change of ohmic internal resistance can improve the accuracy of SOH estimation of the battery management system, and provide theoretical support for health state alarm and timely replacement of the battery. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 is the fractional order equivalent circuit model of the lithium ion battery used in the present application.

[0047] Figure 2 is the flow chart of the SOH estimation method of the present application.

[0048] Figure 3 is the capacity comparison graph of the four single batteries in the present application.

[0049] Figure 4 is the input current and terminal voltage measurement curve under the dynamic stress test (DST) working condition in the present application.

[0050] Figure 5 is the comparison of the battery terminal voltage prediction results of the new battery under the DST working condition in the present application.

[0051] Figure 6 is the comparison of the terminal voltage prediction error of the new battery under the DST working condition in the present application.

[0052] Figure 7This is a comparison of the SOC estimation results of the new battery of this invention under DST conditions;

[0053] Figure 8 This invention compares the SOC error of the new battery under DST conditions.

[0054] Figure 9 This is a comparison of the voltage prediction results of the aged No. 1 battery under DST conditions according to the present invention;

[0055] Figure 10 This invention compares the voltage error of aged No. 1 batteries under DST conditions.

[0056] Figure 11 This is a comparison of the SOC estimation results of the aged No. 1 battery under DST conditions according to the present invention;

[0057] Figure 12 This invention compares the SOC error of aged No. 1 batteries under DST conditions.

[0058] Figure 13 This invention compares the voltage prediction results of aged No. 3 batteries under DST conditions.

[0059] Figure 14 This invention compares the voltage error of aged No. 3 batteries under DST conditions.

[0060] Figure 15 This invention compares the SOC estimation results of aged No. 3 batteries under DST conditions.

[0061] Figure 16 This invention compares the SOC error of aged No. 3 batteries under DST conditions.

[0062] Figure 17 This is the relative error in the prediction of the terminal voltage of different batteries in this invention.

[0063] Figure 18 These are the ohmic internal resistance curves of different batteries in this invention.

[0064] Figure 19 This is a histogram comparing the state of oxygen (SOH) of batteries with different discharge depths and different ages according to the present invention.

[0065] In the diagram, 1. Power supply. Detailed Implementation

[0066] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0067] This invention provides a method for jointly estimating the state of charge and state of health of a lithium-ion battery, such as... Figures 1-2 As shown, please follow these steps:

[0068] Step 1: establish a fractional order equivalent circuit model of a lithium ion battery, and then establish a mathematical model of the model parameters;

[0069] Step 1 is specifically:

[0070] Step 1.1, establish a fractional order equivalent circuit model of a lithium ion battery:

[0071] Figure 1 is the fractional order equivalent circuit model of the lithium ion battery used by the application, and the model structure includes a power supply (1), an ohmic resistor R0, an electrochemical polarization resistor R1, a concentration polarization resistor R2 and a Warburg element C connected in sequence W , and further includes a polarization capacitor CPE1 connected in parallel with the electrochemical polarization resistor R1 and a concentration polarization capacitor CPE2 connected in parallel with the concentration polarization resistor R2;

[0072] Figure 1 , wherein U o is the voltage across the model, I L is the current flowing into the circuit, R0 represents the ohmic resistance, which is represented as the intersection point of the impedance spectrum curve and the real axis; R1, CPE1, R2 and CPE2 respectively represent the electrochemical polarization resistance and the polarization capacitor, the concentration polarization resistance and the concentration polarization capacitor, U1 and U2 are respectively the voltages of the constant phase elements CPE1 and CPE2, and the above physical quantities reflect the charge transfer reaction and the double layer effect; U3 is the voltage across it; U ocv represents the open circuit voltage of the battery.

[0073] , wherein CPE1 and CPE2 are constant phase elements, and their impedance is shown in formula (1):

[0074]

[0075] In the formula, α and β are respectively the fractional orders of the constant phase elements CPE1 and CPE2, and s is the Laplace operator.

[0076] The Warburg element C W represents the positive and negative electrode solid phase diffusion reaction, and its impedance is shown in formula (2):

[0077]

[0078] In the formula, γ represents the fractional order of the Warburg element C W .

[0079] Step 1.2, establish a mathematical model of the parameters of the fractional order equivalent circuit model of the lithium ion battery, specifically as follows:

[0080]

[0081] In the formula, η is the charge and discharge efficiency; Qn is the maximum available capacity of the battery; T s is the sampling time; a, b are the fractional orders of the constant phase elements CPE1 and CPE2, respectively; g is the fractional order of the element C W ; I L (t) represents the battery discharge current; U1(t) represents the voltage of the constant phase element CPE1; U2(t) represents the voltage of the constant phase element CPE2; U3(t) represents the voltage of the element C W ; SOC(t) represents the state of charge of the battery; U o (k) represents the battery terminal voltage at the kth moment;

[0082] Step 1.3, according to the Kirchhoff voltage law and the fractional calculus equation of the fractional order equivalent circuit model of the lithium ion battery, the discretization is as follows: formula (4) is obtained, as shown below:

[0083]

[0084] In the formula, I L (k) represents the battery discharge current sampled at the kth moment; U1(k), U1(k+1) represent the voltage of the constant phase element CPE1 sampled at the kth and k+1th moments, respectively; U2(k), U2(k+1) represent the voltage of the constant phase element CPE2 sampled at the kth and k+1th moments, respectively; U3(k), U3(k+1) represent the voltage of the element C W sampled at the kth and k+1th moments, respectively; SOC(k), SOC(k+1) represent the state of charge of the battery at the kth and k+1th moments, respectively; U o (k) represents the battery terminal voltage at the kth moment;

[0085] In order to accurately describe the behavior of the battery under different working conditions, the ohmic resistance R0 and the maximum available capacity Q n of the battery under the current state, the parameters R1, CPE1, R2, CPE2, C W of the fractional order model, and the fractional orders a, b, g will be identified online.

[0086] Step 1.4, define x k = [SOC(k), U1(k), U2(k), U3(k), R0(k), 1 / Q n (k)] T , 0 k = [1 / R 1,k , 1 / CPE 1,k , a k , 1 / R 2,k , 1 / CPE 2,k , b k , 1 / CW,k , γ k ], u k = I L (k), y k = U o (k); formula (4) is arranged into formula (5):

[0087]

[0088] Wherein:

[0089]

[0090]

[0091] In the formula, x i,k represents the i-th element of x k , θ i,k represents the i-th element of θ k .

[0092] Step 2: using double extended Kalman filter algorithm (DEKF) to estimate the fractional order equivalent circuit model parameters and the current state of SOC, battery maximum available capacity online;

[0093] Step 2.1: in the double extended Kalman filter algorithm, the first extended Kalman filter EKF1 is used to estimate x k parameters, and the second extended Kalman filter EKF2 is used to estimate θ k parameters, therefore, the DEKF algorithm can estimate SOC and SOH simultaneously, and can also update the battery fractional order model parameters synchronously; EKF2 estimates θ k parameters, and the θ k parameters estimated by EKF2 are brought into EKF1 to estimate x n parameters of SOC(k), ohmic resistance R0(k) and battery maximum available capacity Q k (k);

[0094] Step 2.2: according to the x k parameters estimated by EKF1, and combined with the measured battery discharge current I L (k) and terminal voltage U o (k), the current state of SOC(k), ohmic resistance R0(k) and battery maximum available capacity Q n (k) are estimated, and the above-mentioned SOC(k), R0(k) and Q n (k) are recorded.(k), while recording the discharge current and discharge rate of the battery, measuring the amount of time the battery takes at the same discharge current and discharge rate; if the discharge current and discharge rate change during the discharge process, the time corresponding to the different discharge current and discharge rate is counted;

[0095] wherein EKF2 and EKF1 need to repeat the recording period to determine whether the discharge current and discharge rate of the battery change, if the discharge current and discharge rate do not change, the corresponding discharge time is continuously recorded; if the discharge current and discharge rate change, the time recording of the last time is ended, and the time recording corresponding to the new discharge current and discharge rate is started; the estimation process and time recording process of EKF2 and EKF1 are cycled until the battery is completely discharged or the discharge process is ended; EKF1 is repeated N times to obtain a series of ohmic internal resistance, which are respectively represented as R0(1), R0(2), …, R0(N), in addition, a batch of data can be obtained as (Q n (1), t1, SOC(1)), (Q n (2), t2, SOC(2)), (Q n (3), t3, SOC(3)), …, (Q n (N), t N , SOC(N)).

[0096] Step 3: The average battery maximum available capacity and SOH are calculated by using the time series weighting method, and the health status of the battery is comprehensively judged in combination with the estimated ohmic internal resistance and the multiple relationship of the new battery ohmic internal resistance.

[0097] Step 3 is specifically:

[0098] Step 3.1, using a batch of N groups of data (Q n (1), t1, SOC(1)), (Q n (2), t2, SOC(2)), (Q n (3), t3, SOC(3)), …, (Q n (N), t N , SOC(N)) obtained by EKF1 in step 2.2, the average battery maximum available capacity Q0 is calculated by using the time series weighting method represented by formula (7), Q0 is specifically as follows:

[0099]

[0100] In the formula, Q n (1), Q n (2), … Q n (N) are the maximum available capacities of the battery at each discharge current and discharge rate estimated by EKF1, tn represents the time experienced by the battery at each corresponding discharge current and discharge rate, SOC(N) is the SOC at the end of the discharge process; (1-SOC(N)) aims to solve the problem of inaccurate estimation of the maximum available capacity of the battery at variable discharge depth, Q n (1), Q n (2),... Q n (N) solves the problem of inaccurate estimation of the maximum available capacity of the battery at different discharge rates in a single discharge process.

[0101] Step 3.2, calculate SOH using formula (8), as follows:

[0102]

[0103] In the formula, Q rate is the rated capacity of the battery.

[0104] Step 3.3, the SOH of step 3.2 represents the health state of a discharge depth, analyze SOH: if there are w health states of the same or different discharge depths in a long time range (e.g. one week), then take the average of the w health states as the global health state SOH whole , as shown in formula (9):

[0105]

[0106] In the formula, SOH1, SOH2,..., SOH w represent the health states of the same or different discharge depths in a long time range (e.g. one week), respectively.

[0107] At the same time, a series of ohmic internal resistances estimated by EKF1 in step 2.2 are represented as R0(1), R0(2),..., R0(N) for auxiliary evaluation. Since it is a discharge process, the degree of aging of the battery gradually deepens over time, so the ohmic internal resistance gradually increases. Therefore, under normal circumstances, R0(N) is the largest. Compare R0(N) with the ohmic internal resistance of a new battery. If R0(N) is not less than m times (e.g. m = 2) the ohmic internal resistance of a new battery, it is considered that the battery has aged and can be retired; if R0(N) is less than m times (e.g. m = 2) the ohmic internal resistance of a new battery, the degree of aging of the battery needs to be determined in combination with the global health state SOH whole . If SOH whole ≥ 50%, although it has aged, it can be used in echelon; if SOH whole < 50%, it is considered that the battery has aged and can be retired.

[0108] The average SOH value is calculated using a time-series weighted method, and the health status of the battery is comprehensively judged by combining the estimated ohmic internal resistance with the multiple relationship between the ohmic internal resistance of the new battery.

[0109] Experimental verification:

[0110] Taking a single lithium iron phosphate battery as an example, experimental verification was conducted. The basic parameters of a single lithium-ion power battery are shown in Table 1.

[0111] Table 1 Parameters of a single lithium iron phosphate battery

[0112]

[0113] Figure 3 The capacity of the four batteries was measured by completely discharging them. Figure 3 As can be seen, the top battery is the new battery with the rated capacity. The other three batteries are batteries with different degrees of aging, among which the aged battery No. 3 is the most aged, with only half the capacity of the new battery. Since the aging levels of aged batteries No. 1 and No. 2 are similar, dynamic stress testing (DST) is performed using the new battery, aged battery No. 1, and aged battery No. 3.

[0114] Figure 4 The battery discharge current and terminal voltage curves were obtained in the DST experiment at an ambient temperature of 25℃, with a sampling time of 10s.

[0115] Figures 5-8 The estimated terminal voltage and SOC of the new battery under DST conditions are as follows: Figure 6 and Figure 8 They are respectively Figure 5 and Figure 7 The relative error values ​​show that the errors in the predicted terminal voltage and the SOC estimation are both very small.

[0116] Figures 9-12 The waveforms are related to the aging of battery No. 1, among which Figure 9 This is a comparison chart of the estimated and actual battery terminal voltages. Figure 11 This is a comparison chart of the estimated SOC and the actual SOC. Figure 10 and Figure 12 They are respectively Figure 9 and Figure 11 Relative error value, from Figure 10 and Figure 12 It can be seen that the predicted terminal voltage error and SOC error of the No. 1 aging battery are both very small.

[0117] Figures 13-16The comparative curves of the relative error of the terminal voltage of each battery under different aging degrees. With the increase of aging degree, the terminal voltage error gradually increases, but the maximum estimation error is only within 0.3%, indicating that the estimation effect of DEKF algorithm under different aging degrees is good.

[0118] From Figure 17 It can be seen that with the deepening of the aging degree of the battery, the ohmic resistance of the battery increases. But in the early stage of dynamic working condition, the internal resistance of the battery is basically the same, and with the decrease of the state of charge of the battery, the change of the internal resistance of the battery becomes larger.

[0119] From Figure 18 It can be seen that with the deepening of the aging degree of the battery, the ohmic resistance of the battery increases. The final ohmic resistance of the battery of aging 1 is about twice that of the new battery, and the final ohmic resistance of the battery of aging 3 is more than 4 times that of the new battery, indicating that the batteries of aging 1 and aging 3 have been aged and have poor health status and should be retired.

[0120] Figure 19 SOH comparison of batteries with different aging degrees under different discharge depths. Taking the SOH estimation value of the new battery as an example, the blue column is the reference value, and the SOH estimation value under the condition of 100% discharge depth, i.e. the result of full discharge of the battery, is used as a reference. The SOH of the battery under different discharge depths is compared with that under 100% discharge depth, and the global health SOH whole compared with the reference value, the error is not more than 2%. The aging degree of aging 1 battery and aging 2 battery is not much different, through comparison, the error of SOH whole of the two is not more than 1%, so it can be explained that the aging degree of aging 1 battery and aging 2 battery is not much different. Aging 3 battery is the battery with the deepest aging degree, through comparison of the estimation value of discharge depth 100% and global health SOH whole , it can be known that the error is not more than 1%, which verifies the effectiveness of the algorithm.

Claims

1. A method for combined state-of-charge and state-of-health estimation of a lithium-ion battery, characterized in that, The following steps are implemented in detail: Step 1: Establish a fractional order equivalent circuit model of the lithium ion battery, and then establish a mathematical model of the model parameters; Step 2: Estimate the fractional order equivalent circuit model parameters and the current state of SOC and the maximum available capacity of the battery online by using the double extended Kalman filter algorithm; Step 2 is specifically as follows: Step 2.

1. The first extended Kalman filter EKF1 is used in the dual extended Kalman filter algorithm to estimate x k parameters, the second extended Kalman filter EKF2 estimates simultaneously θ k parameters; EKF2 estimates θ k parameters, the EKF2 estimated θ k parameters are brought into EKF1 to estimate SOC k ), the ohmic internal resistance R 0 k ), and the maximum available capacity of the battery Q n k x k parameters;​​​ Step 2.2, based on the estimated value from EKF1 x k Parameters, and combined with the measured battery discharge current I L ( k and terminal voltage U o ( k Estimate the current state. SOC ( k Ohmic internal resistance R 0( k ) and maximum usable battery capacity Q n ( k Record the above. SOC ( k ), R 0( k )and Q n ( k Simultaneously, the battery's discharge current and discharge rate are recorded, and the time taken by the battery under the same discharge current and discharge rate is measured. In the discharge process, if the discharge current and the discharge rate change, the time corresponding to different discharge currents and discharge rates is taken into account; Step 3: Calculate the average maximum available capacity and SOH of the battery by using the time series weighting method, and combine the estimated ohmic resistance with the multiple relationship of the new battery ohmic resistance to comprehensively judge the health state of the battery; Step 3 is specifically as follows: Step 3.1, using the set of data obtained by the EKF1 in Step 2.2 N Q n (1), t 1, SOC (1), (2) Q n (2), t 2, SOC (2), (3) Q n (3), t 3, SOC (3), (4), (5) Q n N , t N , SOC N ), the average battery maximum available capacity is calculated using the time series weighting method represented by Equation (7) Q 0, Q 0Specifically as follows:​​​ wherein Q n (1), Q n (2),… Q n ( N ) are the maximum available capacities of the battery at each discharge current and discharge rate estimated by the EKF1, respectively, denotes the time experienced by the battery at each corresponding discharge current and discharge rate, SOC ( N ) is the SOC at the end of the discharge process; Step 3.2, calculate SOH by using formula (8), specifically as follows: (8) In the formula, Q rate C is the rated capacity of the battery; Step 3.

3. SOH of step 3.2 represents a state of health for a depth of discharge. SOH is analyzed: if there are w states of health for the same or different depths of discharge, then the average of the w states of health is taken as the global state of health over a long time range as shown in equation (9): (9) wherein SOH 1、 SOH 2、┅┅、 SOH w respectively represent the health state at the same or different discharge depth in a long time range.

2. The lithium-ion battery state-of-charge and state-of-health combined estimation method of claim 1, wherein, Step 1 is specifically as follows: Step 1.1, establishing a fractional order equivalent circuit model of a lithium ion battery, including a power source (1), an ohmic resistance R 0, an electrochemical polarization resistance R 1, a concentration polarization resistance R 2, and a Warburg element C W , further including a polarization capacitance R 1 in parallel with the electrochemical polarization resistance CPE 1, and a concentration polarization capacitance R 2 in parallel with the concentration polarization resistance CPE 2; Wherein, CPE1 and CPE2 are constant phase elements, and their impedances are as shown in formula (1): （1） wherein α、β are constant phase elements CPE 1 and CPE 2 are fractional orders, s is the Laplace operator; Volberg element C W represents the positive and negative electrode solid-phase diffusion reaction, and the impedance thereof is shown as formula (2): （2） wherein γ denotes a Volberg element C W fractional order; Step 1.2, establish a mathematical model of the fractional order equivalent circuit model parameters of the lithium ion battery, specifically as follows: （3） wherein η is the charge and discharge efficiency; Q n is the maximum available capacity of the battery; T s is the sampling time; I L (t) is the battery discharge current; U 1( t ) is the voltage of the constant phase element CPE 1; U 2( t ) is the voltage of the constant phase element CPE 2; U 3( t ) is the voltage of the element C W ; SOC ( t ) is the state of charge of the battery; U o ( t ) is the battery terminal voltage at the t moment. Step 1.3, according to the Kirchhoff voltage law and the fractional order integral equation of the fractional order equivalent circuit model of the lithium ion battery, after discretization, formula (4) is obtained, as shown below: In the formula, I L ( k ) indicates the first k The battery discharge current sampled at constant times; U 1( k ), U 1( k +1) respectively represent the first k , k The constant-phase element sampled at time +1 CPE 1 voltage; U 2( k ), U 2( k +1) respectively represent the first k , k The constant-phase element sampled at time +1 CPE 2. Voltage; U 3( k ), U 3( k +1) respectively represent the first k , k The element sampled at time +1 C W The voltage; SOC ( k ), SOC ( k +1) respectively represent the first k , k Battery state of charge at time +1; U o ( k ) indicates the first k Battery terminal voltage at all times; Step 1.4, Definition x k =[ SOC ( k ), U 1( k ), U 2( k ), U 3( k ), R 0( k ), 1 / Q n ( k )] T , θ k =[1 / R 1,k , 1 / CPE 1,k , α k ， 1 / R 2,k , 1 / CPE 2,k , β k ， 1 / C W,k , γ k ], u k = I L ( k ), y k = U o ( k Equation (4) is rearranged into equation (5): Wherein: （6） wherein x i,k denotes x k the first i element of the sequence θ i,k denotes θ k the first i element of the sequence 3. The lithium-ion battery state-of-charge and state-of-health combined estimation method of claim 1, wherein, In step 2.2, EKF2 and EKF1 need to be repeated for each estimation, and it is necessary to judge whether the discharge current and discharge rate of the battery change during the recording period. If the discharge current and discharge rate do not change, the corresponding discharge time is continuously recorded. If the discharge current and discharge rate change, the time recording of the last time is ended, and the corresponding time recording under the new discharge current and discharge rate is started. The estimation process and time recording process of EKF2 and EKF1 are cycled in this way until the battery is completely discharged or the discharge process is ended. EKF1 is repeated N After 0(1) seconds, a series of ohmic internal resistances are obtained, and each ohmic internal resistance is represented as R 0(2),... R 0(3),... R 0(4),... N 0(5),...In addition, a batch of data is obtained, which is respectively Q n (1), t 1, SOC (1), Q n (2), t 2, SOC (2), Q n (3), t 3, SOC (3), Q n N , t N , SOC N .​ 4. The lithium-ion battery state-of-charge and state-of-health combined estimation method of claim 3, wherein, At the same time, the ohmic internal resistance estimated by EKF1 in step 2.2 is represented as R 0(1)、 R 0(2)┅┅、 R 0( N )auxiliary evaluation, as the discharge process, so over time, the degree of aging of the battery gradually deepened, so the ohmic internal resistance will gradually become larger, so under normal circumstances R 0( N )maximum, R 0( N )and the ohmic internal resistance of the new battery, if R 0( N )is not less than m times the ohmic internal resistance of the new battery, it is considered that the battery has been aged; If R 0( N ) is less than m times the new battery ohmic resistance, the degree of battery aging needs to be combined with the global health status SOH whole determined, if SOH whole ≥ 50%, although it has aged, but can be used in echelon; if SOH whole < 50%, it is considered that the battery has aged.

Images