Energy storage battery state on-line monitoring method and system

Abstract

The invention discloses an energy storage battery state online monitoring method and system, and belongs to the technical field of battery management. According to the method, the accurate ampere-hour model and the capacity model are constructed, so that a solid foundation is provided for evaluation of the battery state. By introducing a double-extended Kalman filtering method, an EKF1 module and an EKF2 module are mutually iterated, Q and R matrixes are dynamically adjusted through AMPSO, the monitoring precision of the battery state is effectively improved, the dynamic change of the battery can be captured in real time, key parameters such as the residual capacity and the internal resistance of the battery can be accurately estimated, and powerful support is provided for health management and maintenance of the battery.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and more particularly to an online monitoring method and system for energy storage battery status. Background Art

[0002] With the development of lithium batteries and energy storage technology, the requirements for battery management systems are becoming increasingly higher. The detection of battery status in battery management systems plays an important role and has a significant impact on battery performance and battery safety.

[0003] At present, although a variety of battery status monitoring technologies have been applied to the market, these technologies still expose many shortcomings in practical applications. On the one hand, the estimation of SOC is often affected by multiple factors such as battery model accuracy, sensor error, and changes in operating conditions, resulting in deviations in the estimation results. This deviation not only affects the decision-making accuracy of the battery management system, but may also cause safety problems such as battery overcharging and over-discharging. On the other hand, the evaluation of SOH also faces many difficulties, such as the complexity of the battery aging mechanism, the limitations of test conditions, and the inconsistency of evaluation standards, all of which make the accurate evaluation of SOH particularly difficult.

[0004] In addition, existing battery status monitoring methods also have certain limitations in terms of real-time performance, stability and reliability. With the continuous advancement of battery technology and the widespread application of energy storage systems, the requirements for battery status monitoring are becoming increasingly higher, and it is necessary to obtain battery status information more quickly, accurately and stably. However, existing monitoring methods often find it difficult to meet these requirements, especially in complex working conditions and harsh environments, where the performance of the monitoring system is often seriously affected.

[0005] Therefore, how to propose an online monitoring method and system for energy storage battery status to improve the detection accuracy of battery status is a problem that technical personnel in this field need to solve urgently. Summary of the invention

[0006] In view of this, the present invention provides a method and system for online monitoring of the state of an energy storage battery, which realizes the coordinated estimation of SOC and SOH and performs high-precision monitoring of the battery state.

[0007] In order to achieve the above object, the present invention adopts the following technical solution:

[0008] In one aspect, the present invention discloses a method for online monitoring of energy storage battery status, comprising the following steps:

[0009] Construct the battery's ampere-hour model and capacity model;

[0010] A double extended Kalman filter method is adopted, wherein the double extended Kalman filter method uses two EKF modules, including EKF1 and EKF2; the output of EKF1 is used as the input of EKF2, and the output of EKF2 is used as the input of the next iteration cycle of EKF1 to correct the output of EKF1 until the iteration ends; wherein EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices;

[0011] The energy storage battery state is estimated based on the ampere-hour model, the capacity model and the dual extended Kalman filter structure.

[0012] Preferably, the ampere-hour model is as follows:

[0013]

[0014] Formula (2) is the SOC state equation, and formula (3) is the SOC observation equation; where SOC t is the SOC value at time t in the continuous time, SOC0 is the initial value of SOC, Q0 is the rated capacity of the battery, η is the coulomb efficiency, I is the current value, and t is the time; SOC k For SOC t The SOC value at time k after discretization sampling, U 1,k is the value of U1 at time k, U 2,k is the U2 value at time k, Δt is the sampling time interval, τ is the time constant, and w k is the system noise parameter, v k is the observation noise parameter, I k is the current value at time k, U L,k+1 is the terminal voltage value at time k+1, U OC,k+1 is the open circuit voltage value at time k+1, U1 is the voltage of the electrochemical polarization capacitor in the second-order RC equivalent circuit model, and U2 is the voltage of the concentration difference polarization capacitor in the second-order RC equivalent circuit model.

[0015] Preferably, the capacity model is as follows:

[0016]

[0017] Formula (4) is the battery health state equation, formula (5) is the battery capacity state equation, and formula (6) is the battery capacity prediction equation; SOH Q is the SOH value of the capacity model, Qn is the current available capacity of the battery, Q0 is the initial rated capacity of the battery; Q k represents the battery capacity at the kth moment, Q k+1 represents the battery capacity at the next moment k+1, r is the system noise; d k A measure of charge or charge imbalance, SOC kis the state of charge of the battery at time k, SOC k-1 is the state of charge of the battery at time k-1, η is the charge and discharge efficiency coefficient, I t,k-1 is the current value at time k-1, T is the time interval, Q k-1 is the capacity of the battery at time k-1, and ξ is the measurement noise.

[0018] Preferably, the double extended Kalman filter method comprises the following steps:

[0019] S31. Based on formula (2) and formula (3), SOC is used as the only variable in the SOC state equation, namely, the EKF1 state variable, U L As a variable of the SOC observation equation, Q0 is used as a constant, and the SOC estimation value is output by EKF1;

[0020] S32. Based on formula (5) and formula (6), the SOC estimation value is used as the input of EKF2, Q n As the variable of the capacity state equation, namely the EKF2 state variable, d k+1 As a variable of the capacity observation equation, use EKF2 to output Q n Estimated value;

[0021] S33. Using the Q n Estimated value, calculate the estimated value of SOH according to formula (4);

[0022] S34. Using the Q at the current moment n The estimated value replaces Q0 in EKF1, outputs the SOC estimate for the next moment, and repeats S32 and S33 to obtain Q n The estimated value and SOH estimated value are iteratively updated accordingly until the iteration ends.

[0023] Preferably, the workflow of EKF1 is as follows:

[0024] a1) Parameter initialization:

[0025] Where x0 is the initial value of the EKF1 state variable, is the expectation of x0, is the initial value of the covariance in EKF1, Q0 is the initial value of the state variable in EKF2, is the expectation of Q0, is the initial value of the covariance in EKF2;

[0026] a2) According to formula (2) and formula (3), the SOC state equation and SOC observation equation are simplified as follows:

[0027] x k+1 =Ax x k +B x I k +w k

[0028] y k+1 =C x x k+1 +D x I x+1 +v k

[0029] Among them, x k and x k+1 denote the EKF1 state variables at time k and time k+1 respectively, A x , B x is the coefficient matrix of the SOC state equation, y k+1 is the SOC observation value at time k+1, C x , D x is the coefficient matrix of the SOC observation equation, I k+1 is the current at time k+1;

[0030] a3) The state prediction and covariance prediction of SOC in EKF1 are as follows:

[0031]

[0032] is the initial estimate of the EKF1 state variable at time k+1, is the estimated value of the EKF1 state variable at time k, P x (k+1|k) is the error covariance matrix at time k+1, P x (k|k) is the error covariance matrix at time k, and W(k) is w k expectations.

[0033] a4) Calculate the Kalman gain K x

[0034]

[0035] In the formula, K x (k+1) is the Kalman gain at time k+1 in EKF1, V x v k expectations;

[0036] a5) SOC state update and covariance update

[0037]

[0038] P x (k+1|k+1)=[HKx (k+1)C x (k+1)]P x (k+1|k)

[0039] Where y(k+1) is the SOC observation value at time k+1, is the estimated SOC observation value at time k+1, is the updated value of the EKF1 state variable at time k+1, P x (k+1|k+1) is the updated value of the covariance in EKF1 at time k+1, and H represents the unit matrix.

[0040] Preferably, the workflow of EKF2 is as follows:

[0041] b1) According to formula (5) and formula (6), the capacity state equation and capacity observation equation are simplified as follows:

[0042] q k+1 =q k +r

[0043] d k+1 =C Q (x k+1 -x k )+D Q I k +ξ

[0044] Among them, q k and q k+1 are the capacity state variables at time k and time k+1, C Q , D Q is the coefficient matrix of the capacity observation equation;

[0045] b2) State prediction and covariance prediction of capacity in EKF2

[0046]

[0047] P Q (k+1|k)=P Q (k|k)+R(k)

[0048] In the formula, is the initial estimate of the EKF2 state variable at time k+1, is the estimated value of the EKF2 state variable at time k, P Q (k+1|k) is the error covariance matrix at time k+1 in EKF2, P Q (k|k) is the error covariance matrix at time k, and R(k) is the expectation of r;

[0049] b3) Calculate the Kalman gain K Q

[0050]

[0051] In the formula, K Q (k+1) is the Kalman gain at time k+1 in EKF2, V Q is the expectation of ξ;

[0052] b4) Capacity update and covariance update

[0053]

[0054] P Q (k+1|k+1)=[HK Q (k+1)C Q (k+1)]P Q (k+1|k)

[0055] In the formula, y Q (k+1) is the capacity observation value at time k+1 in EKF2, is the capacity estimation observation value at time k+1, represents the updated value of the EKF2 state variable at time k+1, P Q (k+1|k+1) represents the updated value of the covariance in EKF2 at time k+1.

[0056] Preferably, EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices, including:

[0057] 1) Initialize the particle swarm, give the initial values ​​of the position and velocity of the diagonal elements of the Q and R matrices, initialize the fitness value, and set the best position P of the individual best-id Set as the current position, the global best position G best-d Set it to the position with the minimum fitness value;

[0058] 2) Update the diagonal element values ​​of the Q and R matrices;

[0059] 3) Connect the diagonal elements of the Q and R matrices through the matrix connection module to obtain the optimized Q and R matrices of EKF1 and EKF2 respectively;

[0060] 4) Calculate the fitness value of EKF1 and EKF2 according to the current estimated value and observed value, and update the current individual extreme value, group extreme value and learning radius;

[0061] 5) Perform mixed mutation operations on extreme values ​​according to design rules;

[0062] 6) When the maximum number of iterations is reached, the group extreme value is output, which is the optimized Q and R matrices. Otherwise, go to 2).

[0063] On the other hand, the present invention also proposes an energy storage battery status online monitoring system, which is used to implement the above energy storage battery status online monitoring method, comprising:

[0064] Battery model building module, building battery ampere-hour model and capacity model;

[0065] The EKF module adopts a double extended Kalman filter method, which uses two EKF modules, including EKF1 and EKF2; the output of EKF1 is used as the input of EKF2, and the output of EKF2 is used as the input of the next iteration cycle of EKF1 to correct the output of EKF1 until the iteration ends; wherein EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices;

[0066] The battery state estimation module estimates the energy storage battery state based on the ampere-hour model, the capacity model and the dual extended Kalman filter structure.

[0067] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses an online monitoring method and system for the state of an energy storage battery, which provides a solid foundation for the evaluation of the battery state by constructing an accurate ampere-hour model and a capacity model. By introducing a dual extended Kalman filter method, in which the EKF1 and EKF2 modules iterate each other and use AMPSO to dynamically adjust the Q and R matrices, the monitoring accuracy of the battery state is effectively improved, which can not only capture the dynamic changes of the battery in real time, but also accurately estimate the remaining capacity, internal resistance and other key parameters of the battery, providing strong support for the health management and maintenance of the battery. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0069] Figure 1 A schematic diagram of a process flow for estimating the state of an energy storage battery based on a double extended Kalman filter provided by the present invention;

[0070] Figure 2 The block diagram of the principle of optimizing the DEKF noise matrix using AMPSO;

[0071] Figure 3 Flowchart for optimizing the Q and R covariance matrices of DEKF using the asynchronous learning factor AMPSO;

[0072] Figure 4 This is a schematic diagram of the second-order RC equivalent circuit model. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] In one aspect, an embodiment of the present invention discloses a method for online monitoring of a state of an energy storage battery, comprising the following steps:

[0075] S1. Construct the battery's ampere-hour model and capacity model.

[0076] According to the second-order RC equivalent circuit model (ref. Figure 4 ), establish the battery's ampere-hour integral model and capacity model, Figure 4 Where Uoc is the open circuit voltage, U is the terminal voltage, and R0 is the ohmic internal resistance; R1 and C1 are used to simulate the electrochemical polarization reaction, R1 is the electrochemical polarization resistor, and C1 is the electrochemical polarization capacitor; R2 and C2 are used to simulate the concentration difference polarization reaction, R2 is the concentration difference polarization resistor, and C2 is the concentration difference polarization capacitor.

[0077] Among them, the Anshi model is constructed as follows:

[0078]

[0079] Formula (2) is the SOC state equation, and formula (3) is the SOC observation equation; where SOC t is the SOC value at time t in the continuous time, SOC0 is the initial value of SOC, Q0 is the rated capacity of the battery, η is the coulomb efficiency, I is the current value, and t is the time; SOC k For SOC t The SOC value at time k after discretization sampling, U 1,k is the value of U1 at time k, U 2,k is the U2 value at time k, Δt is the sampling time interval, τ is the time constant, and w k is the system noise parameter, v k is the observation noise parameter; I k is the current value at time k, U L,k+1 is the terminal voltage value at time k+1, U OC,k+1 is the open circuit voltage value at time k+1, U1 is the voltage of the electrochemical polarization capacitor in the second-order RC equivalent circuit model, and U2 is the voltage of the concentration difference polarization capacitor in the second-order RC equivalent circuit model.

[0080] Furthermore, the capacity model formula is as follows:

[0081]

[0082] Formula (4) is the battery health state equation, formula (5) is the battery capacity state equation, and formula (6) is the battery capacity prediction equation; SOH Q is the SOH value of the capacity model, Q n is the current available capacity of the battery, Q0 is the initial rated capacity of the battery; Q k represents the battery capacity at the kth moment, Q k+1 represents the battery capacity at the next moment k+1, r is the system noise; d k A measure of charge or charge imbalance, SOC k is the state of charge of the battery at time k, SOC k-1 is the state of charge of the battery at time k-1; η is the charge and discharge efficiency coefficient, which is used to indicate energy loss or efficiency; I t,k-1 is the current value at time k-1; T is the time interval, indicating the discrete time step; Q k-1 is the capacity of the battery at time k-1, and ξ is the measurement noise.

[0083] S2. A double extended Kalman filter method is adopted. The double extended Kalman filter method uses two EKF modules, including EKF1 and EKF2; the output of EKF1 is used as the input of EKF2, and the output of EKF2 is used as the input of the next iteration cycle of EKF1 to correct the output of EKF1 until the iteration ends; among them, EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices.

[0084] There is a coupling effect between SOC and SOH. Estimating SOC or SOH alone cannot guarantee the accurate and true state value of the battery. Therefore, this embodiment realizes the coordinated estimation of SOC and SOH by constructing a double extended Kalman filter. The first EKF realizes the estimation of SOC, and uses the estimation result as the input of the second EKF to realize the progressive estimation of SOH. Then, the estimated SOH value at the current moment obtained by the second EKF is used as the input of the next iterative cycle, and the estimated value of SOC at the next moment is further corrected to form a closed loop until the iteration ends. After the mutual correction feedback between the two, the coordinated estimation of the two is finally realized.

[0085] The process of the double extended Kalman filter method is as follows Figure 1 As shown in the figure, it mainly includes two modules: EKF1 and EKF2. Among them, EKF1 is responsible for the prediction of SOC, and EKF2 is responsible for the prediction of SOH. The two complement each other and provide feedback correction to each other, achieving the purpose of collaborative estimation of SOC and SOH, laying the foundation for the overall establishment of the collaborative estimation algorithm model in the following embodiment.

[0086] S3. Estimate the state of the energy storage battery based on the ampere-hour model, capacity model and dual extended Kalman filter structure. S31. Based on formula (2) and formula (3), SOC is used as the only variable in the SOC state equation, namely the EKF1 state variable, U L As a variable of the SOC observation equation, Q0 is used as a constant, and the SOC estimation value is output by EKF1;

[0087] S32. Based on formula (5) and formula (6), the SOC estimation value is used as the input of EKF2, Q n As the variable of the capacity state equation, namely the EKF2 state variable, d k+1 As a variable of the capacity observation equation, use EKF2 to output Q n Estimated value;

[0088] S33. Use Q n Estimated value, calculate the estimated value of SOH according to formula (4);

[0089] S34. Use the current Q n The estimated value replaces Q0 in EKF1, outputs the SOC estimate for the next moment, and repeats S32 and S33 to obtain Q n The estimated value and SOH estimated value are iteratively updated accordingly until the iteration ends.

[0090] Preferably, the workflow of EKF1 is as follows:

[0091] a1) Parameter initialization:

[0092] Where x0 is the initial value of the EKF1 state variable, is the expectation of x0, is the initial value of the covariance in EKF1, Q0 is the initial value of the state variable in EKF2, is the expectation of Q0, is the initial value of the covariance in EKF2;

[0093] a2) According to formula (2) and formula (3), the SOC state equation and SOC observation equation are simplified as follows:

[0094] x k+1 =A x x k +B x I x +w k

[0095] y k+1 =C x x k+1 +D x Ik+1 +v k

[0096] Among them, x k and x k+1 denote the EKF1 state variables at time k and time k+1 respectively, A x , B x is the coefficient matrix of the SOC state equation, y k+1 is the SOC observation value at time k+1, C x , D x is the coefficient matrix of the SOC observation equation, I k+1 is the current at time k+1;

[0097] a3) The state prediction and covariance prediction of SOC in EKF1 are as follows:

[0098]

[0099] is the initial estimate of the EKF1 state variable at time k+1, is the estimated value of the EKF1 state variable at time k, P x (k+1|k) is the error covariance matrix at time k+1, P x (k|k) is the error covariance matrix at time k, and W(k) is w k expectations.

[0100] a4) Calculate the Kalman gain K x

[0101]

[0102] In the formula, K x (k+1) is the Kalman gain at time k+1 in EKF1, V x v k expectations;

[0103] a5) SOC state update and covariance update

[0104]

[0105] P x (k+1|k+1)=[HK x (k+1)C x (k+1)]P x (k+1|k)

[0106] Where y(k+1) is the SOC observation value at time k+1, is the estimated SOC observation value at time k+1, is the updated value of the EKF1 state variable at time k+1, P x (k+1|k+1) is the updated value of the covariance in EKF1 at time k+1, and H represents the unit matrix.

[0107] Preferably, the workflow of EKF2 is as follows:

[0108] b1) According to formula (5) and formula (6), the capacity state equation and capacity observation equation are simplified as follows:

[0109] q k+1 =q k +r

[0110] d k+1 =C Q (x k+1 -x k )+D Q I k +ξ

[0111] Among them, q k and q k+1 are the capacity state variables at time k and time k+1, C Q , D Q is the coefficient matrix of the capacity observation equation;

[0112] b2) State prediction and covariance prediction of capacity in EKF2

[0113]

[0114] P Q (k+1|k)=P Q (k|k)+R(k)

[0115] In the formula, is the initial estimate of the EKF2 state variable at time k+1, is the estimated value of the EKF2 state variable at time k, P Q (k+1|k) is the error covariance matrix at time k+1 in EKF2, P Q (k|k) is the error covariance matrix at time k, and R(k) is the expectation of r;

[0116] b3) Calculate the Kalman gain K Q

[0117]

[0118] In the formula, K Q (k+1) is the Kalman gain at time k+1 in EKF2, V Q is the expectation of ξ;

[0119] b4) Capacity update and covariance update

[0120]

[0121] P Q (k+1|k+1)=[HK Q (k+1)C Q (k+1)]P Q (k+1|k)

[0122] In the formula, y Q (k+1) is the capacity observation value at time k+1 in EKF2, is the capacity estimation observation value at time k+1, represents the updated value of the EKF2 state variable at time k+1, P Q (k+1|k+1) represents the updated value of the covariance in EKF2 at time k+1.

[0123] Preferably, the principle block diagram of this embodiment using asynchronous learning factor AMPSO to optimize the DEKF noise matrix is ​​as follows: Figure 2 As shown in Figure 2, the optimization idea is to use AMPSO to dynamically adjust the Q and R matrices to minimize the error between the EKF state estimate and the measured value (minimum fitness value). Figure 3 , the algorithm steps are as follows:

[0124] 1) Initialize the particle swarm, give the initial values ​​of the position and velocity of the diagonal elements of the Q and R matrices, initialize the fitness value, and set the best position P of the individual best-id Set as the current position, the global best position G best-d Set it to the position with the minimum fitness value;

[0125] 2) Update the diagonal element values ​​of the Q and R matrices;

[0126] 3) Connect the diagonal elements of the Q and R matrices through the matrix connection module to obtain the optimized Q and R matrices of EKF1 and EKF2 respectively;

[0127] 4) Calculate the fitness value of EKF1 and EKF2 according to the current estimated value and observed value, and update the current individual extreme value, group extreme value and learning radius;

[0128] 5) Perform mixed mutation operations on extreme values ​​according to design rules;

[0129] 6) When the maximum number of iterations is reached, the group extreme value is output, which is the optimized Q and R matrices. Otherwise, go to 2).

[0130] On the other hand, the present invention also proposes an energy storage battery status online monitoring system, which is used to implement the above energy storage battery status online monitoring method, comprising:

[0131] Battery model building module, building battery ampere-hour model and capacity model;

[0132] The EKF module adopts a double extended Kalman filter method. The double extended Kalman filter method uses two EKF modules, including EKF1 and EKF2. The output of EKF1 is used as the input of EKF2, and the output of EKF2 is used as the input of the next iteration cycle of EKF1 to correct the output of EKF1 until the iteration ends. Among them, EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices.

[0133] The battery state estimation module estimates the energy storage battery state based on the ampere-hour model, capacity model and double extended Kalman filter structure.

[0134] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0135] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for online monitoring of energy storage battery status, characterized in that: The following steps are involved: Construct the battery's ampere-hour model and capacity model; A double extended Kalman filter method is adopted, wherein the double extended Kalman filter method uses two EKF modules, including EKF1 and EKF2; the output of EKF1 is used as the input of EKF2, and the output of EKF2 is used as the input of the next iteration cycle of EKF1 to correct the output of EKF1 until the iteration ends; wherein EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices; The energy storage battery state is estimated based on the ampere-hour model, the capacity model and the dual extended Kalman filter structure.

2. The method for online monitoring of energy storage battery status according to claim 1, characterized in that: The ampere-hour model is as follows: Formula (2) is the SOC state equation, and formula (3) is the SOC observation equation; where SOC t is the SOC value at time t in the continuous time, SOC0 is the initial value of SOC, Q0 is the rated capacity of the battery, η is the coulomb efficiency, I is the current value, and t is the time; SOC k For SOC t The SOC value at time k after discretization sampling, U 1,k is the value of U1 at time k, U 2,k is the U2 value at time k, Δt is the sampling time interval, τ is the time constant, and w k is the system noise parameter, v k is the observation noise parameter; I k is the current value at time k, U L,k+1 is the terminal voltage value at time k+1, U OC,k+1 is the open circuit voltage value at time k+1, U1 is the voltage of the electrochemical polarization capacitor in the second-order RC equivalent circuit model, and U2 is the voltage of the concentration difference polarization capacitor in the second-order RC equivalent circuit model.

3. The method for online monitoring of energy storage battery status according to claim 2, characterized in that: The capacity model is as follows: Formula (4) is the battery health state equation, formula (5) is the battery capacity state equation, and formula (6) is the battery capacity prediction equation; In formula (4), SOH Q is the SOH value of the capacity model, Q n is the current available capacity of the battery, Q0 is the initial rated capacity of the battery; (5) where Q k represents the battery capacity at the kth moment, Q k+1 represents the battery capacity at the next moment k+1, r is the system noise; (6) where d k A measure of charge or charge imbalance, SOC k is the state of charge of the battery at time k, SOC k-1 is the state of charge of the battery at time k-1, η is the charge and discharge efficiency coefficient, I t,k-1 is the current value at time k-1, T is the time interval, Q k-1 is the capacity of the battery at time k-1, and ξ is the measurement noise.

4. The method for online monitoring of energy storage battery status according to claim 3, characterized in that: The double extended Kalman filter method comprises the following steps: S31. Based on formula (2) and formula (3), SOC is used as the only variable in the SOC state equation, namely, the EKF1 state variable, U L As a variable of the SOC observation equation, Q0 is used as a constant, and the SOC estimation value is output by EKF1; S32. Based on formula (5) and formula (6), the SOC estimation value is used as the input of EKF2, Q n As the variable of the capacity state equation, namely the EKF2 state variable, d k+1 As a variable of the capacity observation equation, use EKF2 to output Q n Estimated value; S33. Using the Q n Estimated value, calculate the estimated value of SOH according to formula (4); S34. Using the Q at the current moment n The estimated value replaces Q0 in EKF1, outputs the SOC estimate for the next moment, and repeats S32 and S33 to obtain Q n The estimated value and SOH estimated value are iteratively updated accordingly until the iteration ends.

5. The method for online monitoring of energy storage battery status according to claim 4, characterized in that: The workflow of EKF1 is as follows: a1) Parameter initialization: Where x0 is the initial value of the EKF1 state variable, is the expectation of x0, is the initial value of the covariance in EKF1, Q0 is the initial value of the state variable in EKF2, is the expectation of Q0, is the initial value of the covariance in EKF2; a2) According to formula (2) and formula (3), the SOC state equation and SOC observation equation are simplified as follows: x k+1 =A x x k +B x I k +w k y k+1 =C x x k+1 +D x I x+1 +v k Among them, x k and x k+1 denote the EKF1 state variables at time k and time k+1 respectively, A x , B x is the coefficient matrix of the SOC state equation, y k+1 is the SOC observation value at time k+1, C x , D x is the coefficient matrix of the SOC observation equation, I k+1 is the current at time k+1; a3) The state prediction and covariance prediction of SOC in EKF1 are as follows: is the initial estimate of the EKF1 state variable at time k+1, is the estimated value of the EKF1 state variable at time k, P x (k+1|k) is the error covariance matrix at time k+1, P x (k|k) is the error covariance matrix at time k, and W(k) is w k expectations. a4) Calculate Kalman gain K x In the formula, K x (k+1) is the Kalman gain at time k+1 in EKF1, V x v k expectations; a5) SOC state update and covariance update P x (k+1|k+1)=[H-K x (k+1)C x (k+1)]P x (k+1|k) Where y(k+1) is the SOC observation value at time k+1, is the estimated SOC observation value at time k+1, is the updated value of the EKF1 state variable at time k+1, P x (k+1|k+1) is the updated value of the covariance in EKF1 at time k+1, and H represents the unit matrix.

6. The method for online monitoring of energy storage battery status according to claim 5, characterized in that: The workflow of EKF2 is as follows: b1) According to formula (5) and formula (6), the capacity state equation and capacity observation equation are simplified as follows: q k+1 =q k +r d k+1 =C Q (x k+1 -x k )+D Q I k +ξ Among them, q k and q k+1 are the capacity state variables at time k and time k+1, C Q , D Q is the coefficient matrix of the capacity observation equation; b2) State prediction and covariance prediction of capacity in EKF2 P Q (k+1|k)=P Q (k|k)+R(k) In the formula, is the initial estimate of the EKF2 state variable at time k+1, is the estimated value of the EKF2 state variable at time k, P Q (k+1|k) is the error covariance matrix at time k+1 in EKF2, P Q (k|k) is the error covariance matrix at time k, and R(k) is the expectation of r; b3) Calculate the Kalman gain K Q In the formula, K Q (k+1) is the Kalman gain at time k+1 in EKF2, V Q is the expectation of ξ; b4) Capacity update and covariance update P Q (k+1|k+1)=[H-K Q (k+1)C Q (k+1)]P Q (k+1|k) In the formula, y Q (k+1) is the capacity observation value at time k+1 in EKF2, is the capacity estimation observation value at time k+1, represents the updated value of the EKF2 state variable at time k+1, P Q (k+1|k+1) represents the updated value of the covariance in EKF2 at time k+1.

7. The method for online monitoring of energy storage battery status according to claim 1, characterized in that: EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices, including: 1) Initialize the particle swarm, give the initial values ​​of the position and velocity of the diagonal elements of the Q and R matrices, initialize the fitness value, and set the best position P of the individual best-id Set as the current position, the global best position G best-d Set it to the position with the minimum fitness value; 2) Update the diagonal element values ​​of the Q and R matrices; 3) Connect the diagonal elements of the Q and R matrices through the matrix connection module to obtain the optimized Q and R matrices of EKF1 and EKF2 respectively; 4) Calculate the fitness value of EKF1 and EKF2 according to the current estimated value and observed value, and update the current individual extreme value, group extreme value and learning radius; 5) Perform mixed mutation operations on extreme values ​​according to design rules; 6) When the maximum number of iterations is reached, the group extreme value is output, which is the optimized Q and R matrices. Otherwise, go to 2).

8. An online monitoring system for energy storage battery status, characterized in that: include: Battery model building module, building battery ampere-hour model and capacity model; The EKF module adopts a double extended Kalman filter method, which uses two EKF modules, including EKF1 and EKF2; the output of EKF1 is used as the input of EKF2, and the output of EKF2 is used as the input of the next iteration cycle of EKF1 to correct the output of EKF1 until the iteration ends; wherein EKF1 and EKF2 use AMPSO to dynamically adjust the Q and R matrices; The battery state estimation module estimates the energy storage battery state based on the ampere-hour model, the capacity model and the dual extended Kalman filter structure.

Images