Multi-time scale balancing strategy based on state of charge and capacity

Abstract

The present application discloses a multi-time scale balancing strategy based on state of charge and capacity for a passive balancing circuit, including first establishing a single-cell minimum capacity-difference model; wherein, it includes a single-cell minimum capacity model and a single-cell difference model; then, based on the single-cell minimum capacity-difference model, the state of charge and capacity of each single cell in the battery pack are estimated at multiple time scales through a double extended Kalman filter; finally, based on the state of charge and capacity estimation results of each single cell, the ampere-hour integral required for balancing of each battery is calculated, and the balancing current is output. In this way, by establishing a model, the dynamic behavior of the MCC and other single cells in the battery pack is accurately described, and at the same time, based on the model and the double extended Kalman filter, the single cell SOC and capacity in the battery pack are accurately estimated at multiple time scales. This solves the problem in the prior art that it is difficult to make accurate capacity estimation, and it must be based on accurate SOC estimation.

Description

Technical Field

[0001] The present invention belongs to the technical field of battery management, and in particular relates to a multi-time scale balancing strategy based on state of charge and capacity. Background Art

[0002] According to the balancing variables, the balancing strategies can be divided into voltage-based, state of charge (SOC)-based and capacity-based balancing strategies. The voltage-based balancing strategy uses the voltage at the positive and negative terminals of the battery as the control variable during operation, and formulates a strategy to achieve the voltage of each cell in the battery pack to be consistent or within the threshold range. This method has the advantages of direct measurement, high measurement accuracy and simple control algorithm. However, the voltage cannot accurately represent the differences in the internal state of the battery, which can easily lead to over-balancing, increasing the balancing energy consumption and balancing time. The SOC-based balancing strategy uses the battery's SOC as the balancing variable, and its balancing goal is to achieve the SOC of each cell in the battery pack to be consistent or within the threshold range. This method has the advantages of fully utilizing the battery pack's electrical energy, avoiding differences in aging speed caused by different discharge depths, and shortening the balancing time. However, during the battery's full life cycle, as the battery degrades, its characteristic parameters will also change, and accurate estimation of SOC is relatively difficult. The capacity-based balancing strategy uses the total capacity, rechargeable capacity or releasable capacity as the balancing variable, and the balancing goal is to maximize the total capacity of the battery pack and improve the capacity utilization of the battery pack. This method can maximize the capacity of the battery pack and avoid the problems of high equalization energy consumption and long equalization time caused by over-equalization. However, accurate capacity estimation is difficult to achieve and requires accurate SOC estimation. Summary of the invention

[0003] In view of this, the purpose of the present invention is to overcome the shortcomings of the prior art and provide a multi-time scale SOC and capacity-based balancing strategy for a passive balancing circuit to solve the problems in the prior art of highly accurate estimation of the SOC and capacity of each cell in a battery pack under the premise of reasonable computational complexity, and the traditional voltage and SOC-based balancing strategy can only achieve partial performance indicators such as maximizing the battery pack capacity, minimizing the balancing energy consumption and balancing time, and it is difficult to take into account the above performance indicators and realize a high-efficiency balancing strategy.

[0004] To achieve the above objectives, the present invention adopts the following technical solution: a multi-time scale SOC and capacity-based balancing strategy for a passive balancing circuit, comprising:

[0005] Establishing a monomer minimum capacity-difference model; the monomer minimum capacity difference model includes a monomer minimum capacity model and a monomer difference model;

[0006] Based on the cell minimum capacity-difference model, the state of charge and capacity of each cell in the battery pack are estimated at multiple time scales through a double extended Kalman filter;

[0007] Based on the state of charge and capacity estimation results of each cell, the ampere-hour integral required for balancing each battery is calculated and the balancing current is output.

[0008] Furthermore, the establishment of the monomer minimum capacity-difference model includes:

[0009] Establish a single-cell minimum capacity model based on the first-order RC model;

[0010] Establish a monomer difference model based on the Rint model;

[0011] Based on the monomer minimum capacity model and the monomer difference model, the monomer minimum capacity-difference model is obtained.

[0012] Furthermore, the method of estimating the state of charge and capacity of each cell in the battery pack at multiple time scales based on the cell minimum capacity-difference model through a double extended Kalman filter includes:

[0013] Based on the single-cell minimum capacity model, the state of charge and parameters of the MCC are estimated by using a dual extended Kalman filter, and based on the state of charge estimation result, the capacity of the MCC is estimated by using an extended Kalman filter;

[0014] Based on the monomer difference model, the ΔSOC and Δ parameter of other monomers are estimated through a double extended Kalman filter, and the ΔSOC of the other monomers and the SOC of the MCC are jointly loaded into the capacity estimation EKF to obtain the remaining monomer capacity estimation.

[0015] Furthermore, the method of calculating the ampere-hour integral required for balancing of each battery cell based on the state of charge and capacity estimation result of each battery cell and outputting the balancing current includes:

[0016] When the equilibrium state is not reached, the balancing current is continuously outputted until the equilibrium state is reached; wherein the equilibrium state is that under passive balancing conditions, the battery pack capacity is equal to the MCC capacity.

[0017] Furthermore, the battery pack capacity is equal to the MCC capacity:

[0018] The releasable capacity of the MCC is equal to the minimum releasable capacity of all monomers, and the rechargeable capacity of the MCC is equal to the minimum rechargeable capacity of all monomers;

[0019] Or the releasable capacity of the battery pack or the rechargeable capacity of the battery pack is consistent.

[0020] Furthermore, the state space equation of the monomer minimum capacity model established based on the first-order RC model is:

[0021]

[0022] U T，k =U OCV，k (SOC k )-U P，k -R O，k (I k +I Equ，k-1 );

[0023] Where k is the current execution cycle, k-1 is the previous execution cycle, and Δt is the execution cycle duration. k is the MCC state of charge. η is the MCC coulombic efficiency. C k-1 is the capacity of MCC. k and I Equ，k-1 They are the current in the main circuit and the balancing current, positive for discharge and negative for charge. T，k Represents the MCC terminal voltage. P，k and τ k are the polarization internal resistance and time constant ofRC cirtuit respectively. P，k is the terminal voltage of the RC circuit. O，k U is the ohmic internal resistance. OCV，k Represents OCV. The relationship between OCV and SOC can be fitted by a polynomial function, specifically:

[0024] U OCV =α1SOC 6 +α2SOC 5 +α3SOC 4 +α4SOC 3 +α5SOC 2 +α6SOC 1 +α7;

[0025] Among them, α1~α7 are fitting coefficients.

[0026] Furthermore, the state space equation of the monomer difference model established based on the Rint model is:

[0027] ΔU T，i，k =ΔU OCV，i，k (ΔSOC i，k )-ΔR O，i，k (I k +I Equ，i，k-1 );

[0028] Among them, ΔU T，i，kRepresents the terminal voltage difference between the i-th battery and the MCC (a positive sign indicates that it is greater than U T，k , the negative sign indicates that it is less than U T，k ). ΔU OCV，i，k Represents the OCV difference between the i-th battery and the MCC (a positive sign indicates that it is greater than U OCV，k , the negative sign indicates that it is less than U OCV，k ). ΔSOC i，k Represents the SOC difference between the i-th battery and the MCC (a positive sign indicates a value greater than SOC k , the negative sign indicates that it is less than SOC k ). ΔR O，i，k Represents the ohmic internal resistance difference between the i-th battery and the MCC (a positive sign indicates that it is greater than R O，k , the negative sign indicates that it is less than R O，k ).

[0029] Furthermore, based on the single-cell minimum capacity model, the charge state and parameters of the MCC are estimated by using a double extended Kalman filter, and the nonlinear state space equation of the joint system is:

[0030]

[0031] The nonlinear state-space equation for capacity estimation is as follows:

[0032] [C k ]=[C k-1 ]+r c，k

[0033]

[0034] Where k is the current execution cycle, k-1 is the previous execution cycle, and Δt is the execution cycle duration. k is the MCC state of charge. η is the MCC coulombic efficiency. C k-1 is the capacity of MCC. k and I Equ，k-1 They are the current in the main circuit and the balancing current, positive for discharge and negative for charge. T，k Represents the MCC terminal voltage. P，k and τ k are the polarization internal resistance and time constant ofRC cirtuit respectively. P，k is the terminal voltage of the RC circuit. O，k U is the ohmic internal resistance. OCV，k Represents OCV. The relationship between OCV and SOC can be fitted by a polynomial function, specifically:

[0035] U OCV =α1SOC 6 +α2SOC5 +α3SOC 4 +α4SOC 3 +α5SOC 2 +α6SOC 1 +α7;

[0036] Among them, α1~α7 are fitting coefficients.

[0037] Furthermore, based on the monomer difference model, the ΔSOC and Δ parameters of other monomers are estimated through the double extended Kalman filter. The nonlinear state space equation of the joint system is:

[0038]

[0039]

[0040] ΔU T，i，k =ΔU OCV，i，k -ΔR O，i，k (I k +I Equ，i，k-1 )+v i，k / e i，k ;

[0041] The nonlinear state space equation for capacity estimation is:

[0042]

[0043]

[0044] Where k is the current execution cycle, k-1 is the previous execution cycle, and Δt is the execution cycle duration. k is the MCC state of charge. η is the MCC coulombic efficiency. C k-1 is the capacity of MCC. k and I Equ，k-1 They are the current in the main circuit and the balancing current, positive for discharge and negative for charge. T，k Represents the MCC terminal voltage. P，k and τ k are the polarization internal resistance and time constant ofRC cirtuit respectively. P，k is the terminal voltage of the RC circuit. O，k U is the ohmic internal resistance. OCV，k Represents OCV. The relationship between OCV and SOC can be fitted by a polynomial function, specifically:

[0045] U OCV =α1SOC 6 +α2SOC 5 +α3SOC 4 +α4SOC3 +α5SOC 2 +α6SOC 1 +α7;

[0046] Among them, α1~α7 are fitting coefficients.

[0047] Furthermore, when the equilibrium condition is that the releasable capacity of the MCC is equal to the minimum releasable capacity of all monomers, and the rechargeable capacity of the MCC is equal to the minimum rechargeable capacity of all monomers, the calculation formula of the equilibrium state is:

[0048] C R，min，k =min(C R，i，k )&C C，min，k =min(C C，i，k );

[0049]

[0050] C C，i，k =C C，i，k-1 +C Equ，R，i，k

[0051]

[0052] C Equ，i，k =C Equ，R，i，k +C Equ，C，i，k ;

[0053] When the balancing condition is that the releasable capacity of the battery pack or the rechargeable capacity of the battery pack is consistent, the calculation formula is:

[0054] C C，1，k =C C，min，k =…=C C，N，k ;

[0055] C Equ，i，k =max(C C，i，k )-C C，i，k ;

[0056] Among them, C R，i，k is the monomer releasable capacity, C C，i，k is the single-cell rechargeable capacity, C Equ，R，i，k is the release equilibrium capacity of the monomer, C Equ，C，i，k is the release equilibrium capacity of the monomer, C Equ，i，k is the total balanced capacity.

[0057] The present invention adopts the above technical scheme, firstly establishes a monomer minimum capacity-difference model; the monomer minimum capacity difference model includes a monomer minimum capacity model and a monomer difference model; then based on the monomer minimum capacity-difference model, the state of charge and capacity of each monomer in the battery pack are estimated at multiple time scales through a double extended Kalman filter; finally, based on the state of charge and capacity estimation results of each monomer, the ampere-hour integral required for balancing of each battery is calculated, and the balancing current is output. In this way, by establishing an MCDM model including CMCD and CDM, the dynamic behavior of the minimum-capacity monomer MCC (minimum-capacity cell, MCC) and other monomers in the battery pack can be accurately described, and at the same time, based on MCDM, the double extended Kalman filter DEKF is used to accurately estimate the SOC and capacity of the monomers in the battery pack at multiple time scales. It solves the problem in the prior art that it is difficult to accurately estimate the capacity, and it must be based on accurate SOC estimation.

[0058] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] 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 some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0060] FIG. 1( a ) is an algorithm framework diagram of a multi-time scale balancing strategy based on state of charge and capacity for a passive balancing circuit provided in an embodiment of the present application;

[0061] FIG1( b ) is an algorithm timing diagram of a multi-time scale balancing strategy based on state of charge and capacity for a passive balancing circuit provided in an embodiment of the present application;

[0062] Figure 2 A schematic diagram of the MCDM selected in the multi-time scale state of charge and capacity-based balancing strategy for a passive balancing circuit provided in the embodiment of the present application;

[0063] Figure 3 A schematic diagram of a balancing strategy in a multi-time scale balancing strategy based on state of charge and capacity for a passive balancing circuit provided in an embodiment of the present application;

[0064] Figure 4 A new battery pack SOC and capacity estimation diagram in the solution provided in the embodiment of the present application;

[0065] Figure 5 The SOC and capacity estimation diagram of the old battery pack in the solution provided in the embodiment of the present application;

[0066] Figure 6 A battery pack capacity comparison diagram based on voltage, SOC and SOC and capacity balancing strategy in the solution provided in the embodiment of the present application;

[0067] Figure 7 A comparison chart of energy consumption based on voltage, SOC and SOC and capacity balancing strategy in the solution provided in the embodiment of the present application;

[0068] Figure 8 A comparison chart of the first balancing time based on voltage, SOC and SOC and capacity balancing strategies in the solution provided in an embodiment of the present application. DETAILED DESCRIPTION

[0069] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0070] After lithium-ion battery cells are integrated into battery packs, the cycle life of the battery pack is much lower than the life of the single cell due to the differences in external use environment and internal characteristic parameters. In order to reduce the differences in battery packs, increase the capacity of battery packs, and extend the cycle life, balancing management is essential. Passive balancing is widely used in electric vehicles and energy storage systems due to its advantages of low cost, simple structure and easy control. Compared with the mature development of passive balancing hardware circuits, the development of high-efficiency balancing strategies is relatively backward.

[0071] In order to solve the above problems, the present application provides a multi-time scale balancing strategy based on state of charge and capacity for a passive balancing circuit, estimates the SOC and capacity of a single cell in a battery pack on multiple time scales, and provides an execution process based on the SOC and capacity balancing strategy. The specific implementation scheme is described in detail through the following embodiments.

[0072] Example:

[0073] FIG. 1 is a multi-time scale balancing strategy based on state of charge and capacity for a passive balancing circuit provided in an embodiment of the present application. As shown in FIG. 1( a ) and FIG. 1( b ), in the multi-time scale balancing strategy based on state of charge and capacity for a passive balancing circuit provided in an embodiment of the present application:

[0074] The algorithm collective structure mainly includes three parts: the first part is the high-precision and high-fidelity multi-cell model (MCM) part, which is used to simulate the dynamic, degradation and balancing behavior of the battery pack throughout the life cycle. The second part is the state of charge (SOC) and capacity estimation part, which estimates the minimum capacity cell SOC and parameters (denoted as: SOC EKF#min and PAREKF#min) based on the cell minimum capacity model (Cell Min Capacity Model, CMCM) using the dual extended Kalman filter (Dual Extend Kalman Filter, DEKF) (denoted as: CMCM DEKF), and loads the posterior SOC into the capacity estimation EKF (denoted as: CAP EKF#min) to obtain the minimum capacity cell capacity estimation. Based on CDM, DEKF (denoted as: CDM DEKF) is used to estimate the ΔSOC and Δ parameters (denoted as: ΔSOC EKF#i and ΔPAR EKF#i) of other cells (denoted as #i, i from #1 to #N-1, excluding #min, N iscell numbers), and the posterior ΔSOC and posterior SOC are loaded into the capacity estimation EKF (denoted as: CAP EKF#i) to obtain the capacity estimation of other cells. The third part is the balancing strategy part, which calculates the ampere-hour integral required for balancing each battery based on the posterior estimation of the SOC and capacity of each cell, and outputs the balancing current to the above parts (denoted as: EQU#min and EQU#i).

[0075] In practical applications, in order to reduce the computational complexity of the algorithm, this application designs multi-time scale execution timing for different parts, as shown in Figure 1(b). The "SCLK" main clock cycle is 1s, and all parts have execution cycles that are integer multiples of the main clock cycle. MCM requires a high-fidelity simulated battery pack, and the "Cell#min+Cell#i" execution cycle is synchronized with the main clock cycle for 1s. The SOC and capacity estimation part includes the CMCM and CDM parts. CMCM DEKF and capacity EKF are the basis of other single-cell EKFs, and the "CMCM DEKF+CAP EKF#min" execution cycle is synchronized with the main clock cycle for 1s. The difference SOC and difference parameters of CDM are slow variables, so the "CDM DEKF+CAP EKF#i" execution cycle is (N-1)s. The equalization part requires all single-cell SOC and capacity estimation results, and the "EQU#min+EQU#i" execution cycle is synchronized with the "CDM DEKF+CAP EKF#i" cycle for (N-1)s. Since capacity estimation takes some time to regress, no balancing is performed during the initial capacity regression process.

[0076] It should be noted that the traditional series battery pack MDM is composed of a CMM and N CDMs. The output voltage of the CMM numerically corresponds to the average value of all monomer voltages, and the average value needs to be calculated each time, which increases the computational complexity. In addition, the CMM has no corresponding physical entity and lacks a clear physical meaning. Drawing on the modeling idea of ​​MDM, MCDM is proposed in this application, and CMM is replaced by CMCM. This not only avoids the calculation of the average value, but also forms a mapping relationship between CMCM and the minimum capacity monomer, which has a clear physical meaning. In addition, the number of CDMs is reduced from N to N-1. In this application, four battery pack models based on Rint, First-order and Second-order are compared and studied. According to previous research results, CMCM is selected as the first-order RC model and CDM is selected as the Rint model. The MCDM selected in this application is as follows Figure 2 shown.

[0077] In some specific implementation processes, this application focuses on the construction of the CMCM model:

[0078] Since the first-order RC model can take into account accuracy, computational complexity and degradation adaptability, it is widely used in battery management algorithms. In this application, the first-order RC model is selected to load the CMCM in MCDM, such as Figure 2 (a), where Figure 2 (a) is MCDM, Figure 2 (b) is CDM. The CMCM state space equation considering the balanced current is written as follows:

[0079]

[0080] U T，k =U OCV，k (SOCk)-U P，k -R 0，k (Ik+I Equ，k-1 ). (2)

[0081] Where k is the current execution cycle, k-1 is the previous execution cycle, and Δt is the execution cycle duration. k is the CMC state of charge. η is the CMC coulombic efficiency. C k-1 is the capacity of CMC. k and I Equ，k-1 They are the current in the main circuit and the balancing current, positive for discharge and negative for charge. T，k Represents the CMC terminal voltage. P，k and τ k are the polarization internal resistance and time constant ofRC cirtuit respectively. P，k is the terminal voltage of the RC circuit.O，k U is the ohmic internal resistance. OC，k1 Represents OCV. The relationship between OCV and SOC can be fitted by a polynomial function, as follows:

[0082] U OCV =α1SOC 6 +α2SOC 5 +α3SOC 4 +α4SOC 3 +α5SOC 2 +α6SOC 1 +α7 (3)

[0083] Among them, α1~α7 are fitting coefficients.

[0084] For the construction of CDM model:

[0085] In the degradation process of the battery, the changes in OCV and ohmic internal resistance are dominant. In addition, in this application, the CDM based on the Rint model can take into account both accuracy and computational complexity. Therefore, the Rint model is selected in this application to establish the CDM, such as Figure 2 (b). The state space equation of CDM is written as follows:

[0086] ΔU T，i，k =ΔU OCV，i，k (ΔSOC i，k )-ΔR O，i，k (I k +I Equ，ik-1 ) (4)

[0087] Among them, ΔU T，i，k Represents the terminal voltage difference between the ith battery and the CMC (a positive sign indicates that it is greater than U T，k , the negative sign indicates that it is less than U T，k ). ΔU OCV，i，k Represents the OCV difference between the i-th battery and the CMC (a positive sign indicates that it is greater than U OCV，k , the negative sign indicates that it is less than U OCV，k ). ΔSOC i，k Represents the SOC difference between the i-th battery and the CMC (a positive sign indicates a value greater than SOC k , the negative sign indicates that it is less than SOC k ). ΔR O，i，k Represents the ohmic internal resistance difference between the i-th battery and the CMC (a positive sign indicates that it is greater than R O，k , the negative sign indicates that it is less than R O，k ).

[0088] After the above model is established, it is necessary to estimate the SOC and capacity of the battery pack based on the multi-time scale of the model movement. First, based on the CMCM model, perform SOC and parameter estimation:

[0089] State variable SOC in CMCM k and U P，k It not only depends on the previous short-term history, but also changes rapidly with the changes of the system control variables. In contrast, the parameters in CMCM change slowly as the battery ages, indicating that the system parameters are stable in a short period of time. When selecting an algorithm, it is necessary to consider the situation where the rate of change of the battery state and the parameters is very different, and it is best to transfer and correct them separately. Therefore, the selection of the DEKF algorithm provides an effective method for establishing a more independent system parameter and state estimation path.

[0090] The process of DEKF is as follows Figure 2 As shown, is the prior estimate of the system’s state vector; is the prior estimate of the system's parameter vector, and the two are updated alternately and iteratively. Table 1 provides the details of the DEKF algorithm, which consists of two carefully integrated EKFs. First, based on the selected CMCM model, the state equation, parameter equation and measurement equation are established using the DEKF method. This application gives the state equation and state measurement equation of the CMCM in the above process. Here, the state matrix x k =[SOC k U P，k ], input matrix u k =I k-1 +I Equ，k-1 , the output matrix y k =U T，k , and increase the real system ω k and v k Noise input. For parametric equations, the parameter matrix θ k =[R O，k R P，k τ k ]. It is assumed that the values ​​of the parameters do not change between two sampling points, but they may change slowly over time due to a small fictitious process noise r k Modeling. The parameter measurement equation is the CMCM output estimate plus the estimation error e k The nonlinear state space equations of the joint system will be composed of equations (5)-(7).

[0091]

[0092]

[0093] U T，k =UOC，k -U P，k -R O，k (I k +I Equ，k-1 )+v k / e k (7)

[0094] During the calculation process, the Jacobian ratio matrix defined in Table 1 is written as formulas (8)-(12), where Δt is each execution cycle.

[0095]

[0096]

[0097]

[0098]

[0099]

[0100] Table 1 CMCM SOC and parameter estimation based on dual extended Kalman filter

[0101]

[0102]

[0103] Furthermore, based on the results obtained in the above steps, capacity estimation is performed:

[0104] Although capacity can also be used as a battery parameter to be estimated in DEKF, as an important control variable of the balancing strategy, the accuracy, stability and reliability of capacity are crucial. Therefore, in this application, capacity is separated from other parameters and estimated separately. Drawing on the parameter estimation method in DEKF, capacity can also be estimated as a slowly varying parameter. In the parameter equation, the parameter matrix θ C，k =[C k ]. In the measurement equation, the input matrix Output matrix y C，k =ηΔt(I k-1 +I Equ，k-1 ). The nonlinear state space equations for capacity estimation and the corresponding Jacobian ratio matrix will be composed of equations (13)-(15).

[0105] [C k ]=[C k-1 |+r C，k (13)

[0106]

[0107]

[0108] Then, SOC and capacity estimation are performed based on CDM, including SOC and parameter estimation and capacity estimation.

[0109] For SOC and parameter estimation: Similar to the SOC and parameter estimation based on CMCM, DEKF is also selected to estimate the state and parameters of CDM. Table 2 provides the details of the DEKF algorithm based on CDM. First, the state equation, parameter equation and measurement equation need to be determined. In the state equation, the state matrix is ​​Δx i，k =[ΔSOC i，k ]. Similar to the parameter estimation principle, it is assumed that the SOC difference between cell #i and CMC remains constant during the two execution cycles, but changes slowly over time, represented by a small virtual noise ω i，k To drive. In the above process of this application, the state measurement equation of CDM is given, and the matrix u is input here. i，k =I k +I Equ，i，k-1 , the output matrix Δy i，k =ΔU T，i，k , and increase the real system v i，k Noise input. For parametric equations, the parameter matrix Δθ i，k =|ΔR O，i，k |. It is assumed that the values ​​of the parameters do not change between two sampling points, but they may change slowly over time due to a small fictitious process noise r i，k Modeling. The parameter measurement equation is the CMCM output estimate plus the estimation error e i，k The nonlinear state space equations of the joint system will be composed of equations (16)-(18).

[0110]

[0111]

[0112] ΔU T，i，k =ΔU OCV，i，k -ΔR O，i，k (I k +I Equ，i，k-1 )+u i，k / e i，k (18)

[0113] During the calculation process, the Accord ratio matrix defined in Table 2 is written as shown in formulas (19)-(22).

[0114]

[0115]

[0116]

[0117]

[0118] Table 2 CDM SOC and parameter estimation based on dual extended Kalman filter

[0119]

[0120]

[0121] For capacity estimation:

[0122] Based on CMCM, capacity estimation is carried out. In CDM, capacity can also be estimated as a slowly varying parameter. In the parameter equation, the parameter matrix In the measurement equation, the input matrix Can be decomposed into Output matrix y C，i，k =ηΔt(I k-1 +I Equ，i，k-1 ). The nonlinear state space equations for capacity estimation and the corresponding Jacobian ratio matrix will be composed of equations (23)-(26).

[0123]

[0124]

[0125]

[0126]

[0127] Furthermore, based on the accurate estimation of the SOC and capacity of each cell, the present application further designs a balancing strategy based on SOC and capacity. The present application selects a passive balancing circuit, and the following discussion is based on the premise of passive balancing. For example, three cells with different capacities and different initial SOCs are selected as an example to illustrate this process. The three cells are numbered #1 to #3, and the total capacity is 90Ah, 95Ah, and 100Ah, respectively. Figure 3 As shown in Figure 3, where Figure 3 is a balancing strategy based on SOC and capacity. Figure 3 (a) shows the balance before C T，Pack，k <C T，min，k , Figure 3 (b) After equilibrium, C R，min，k =min(C R，i，k )&C C，min，k =min(C C，i，k ), Figure 3 (c) After equilibrium C C，1，k =C C，min，k=... =C C，N，k Before balancing, the initial SOC of the three batteries were 88.89%, 78.95%, and 95%, respectively. According to formulas (27)-(28), the releasable capacity C of the single cell is calculated. R，i，k They are 80Ah, 75Ah, 95Ah respectively, and the single-cell rechargeable capacity is C C，i，k They are 10Ah, 20Ah, and 5Ah respectively. According to formulas (29)-(31), the battery pack can release capacity C R，Pack，k , battery pack rechargeable capacity C C，Pack，k And the total battery capacity C T，Pack，k It is divided into 75Ah, 5Ah and 80Ah, such as Figure 3 (a) Under passive balancing conditions, the battery pack capacity is less than the CMC capacity, indicating that the battery pack is not in a balanced state and needs further balancing.

[0128] C R，i，k =C T，i，k *SOC (27)

[0129] C C，i，k =C T，i，k *(1-SOC) (28)

[0130] C R，Pack，k =min(C R，i，k ) (29)

[0131] C C，Pack，k =min(C C，i，k ) (30)

[0132] C T，Pack，k =C R，Pack，k +C C，Pack，k (31)

[0133] Under passive balancing conditions, the battery pack capacity is equal to the CMC capacity when the equilibrium state is reached. This requires that two conditions are met at the same time: 1. The CMC releasable capacity is equal to the minimum releasable capacity of all monomers; 2. The CMC rechargeable capacity is equal to the minimum rechargeable capacity of all monomers, as shown in formula (32). Further, it is necessary to calculate the balancing capacity required for each monomer. According to condition 1, the calculation formula for the releasable balancing capacity of each monomer is shown in (33). According to condition 2, the calculation formula for the rechargeable balancing capacity of each monomer is shown in (34)-(35). The total balancing capacity of each monomer is the sum of the releasable balancing capacity and the rechargeable balancing capacity, and the calculation formula is shown in (36). Taking 3 sample batteries as an example, the releasable balancing capacity C of each monomer is Equ，R，i，k They are 5Ah, 0Ah, and 0Ah respectively. Each monomer can release a balanced capacity C Equ，C，i，k They are 0Ah, 0Ah, 10Ah respectively, and the total balanced capacity is CEqu，i，k They are 5Ah, 0Ah, and 10Ah respectively. Figure 3 As shown in (b), the battery pack releasable capacity of 75Ah is equal to the CMC releasable capacity of 75Ah, the battery pack rechargeable capacity of 15Ah is equal to the CMC rechargeable capacity of 15Ah, and the battery pack total capacity of 90Ah is equal to the CMC total capacity of 90Ah. The balanced state of the battery pack under passive balancing conditions is achieved.

[0134] C R，min，k =min(C R，i，k )&C C，min，k =min(C C，i，k ) (32)

[0135]

[0136] C C，i，k =C C，i，k-1 +C Equ，R，i，k (34)

[0137]

[0138] C Equ，i，k =C Equ，R，i，k +C Equ，C，i，k (36)

[0139] from Figure 3 (b) It can be seen that the releasable capacity and rechargeable capacity of each cell in the battery pack are not consistent. This is a balancing strategy with wide constraints, and the calculation process is relatively complex. However, in engineering practice, a special case of the above balancing strategy is usually adopted, that is, the releasable capacity or rechargeable capacity is consistent. Taking the consistency of rechargeable capacity as an example, the process of releasable capacity consistency is similar and will not be repeated. The judgment condition for the consistency of the rechargeable capacity of each cell is shown in (37), and the balancing capacity calculation formula is shown in (38). Similarly, taking the sample in Section 3 as an example, the rechargeable capacity C C，i，k They are 10Ah, 20Ah, 5Ah, and balanced capacity C Equ，i，k They are 10Ah, 0Ah, 15Ah respectively. Figure 3 Although the power consumption of the first balancing is larger than that of the previous balancing strategy, the computational complexity is reduced, so the latter is adopted in the solution of this application.

[0140] C C，1，k =C C，min，k =... =C C，N，k (37)

[0141] C Equ，i，k =max(C C，i，k )-C C，i，k (38)

[0142] The present application provides a passive balancing strategy for battery pack SOC and capacity estimation based on multiple time scales proposed in this article. First, a single cell minimum capacity-difference model is established, which includes a single cell minimum capacity model and a single cell difference model, which can accurately describe the dynamic behavior of the minimum capacity single cell and all other single cells in the battery pack. Secondly, based on MCDM and using a double extended Kalman filter, the single cell SOC and capacity in the battery pack are estimated on multiple time scales, and the execution process of the SOC and capacity balancing strategy is given. Taking into account the indicators of maximizing the battery pack capacity, minimizing the balancing energy consumption, and minimizing the first balancing time, high-efficiency balancing over the entire life cycle is achieved.

[0143] The solution provided in this application will be verified by means of simulation experiments. There are huge challenges in experimental verification of the full life cycle of the battery pack, including: a large number of cells, long experimental time and large experimental data, which are not conducive to verification at the principle stage. In addition, it is difficult to obtain accurate SOC and capacity of the cells in experimental verification, which will lead to the lack of benchmark values ​​for the estimation algorithm. Therefore, the proposed estimation algorithm and balancing strategy will be verified by means of simulation.

[0144] First, the experimental design is carried out: In order to simulate the dynamic behavior of the battery pack over its entire life cycle with high fidelity, we use a first-order RC-based MCM. In the simulation conditions, the dynamic stress test (DST) is used as the discharge condition, and the battery pack is discharged until the lowest voltage of the single cell reaches the lower cut-off voltage or the lowest SOC reaches 0%. The 1C constant current charge is used as the charging condition, and the battery pack is charged until the highest voltage of the single cell reaches the upper cut-off voltage or the highest SOC reaches 100%. In order to ensure the fidelity of the simulation, the setting of the model parameters is guided by the parameters of the degradation process of the real battery pack over its entire life cycle. The battery pack consists of four cells connected in series, with a lithium (Ni0.8Co0.1Mn0.1) oxygen cathode and a graphite anode. The rated capacity and voltage of each cell are 177Ah and 3.61V, respectively. The upper and lower cut-off voltages are 4.2 and 2.8V, respectively. The battery pack undergoes about 700 cycles, and the capacity of the CMC reaches the end-of-life condition, that is, 80% of the initial capacity. Under the guidance of the real battery pack parameters, the simulation MCM parameters are set as follows:

[0145] Initial SOC: Normal random given, mean 93.5%, variance 1.3%; average value: 93.5%, maximum value 95%, minimum value 92%, difference: 3.16%.

[0146] Initial capacity C T，i，1: Normal random given, mean 177Ah, variance 2Ah; average value: 176.923Ah, maximum value 180.186Ah, minimum value 173.843Ah, difference: 3.52%.

[0147] Capacity degradation rate C Loss，i，k The capacity attenuation model in this paper adopts the empirical model proposed by Wang et al.

[19] , and its formula is:

[0148] C Loss =k C Ah 0.55 (39)

[0149] Among them, k c is the capacity attenuation coefficient, which follows a normal distribution in the simulation experiment. The battery pack undergoes 700 cycles, and the capacity loss is about 20%, corresponding to a mean attenuation coefficient of 0.0413. Capacity attenuation is usually not screenable, so the variance is given as 0.005. The maximum k c is 0.0448, the minimum k c The value is 0.0378, and the mean is 0.0412, which is about 15.55% difference.

[0150] Initial ohmic resistance R O，i，1 :The mean of the initial ohmic resistance of the real battery pack is 0.349mΩ, and the variance is 0.005mΩ. In the simulation experiment, a normal distribution is given, and the mean and variance are the same as the real value. The initial ohmic resistance has a maximum value of 0.354mΩ and a minimum value of 0.336, with a difference of 5.07%.

[0151] Ohmic resistance degradation rate R Loss，i，k :The ohmic internal resistance growth model in this paper adopts the linear empirical model proposed by Matsushima et al., and its formula is:

[0152] R Loss =k R Ah (40)

[0153] Among them, k s is the ohmic internal resistance growth coefficient. In the simulation experiment, the ohmic internal resistance growth coefficient is related to the capacity attenuation coefficient. The maximum k R is 0.0448, the minimum k R The value is 0.0378, and the mean is 0.0412, which is about 15.55% difference.

[0154] Polarization resistance R P，i : The polarization resistance simulation experiment of 4 single cells is the same as the initial true value and remains unchanged throughout the life cycle.

[0155] Time constant τ i: The time constant simulation experiment of 4 single-cell batteries is the same as the initial true value, and remains unchanged throughout the life cycle.

[0156] Coulomb efficiency η i :The true value of the coulombic efficiency of 4-cell battery is very close to 1. In the simulation experiment, a normal distribution is given with a mean of 0.9995 and a variance of 0.0001. The simulated coulombic efficiency of 4 cells has a maximum value of 0.99967% and a minimum value of 0.99930%, with a difference of 0.04%, which remains unchanged throughout the life cycle.

[0157] Balancing current: The balancing current of the four cells is 100 mA and remains unchanged throughout the life cycle. The battery pack simulation model parameters are shown in Table 3.

[0158] Table 3 Battery pack simulation model parameters

[0159]

[0160]

[0161] It should be noted that the experimental outline includes two parts: battery pack SOC and capacity estimation and balancing performance. Part 1 Battery pack SOC and capacity estimation uses the SOC and capacity simulation data of the new battery pack in the first cycle and the old battery pack in the 672nd cycle (the second to last cycle, the last cycle is incomplete) as reference values ​​to verify the multi-time scale battery pack SOC and capacity estimation algorithm.

[0162] Part 2 Balancing Performance uses the full life cycle battery pack capacity, balancing energy consumption and first balancing time performance indicators to conduct a comprehensive comparative analysis of three balancing strategies based on voltage, based on SOC, and based on SOC and capacity. In order to make each balancing strategy achieve the consistency of balancing variables as much as possible, the balancing cutoff threshold is set as small as possible to avoid the degradation of balancing performance caused by it. The complete balancing performance test outline is shown in Table 4, where Figure 4 SOC and capacity estimation for new battery pack. Figure 4 (a) is the new monomer SOC estimation curve, Figure 4 (b) is the new monomer SOC estimation error curve, Figure 4 (c) is the new monomer capacity estimation curve, Figure 4 (d) is the error curve of new monomer capacity estimation.

[0163] Table 4 Balanced performance test outline and comparative performance

[0164]

[0165] After completing the simulation experiment design, obtain and analyze the battery pack SOC and capacity estimation results.

[0166] Specifically, the SOC and capacity estimation results of the new battery pack in the first cycle are as follows: Figure 4 As shown, the estimated results of the old battery pack at the 672nd cycle are as follows Figure 5 shown. Figure 4 (a), (b), Figure 5 (a) and (b) are the single-cell SOC estimation curve and error curve, respectively. Figure 4 (c), (d), Figure 5 (c) and (d) are the estimated and error curves of the cell capacity, respectively. In this paper, "Cell#1~#4Real" is the real value generated by MCM, "Cell#1~#4Estimation" is the estimated value of the SOC and capacity of the multi-time scale battery pack, and "Cell#1~4Error" is the difference between the real value and the estimated value. The RMSE of the estimated SOC and capacity of the new battery pack and the old battery pack are shown in Table 5.

[0167] like Figure 4 As shown in Figure 2, the estimated SOC and capacity of the new battery pack are given. Figure 4 (a) shows the single-cell SOC estimation result. It can be seen that the multi-time scale estimation method can provide accurate SOC estimation over the entire SOC range. From the sub-figure, it can be seen that the algorithm has a fast regression speed, and the algorithm regresses to the true value within 10 seconds. Figure 4 (b) shows the single-cell SOC estimation error curve. From the results, it can be seen that the method proposed in this paper can keep the SOC estimation error within 0.015%. Figure 4 (c) The cell capacity estimation result shows that the proposed method can regress to the true value around 4000s and has been closely following the true value since then. Figure 4 (d) shows the single cell capacity estimation error curve. From the results, it can be seen that the method proposed in this paper can keep the capacity estimation error within 6Ah (3.39% Full Seale).

[0168] Figure 5 SOC and capacity estimation for old battery packs. Figure 5 (a) is the old cell SOC estimation curve; (b) is the old cell SOC estimation error curve; (c) is the old cell capacity estimation curve; (d) is the old cell capacity estimation error curve. Figure 5 As shown in Figure 2, the estimated SOC and capacity of the old battery pack are given. Figure 5 (a) shows the single-cell SOC estimation result. It can be seen that the proposed method can also provide accurate SOC estimation in the full SOC range and can regress to the true value within 10 s. Figure 5(b) shows the single cell SOC estimation error curve. From the results, it can be seen that the method proposed in this paper can also keep the SOC estimation error within 0.015%. It is similar to the estimation result of the new battery pack. Figure 5 (c) shows the cell capacity estimation result. It can be seen that the proposed method can also regress to the true value around 4000s, and has been closely following the true value since then. Figure 5 (d) shows the single cell capacity estimation error curve. From the results, it can be seen that the method proposed in this paper can keep the capacity estimation error within 4Ah (2.84% Full Scale). This result is more accurate than the estimation result of the new battery pack.

[0169] Table 5 lists the RMSE values ​​of the single-cell SOC and capacity estimation for the new and old battery packs. The maximum RMSE of the single-cell SOC estimation for the new battery pack is 2.3775e-6%, and the average RMSE is 2.1924e-6%. The maximum RMSE of the single-cell SOC estimation for the old battery pack is 2.7529e-6%, and the average RMSE is 2.3547e-6%. The estimation accuracy of both is very high and the difference is almost the same. The maximum RMSE of the single-cell capacity estimation for the new battery pack is 0.1617Ah, and the average RMSE is 0.1471Ah. The maximum RMSE of the single-cell capacity estimation for the old battery pack is 0.1099Ah, and the average RMSE is 0.1059Ah. Both estimation accuracies are very high, and the old battery pack is slightly higher than the new battery pack, which is mainly due to the smaller capacity difference of the old battery pack. Therefore, we can draw the following conclusions: the method proposed in this paper has high accuracy on both new and old battery packs, and it can accurately and effectively estimate the SOC and capacity of each single cell throughout the life cycle.

[0170] Table 5 RMSE of new and aged battery cell SOC(%)and capacity(Ah)estimations.

[0171]

[0172] Next, the balanced performance movement analysis includes:

[0173] For battery capacity: The comparison results of battery pack capacity over the whole life cycle based on voltage, SOC and SOC and capacity balancing strategies are as follows: Figure 6 shown. Figure 6 (a) is the battery pack capacity based on voltage balancing strategy. Figure 6 (b) is the battery pack capacity based on SOC balancing strategy. Figure 6 (c) is the battery pack capacity based on the balancing strategy of SOC and capacity. Figure 6(d) is a comparison of the battery pack capacities of three balancing strategies. In this article, "Cell#1~#4" is the real value of the single cell capacity generated by MCM, "PackTheory" is the theoretical value of the battery pack capacity under passive balancing, which is numerically equal to the minimum single cell capacity, "Pack Real" is the real capacity obtained after balancing by the three balancing strategies, and "Cycle" is the number of battery pack cycles.

[0174] from Figure 6 From the battery pack capacity results in (a), (b), and (c), it can be seen that the three balancing strategies based on voltage, based on SOC, and based on SOC and capacity can quickly return to the theoretical value of the battery pack capacity and closely follow the theoretical value throughout the life cycle. Figure 6 (d) shows the battery pack capacity comparison result. It can be seen that the theoretical and actual values ​​of the battery packs of the three balancing strategies have reached 80% of the nominal capacity of 177Ah, that is, 141.6Ah. However, the number of cycles of the three balancing strategies in the full life cycle are 670, 665 and 673 respectively. The number of cycles of the balancing strategy based on voltage and SOC is less than that of the balancing strategy based on SOC and capacity. This is mainly due to the large balancing energy consumption of the first two balancing strategies. The discharge amount is more than that of the balancing strategy based on SOC and capacity in the full life cycle, resulting in accelerated degradation of the battery pack, which will be analyzed in detail below.

[0175] For balanced energy consumption: The comparison results of balanced energy consumption over the whole life cycle of voltage-based, SOC-based, and SOC and capacity-based balanced strategies are as follows: Figure 7 shown. Figure 7 (a) The voltage-based balancing strategy balances energy consumption. Figure 7 (b) The energy consumption is balanced based on the SOC balancing strategy. Figure 7 (c) The energy consumption is balanced by the balancing strategy based on SOC and capacity. Figure 7 (d) is the comparison of balanced energy consumption of three balanced strategies. In this paper, "Cell#1~#4" is the balanced energy consumption of each cell, and "Sum" is the sum of the balanced energy consumption of all cells.

[0176] from Figure 6 The balanced energy consumption results of (a) and (b) show that all cells are constantly balanced in the whole life cycle of the two balancing strategies based on voltage and SOC. Figure 6 (c) shows the balanced energy consumption result. It can be seen that the balanced strategy based on SOC and capacity will no longer balance after Cell#4 reaches the balanced cutoff threshold condition. Figure 6From the comparison results of balanced energy consumption in (d), it can be seen that no matter it is the balanced energy consumption of a single cell or the balanced energy consumption, the balanced energy consumption based on SOC is the largest, with a total energy consumption of 3435.32Wh. The balanced energy consumption based on voltage is second, with a total energy consumption of 1055.41Wh. The balanced energy consumption based on SOC and capacity is the smallest, with a total energy consumption of 381.64Wh. In addition, the balanced energy consumption of each single cell is related to the coulombic efficiency. Through the comparative analysis of balanced energy consumption, it can be seen that the voltage-based and SOC-based balancing strategies often have the problem of over-balancing. However, the balancing strategy based on SOC and capacity can effectively avoid the problem of over-balancing.

[0177] For the first balancing time: the comparison results of the first balancing time of the balancing strategies based on voltage, based on SOC, and based on SOC and capacity are as follows Figure 8 shown. Figure 8 (a) is the first balancing time of the voltage-based balancing strategy. Figure 8 (b) is the first equilibrium time of the SOC-based equilibrium strategy. Figure 8 (c) is the first balancing time of the balancing strategy based on SOC and capacity. Figure 8 (d) is the comparison of the first balancing time of the three balancing strategies. In this paper, "Cell#1~#4" is the real value of the single cell capacity generated by MCM, "Pack Theory" is the theoretical value of the battery pack capacity under passive balancing, which is numerically equal to the minimum single cell capacity, "Pack Real" is the real capacity obtained after balancing by the three balancing strategies, "Time" is the first balancing time, and "Cycle" is the number of battery pack cycles.

[0178] from Figure 8 The first balancing time results of (a)(b)(c) show that the three balancing strategies based on voltage, based on SOC, and based on SOC and capacity can gradually return to the theoretical value of the battery pack capacity. Figure 8 From the comparison results of the first balancing time in (d), it can be seen that the first balancing time based on voltage is the longest, which is 192856s and the number of cycles is 6. The first balancing time based on SOC is second, which is 118890s and the number of cycles is 3. The first balancing time based on SOC and capacity is the shortest, which is 114022s and the number of cycles is 3. Through the comparative analysis of the first balancing time, it can be seen that the voltage-based balancing strategy will often have balancing gaps, which is mainly because the single cell voltage difference cannot trigger the balancing threshold under low current conditions. However, the balancing strategy based on SOC and SOC and capacity can accurately express the inconsistency and effectively avoid the problem of balancing gaps.

[0179] In summary, the comparison of the balancing performance of the voltage, SOC and SOC and capacity balancing strategies is summarized and listed as shown in Table 6. The actual capacity of the battery pack based on the voltage balancing strategy is close to the theoretical capacity, and the number of cycles is at a medium level; the single cell and total energy consumption are also at a medium level; the first balancing time and the number of cycles are at a low level. The actual capacity of the battery pack based on the SOC balancing strategy is consistent with the theoretical capacity, and the number of cycles is at a low level; the single cell and total energy consumption are also at a low level; the first balancing time and the number of cycles are at a high level. The actual capacity of the battery pack based on the SOC and capacity balancing strategy is consistent with the theoretical capacity, and the number of cycles is at a high level; the single cell and total energy consumption are at a high level; the first balancing time and the number of cycles are also at a high level. We can draw the following conclusions: The traditional voltage and SOC-based balancing strategy cannot take into account the battery pack capacity, balanced energy consumption and first balancing time performance at the same time. However, the SOC and capacity-based balancing strategy can take into account the battery pack capacity maximization, balanced energy consumption and first balancing time minimization performance, and realize a high-efficiency balancing strategy for the entire life cycle.

[0180] The requirements for computing resources for the three balancing strategies based on voltage, SOC, and SOC and capacity are from low to high. If it is for some scenarios that are sensitive to computing resources and costs, it is recommended to adopt a voltage-based balancing strategy, which can take into account both cost and some performance. In addition, you can also consider adopting a balancing strategy based on a combination of voltage and SOC. When the consistency of the battery pack is poor, use the SOC-based balancing strategy to quickly restore the battery pack capacity; when the consistency of the battery pack is good, use the voltage-based balancing strategy to reduce the balancing energy consumption. The above scenarios have high requirements for the heat dissipation performance of the balancing hardware circuit. If it is not sensitive to computing resources and costs, and focuses on balancing performance, it is recommended to adopt a balancing strategy based on SOC and capacity. This scenario has low requirements for the heat dissipation performance of the balancing hardware circuit.

[0181] Table 6 Comparison of balancing performance based on voltage, SOC and SOC and capacity balancing strategies

[0182]

[0183] In view of the problem of increasing parameter differences during the use of battery packs, this application proposes a passive balancing strategy for battery pack SOC and capacity estimation based on multiple time scales. First, an MCDM was established, which includes CMCD and CDM, and can accurately describe the dynamic behavior of the smallest capacity cell and other cells in the battery pack. Secondly, based on MCDM, the double extended Kalman filter DEKF is used to estimate the SOC and capacity of the cells in the battery pack on multiple time scales. Based on this, a balancing strategy based on SOC and capacity is proposed. Finally, using high-fidelity MCM, under the guidance of actual battery pack degradation data, the proposed balancing strategy is simulated and verified over the entire life cycle. Experimental results show that the proposed method can efficiently and accurately estimate the single cell SOC and capacity on new and old battery packs, and achieve high-precision estimation over the entire life cycle. Compared with the traditional balancing strategy based on working voltage and SOC, the proposed balancing strategy can take into account the indicators of maximizing battery pack capacity, minimizing balancing energy consumption, and minimizing balancing time, and achieve high-efficiency balancing over the entire life cycle.

[0184] It can be understood that the same or similar parts of the above embodiments can be referenced to each other, and the contents not described in detail in some embodiments can refer to the same or similar contents in other embodiments.

[0185] It should be noted that, in the description of this application, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, the meaning of "plurality" refers to at least two.

[0186] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0187] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0188] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0189] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0190] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0191] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0192] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A multi-time scale balancing strategy based on state of charge and capacity, characterized in that: include: Establish a monomer minimum capacity-difference model; The monomer minimum capacity difference model includes a monomer minimum capacity model and a monomer difference model; Based on the cell minimum capacity-difference model, the state of charge and capacity of each cell in the battery pack are estimated at multiple time scales through a double extended Kalman filter; Based on the state of charge and capacity estimation results of each cell, the ampere-hour integral required for balancing each battery is calculated, and the balancing current is output; The establishing of the monomer minimum capacity-difference model comprises: establishing the monomer minimum capacity model based on the first-order RC model; establishing the monomer difference model based on the Rint model; and obtaining the monomer minimum capacity-difference model based on the monomer minimum capacity model and the monomer difference model; The state space equation of the monomer minimum capacity model established based on the first-order RC model is: Among them, k is the current execution cycle, k-1 is the previous execution cycle, Δt is the execution cycle duration, SOC k is the MCC state of charge, η is the MCC coulombic efficiency, C k-1 is the capacity of MCC, I k and I Equ，k-1 are the current in the main circuit and the balancing current, respectively. The positive sign indicates discharge, and the negative sign indicates charge. U T，k represents the MCC terminal voltage, R P，k and τ k are polarization internal resistance and time constant, U P，k is the terminal voltage of the RC circuit, R O，k is the ohmic internal resistance, U OCV，k Represents OCV. The relationship between OCV and SOC is fitted by a polynomial function, specifically: U OCV =α1SOC 6 +α2SOC 5 +α3SOC 4 +α4SOC 3 +α5SOC 2 +α6SOC 1 +α7; Among them, α1~α7 are fitting coefficients, and MCC represents the minimum capacity monomer.

2. The multi-time scale balancing strategy based on state of charge and capacity according to claim 1, characterized in that: The method estimates the state of charge and capacity of each cell in the battery pack at multiple time scales based on the cell minimum capacity-difference model through a double extended Kalman filter, including: Based on the single-cell minimum capacity model, the state of charge and parameters of the MCC are estimated by a double extended Kalman filter, and based on the state of charge estimation result, the capacity of the MCC is estimated by an extended Kalman filter; wherein the MCC represents a single-cell with minimum capacity; Based on the monomer difference model, the ΔSOC and Δ parameter of other monomers are estimated through a double extended Kalman filter, and the ΔSOC of the other monomers and the SOC of the MCC are jointly loaded into the capacity estimation EKF to obtain the remaining monomer capacity estimation.

3. The multi-time scale balancing strategy based on state of charge and capacity according to claim 1, characterized in that: The method of calculating the ampere-hour integral required for balancing of each battery cell based on the charge state and capacity estimation result of each battery cell and outputting the balancing current includes: When the equilibrium state is not reached, the balancing current is continuously outputted until the equilibrium state is reached; wherein the equilibrium state is that under passive balancing conditions, the battery pack capacity is equal to the MCC capacity; wherein the MCC is the minimum capacity cell.

4. The multi-time scale balancing strategy based on state of charge and capacity according to claim 3, characterized in that: The battery pack capacity is equal to the MCC capacity: The releasable capacity of the MCC is equal to the minimum releasable capacity of all monomers, and the rechargeable capacity of the MCC is equal to the minimum rechargeable capacity of all monomers; Or the releasable capacity of the battery pack or the rechargeable capacity of the battery pack is consistent.

5. The multi-time scale balancing strategy based on state of charge and capacity according to claim 1, characterized in that: The state space equation of the monomer difference model established based on the Rint model is: D.U. T，i，k =ΔU OCV，i，k (ΔSOC i，k )-ΔR O，i，k (I k +I Equ，i，k-1 ); Among them, ΔU T，i，k Represents the terminal voltage difference between the i-th battery and the MCC, and the positive sign indicates that it is greater than U T，k , the negative sign indicates that it is less than U T，k , ΔU OCV，i，k Represents the OCV difference between the i-th battery and the MCC, and a positive sign indicates that it is greater than U OCV，k , the negative sign indicates that it is less than U OCV，k , ΔSOC i，k Represents the SOC difference between the i-th battery and the MCC, and a positive sign indicates that it is greater than the SOC k , the negative sign indicates that it is less than SOC k , ΔR O，i，k Represents the ohmic internal resistance difference between the i-th battery and the MCC, and the positive sign indicates that it is greater than R O，k , the negative sign indicates that it is less than R O，k ; Among them, MCC represents the minimum capacity monomer.

6. The multi-time scale balancing strategy based on state of charge and capacity according to claim 2, characterized in that: Based on the monomer minimum capacity model, the charge state and parameters of the MCC are estimated by using a double extended Kalman filter. The nonlinear state space equation of the joint system is: U T，k =U OCV，k -U P，k -R O,k (I k +I Equ，k-1 )+v k / e k ; The nonlinear state-space equation for capacity estimation is as follows: [C k ]=[C k-1 ]+r C，k ; Among them, k is the current execution cycle, k-1 is the previous execution cycle, Δt is the execution cycle duration, SOC k is the MCC state of charge, η is the MCC coulombic efficiency, C k-1 is the capacity of MCC, I k and I Equ，k-1 are the current in the main circuit and the balancing current, respectively. The positive sign indicates discharge, and the negative sign indicates charge. U T，k represents the MCC terminal voltage, R P，k and τ k are polarization internal resistance and time constant, U P，k is the terminal voltage of the RC circuit, R O，k is the ohmic internal resistance, U OCV，k Represents OCV. The relationship between OCV and SOC can be fitted by a polynomial function, specifically: U OCV =α1SOC 6 +α2SOC 5 +α3SOC 4 +α4SOC 3 +α5SOC 2 +α6SOC 1 +α7; Among them, α1~α7 are fitting coefficients.

7. The multi-time scale balancing strategy based on state of charge and capacity according to claim 2, characterized in that: Based on the monomer difference model, the ΔSOC and Δ parameters of other monomers are estimated through the double extended Kalman filter. The nonlinear state space equation of the joint system is: ΔU T，i，k =ΔU OCV，i，k -ΔR O，i，k (I k +I Equ，i，k-1 )+v i，k / e i，k ; The nonlinear state space equation for capacity estimation is: Among them, k is the current execution cycle, k-1 is the previous execution cycle, Δt is the execution cycle duration, SOC k is the MCC state of charge, η is the MCC coulombic efficiency, C k-1 is the capacity of MCC, I k and I Equ，k-1 are the current in the main circuit and the balancing current, respectively. The positive sign indicates discharge, and the negative sign indicates charge. U T，k represents the MCC terminal voltage, R P，k and τ k are polarization internal resistance and time constant, U P，k is the terminal voltage of the RC circuit, R O，k is the ohmic internal resistance, U OCV，k Represents OCV. The relationship between OCV and SOC can be fitted by a polynomial function, specifically: U OCV =α1SOC 6 +α2SOC 5 +α3SOC 4 +α4SOC 3 +α5SOC 2 +α6SOC 1 +α7; Among them, α1~α7 are fitting coefficients.

8. The multi-time scale balancing strategy based on state of charge and capacity according to claim 4, characterized in that: When the equilibrium condition is that the releasable capacity of the MCC is equal to the minimum releasable capacity of all monomers, and the chargeable capacity of the MCC is equal to the minimum chargeable capacity of all monomers, the calculation formula of the equilibrium state is: C R，min，k =min(C R，i，k )&C C，min，k =min(C C，i，k ); C C，i，k =C C，i，k-1 +C Equ，R，i，k ; C Equ，i，k =C Equ，R，i，k +C Equ，C，i，k ; When the balancing condition is that the releasable capacity of the battery pack or the rechargeable capacity of the battery pack is consistent, the calculation formula is: C C，1，k =C C，min，k =…=C C，N，k ; C Equ，i，k =max(C C，i，k )-C C，i，k ; Among them, C R，i，k is the monomer releasable capacity, C C，i，k is the single-cell rechargeable capacity, C Equ，R，i，k is the release equilibrium capacity of the monomer, C Equ，C，i，u is the release equilibrium capacity of the monomer, C Equ，i，k is the total balanced capacity.

Images