CN110850299A - SR-UKF-based lithium ion power battery state estimation method - Google Patents

SR-UKF-based lithium ion power battery state estimation method Download PDF

Info

Publication number
CN110850299A
CN110850299A CN201911173741.9A CN201911173741A CN110850299A CN 110850299 A CN110850299 A CN 110850299A CN 201911173741 A CN201911173741 A CN 201911173741A CN 110850299 A CN110850299 A CN 110850299A
Authority
CN
China
Prior art keywords
state
power battery
ion power
lithium ion
charge
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN201911173741.9A
Other languages
Chinese (zh)
Other versions
CN110850299B (en
Inventor
章军辉
祝婉凡
陈大鹏
李庆
庄宝森
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuxi Internet Of Things Innovation Center Co ltd
Original Assignee
Wuxi Internet Of Things Innovation Center Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuxi Internet Of Things Innovation Center Co ltd filed Critical Wuxi Internet Of Things Innovation Center Co ltd
Priority to CN201911173741.9A priority Critical patent/CN110850299B/en
Publication of CN110850299A publication Critical patent/CN110850299A/en
Application granted granted Critical
Publication of CN110850299B publication Critical patent/CN110850299B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/367—Software therefor, e.g. for battery testing using modelling or look-up tables
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/385—Arrangements for measuring battery or accumulator variables
    • G01R31/387—Determining ampere-hour charge capacity or SoC
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/36—Arrangements for testing, measuring or monitoring the electrical condition of accumulators or electric batteries, e.g. capacity or state of charge [SoC]
    • G01R31/392—Determining battery ageing or deterioration, e.g. state of health

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Secondary Cells (AREA)
  • Charge And Discharge Circuits For Batteries Or The Like (AREA)

Abstract

The method adopts a joint estimation strategy to identify and correct parameters of a state-of-charge space model and a state-of-health space model in real time, so that the accuracy and effectiveness of a battery equivalent model are ensured, and the estimation precision is improved. By carrying out quasi-linearization processing on the measurement equations in the charge state space model and the health state space model, the calculation overhead during each traceless transformation can be effectively reduced. In the iterative process, the square root of the state error covariance matrix is used for replacing the state error covariance matrix, the square root is obtained by QR decomposition and first-order updating of Cholesky factors, the problem that a filtering result is diverged due to negative determination of the state error covariance matrix caused by calculation accumulated errors in the iterative process of a standard UKF algorithm is solved well, and the numerical stability of the online rolling estimation of the state of the lithium ion power battery is ensured.

Description

SR-UKF-based lithium ion power battery state estimation method
Technical Field
The application relates to the technical field of lithium ion power batteries, in particular to a state estimation method of a lithium ion power battery based on SR-UKF; and further relates to a lithium ion power battery state estimation device based on SR-UKF.
Background
The lithium ion power battery is an important energy supply source of the new energy electric automobile. Due to fluctuation in the battery production process and non-uniformity of materials, the capacity, internal resistance, self-discharge characteristics and the like of the single battery have certain differences. And along with the increase of the number of charging and discharging cycles and the influence of special working environment in the vehicle, the battery capacity is attenuated to different degrees, so that the difference between the monomers is further aggravated. The accurate estimation of the battery state (including the state of charge and the state of health) can provide a reliable reference basis for battery grouping, Battery Management System (BMS) balancing and other purposes, thereby having important practical significance for full and reasonable utilization of single batteries, prolonging of the service life of a battery pack and improvement of the running efficiency of the whole vehicle.
Currently, the estimation method of the state of charge (SOC) of the battery mainly includes an ampere-hour counting (AH) method, an open-circuit voltage method, an Extended Kalman Filter (EKF) method, an Unscented Kalman Filter (UKF) method, and a particle filter. The unscented Kalman filtering method approximates the posterior probability density distribution of a nonlinear system by constructing a deterministic sample meeting a certain rule, the estimation precision of the unscented Kalman filtering method is obviously improved relative to EKF, but the unscented Kalman filtering method cannot ensure the non-negativity of a state error covariance matrix in the filtering process, the hidden danger of filtering divergence exists, and the estimation precision of the unscented Kalman filtering method is also limited by the accuracy of a battery model to a certain extent. The method for estimating state of health (SOH) of a battery mainly comprises the following steps: based on the prediction of the characteristics, the aging degree of the battery is indirectly predicted through the corresponding relation between the characteristic parameters and the service life of the battery, but the measurement difficulty of the characteristic parameters is higher; based on data-driven prediction, potential rules describing battery performance evolution are mined out by using test data, and further the service life of the battery is predicted. Such as Support Vector Machine (SVM), particle filter, neural network, etc., the methods usually have a certain limitation in engineering application due to the limitation and uncertainty of experimental data.
Therefore, how to accurately estimate the state of the lithium-ion power battery has become a technical problem to be solved by those skilled in the art.
Disclosure of Invention
The method for estimating the state of the lithium ion power battery based on the SR-UKF is capable of ensuring model accuracy, reducing calculation overhead and solving the problem of divergence of a filtering result caused by negative determination of a state error covariance matrix due to calculation accumulated errors, so that the state of the lithium ion power battery is accurately estimated; another object of the present application is to provide a lithium ion power battery state estimation device based on SR-UKF, which also has the above technical effects.
In order to solve the technical problem, the application provides a lithium ion power battery based on SR-UKF, and a state estimation method comprises the following steps:
establishing a charge state space model and a health state space model of the lithium ion power battery;
performing iterative calculation by using an SR-UKF algorithm based on the state of charge space model to obtain the state of charge and the polarization voltage of the lithium ion power battery, and updating the polarization voltage parameter in the state of health space model according to the polarization voltage; in the process of carrying out iterative computation by using an SR-UKF algorithm based on a state-of-charge space model, constructing a Sigma point set according to Cholesky factors of a mean value and a state error covariance of a state vector in the state-of-charge space model, and obtaining a coefficient matrix related to a measurement equation in the state-of-charge space model by inquiring an OCV-SOC mapping table under the current condition;
performing iterative calculation by using an SR-UKF algorithm based on the state of health space model to obtain ohmic internal resistance of the lithium ion power battery, calculating to obtain actual rated capacity of the lithium ion power battery according to the state of charge of the lithium ion power battery, calculating to obtain the state of health of the lithium ion power battery according to the ohmic internal resistance or the actual rated capacity, and updating ohmic internal resistance parameters and rated capacity parameters in the state of charge space model according to the ohmic internal resistance and the actual rated capacity; and constructing a Sigma point set according to Cholesky factors of the mean value and the state error covariance of state variables in the health state space model in the process of carrying out iterative computation by utilizing an SR-UKF algorithm based on the health state space model, and obtaining coefficients related to a measurement equation in the health state space model by inquiring an OCV-SOC mapping table under the current condition.
Optionally, the establishing of the state of charge space model of the lithium ion power battery includes:
based on the definition of the second-order RC equivalent model of the lithium ion power battery and the ampere-hour integral measurement method of the charge state of the lithium ion power battery, x is used1(k)＝[S(k),Ucs(k),Ucl(k)]TAs a state vector, y (k) is Uo(k) As system output, u (k) as control input, obtaining the one-step predicted state-of-charge space model:
wherein u (k) is i (k), w (k) is w1(k),w2(k),w3(k)]TIs process noise, v (k) is observation noise, A, B, C is a coefficient matrix, and
Figure BDA0002289435530000032
Figure BDA0002289435530000033
D＝-Re，Tsfor a sampling period, τs、τlAs time constant, η as charge-discharge efficiency, Q0Is the rated capacity, R, of the lithium ion power batterys、RlS (k) is the polarization internal resistance of the lithium ion power battery, S (k) is the charge state of the lithium ion power battery at the moment k, EtF { s (k) } represents the equilibrium electromotive force E of the lithium ion power batterytMapping relation with the state of charge of the lithium ion power battery, ReIs the ohmic internal resistance of the lithium ion power battery.
Optionally, the establishing a state of health space model of the lithium ion power battery includes:
based on the second-order RC equivalent model of the lithium ion power battery, x is used2(k)＝Re(k) As state variables, the state of health spatial model is obtained with a one-step prediction:
Figure BDA0002289435530000034
where r (k) is process noise, E ═ i (k), F ═ F { s (k) } -Ucs(k)-Ucl(k) I (k) is the discharge current of the lithium ion power battery at the moment k, EtF { s (k) } represents the equilibrium electromotive force E of the lithium ion power batterytFunctional relationship with the state of charge of a lithium ion power cell, Ucs(k) And Ucl(k) And q (k) is observation noise, wherein the voltage drop is generated by the polarization internal resistance of the lithium ion power battery at the moment k.
Optionally, the obtaining of the actual rated capacity of the lithium ion power battery by calculating according to the state of charge of the lithium ion power battery includes:
based on
Figure BDA0002289435530000035
Calculating to obtain the actual rated capacity of the lithium ion power battery;
wherein Q ispresentFor the actual rated capacity of the lithium ion power battery,
Figure BDA0002289435530000036
is t1The state of charge of the lithium ion power battery at the moment,charging for constant current to t2The state of charge of the lithium ion power battery is obtained at the moment and after standing for a period,
Figure BDA0002289435530000042
for charge, η is the charge-discharge efficiency, and I is the constant current charge current.
Optionally, the calculating the health state of the lithium ion power battery according to the ohmic internal resistance includes:
based on
Figure BDA0002289435530000043
Calculating to obtain the health state of the lithium ion power battery;
wherein SOH is the health state, REOLIs the impedance value at the end of the service life of the lithium ion power battery, R0Is the impedance value, R, of the lithium ion power battery when the lithium ion power battery leaves a factoryeThe ohmic internal resistance is obtained by online identification.
Optionally, calculating the state of health of the lithium ion power battery according to the actual rated capacity includes:
based on
Figure BDA0002289435530000044
Calculating to obtain the health state of the lithium ion power battery;
wherein SOH is the health state, QpresentFor said rated capacity, Q0And the rated capacity is the rated capacity of the lithium ion power battery when the lithium ion power battery leaves a factory.
Optionally, the method further includes:
and calibrating the initial value of the state of charge of the lithium ion power battery.
Optionally, the method further includes:
a variance of the process noise is estimated and updated based on the estimated value.
Optionally, the method further includes:
judging whether a filtering result meets the convergence requirement or not in the process of carrying out iterative computation by utilizing the SR-UKF algorithm;
and if not, correcting the state error covariance of the state vector in the state of charge space model and/or correcting the state error covariance of the state variable in the state of health space model.
In order to solve the above technical problem, the present application further provides a lithium ion power battery state estimation device based on SR-UKF algorithm, including:
a memory for storing a computer program;
and the processor is used for realizing the steps of the lithium ion power battery state estimation method based on the SR-UKF algorithm when executing the computer program.
The lithium ion power battery state estimation method based on the SR-UKF algorithm comprises the steps of establishing a state-of-charge space model and a state-of-health space model of the lithium ion power battery; performing iterative calculation by using an SR-UKF algorithm based on the state of charge space model to obtain the state of charge and the polarization voltage of the lithium ion power battery, and updating the polarization voltage parameter in the state of health space model according to the polarization voltage; in the process of carrying out iterative computation by utilizing an SR-UKF algorithm based on a state of charge space model, constructing a Sigma point set according to Cholesky factors of the mean value and the state error covariance of state vectors in the state of charge space model, and obtaining a coefficient matrix related to a measurement equation in the state of charge space model by inquiring an OCV-SOC mapping table under the current condition; performing iterative calculation by using an SR-UKF algorithm based on the state of health space model to obtain ohmic internal resistance of the lithium ion power battery, calculating to obtain actual rated capacity of the lithium ion power battery according to the state of charge of the lithium ion power battery, calculating to obtain the state of health of the lithium ion power battery according to the ohmic internal resistance or the actual rated capacity, and updating ohmic internal resistance parameters and rated capacity parameters in the state of charge space model according to the ohmic internal resistance and the actual rated capacity; and constructing a Sigma point set according to Cholesky factors of the mean value and the state error covariance of state variables in the health state space model in the process of carrying out iterative computation by utilizing an SR-UKF algorithm based on the health state space model, and obtaining coefficients related to a measurement equation in the health state space model by inquiring an OCV-SOC mapping table under the current condition.
Therefore, the lithium ion power battery state estimation method based on the SR-UKF algorithm identifies and updates parameters of a state of charge space model and a state of health space model of the lithium ion power battery in real time, ensures the accuracy and effectiveness of the models, and improves the estimation precision of the state of the lithium ion power battery. In addition, the method obtains the related coefficient matrix and coefficient by a table look-up mode, namely carries out quasi-linearization processing on the measurement equation, and can greatly reduce the calculation overhead during the unscented transformation. In addition, in the iterative computation process, a Sigma point set is constructed according to a Cholesky factor of the state error covariance, namely the state error covariance is replaced by the square root of the state error covariance, and the square root is obtained by QR decomposition and first-order updating of the Cholesky factor, so that the problem of filtering result divergence caused by negative determination of a state error covariance matrix due to the calculation of accumulated errors in the UKF algorithm iterative process can be effectively solved.
The lithium ion power battery state estimation device based on the SR-UKF algorithm has the technical effects.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the prior art and the embodiments are briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
Fig. 1 is a schematic flow chart of a lithium ion power battery state estimation method based on an SR-UKF algorithm according to an embodiment of the present application;
fig. 2 is a schematic diagram of a second-order equivalent model of a lithium ion power battery provided in an embodiment of the present application;
fig. 3 is a block diagram of estimating a state of a lithium-ion power battery according to an embodiment of the present disclosure.
Detailed Description
The core of the application is to provide a lithium ion power battery state estimation method based on SR-UKF, which can ensure model accuracy, reduce calculation overhead and solve the problem of divergence of filtering results caused by negative determination of a state error covariance matrix due to calculation accumulated errors, thereby realizing the purpose of accurately estimating the state of the lithium ion power battery; the other core of the application is to provide a lithium ion power battery state estimation device based on SR-UKF, and the technical effects are also achieved.
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
Referring to fig. 1, fig. 1 is a schematic flow chart of a lithium ion power battery state estimation method based on SR-UKF algorithm according to an embodiment of the present application; referring to fig. 1, the method includes:
s101: establishing a charge state space model and a health state space model of the lithium ion power battery;
specifically, the step aims to establish a state of charge space model and a state of health space model of the lithium ion power battery, and then estimate the state of the lithium ion power battery based on the state of charge space model and the state of health space model.
The establishing of the state of charge space model of the lithium ion power battery can comprise the following steps:
defining a second-order RC equivalent model based on the lithium ion power battery and an ampere-hour integral measurement method of the charge state of the lithium ion power battery by x1(k)＝[S(k),Ucs(k),Ucl(k)]TAs a state vector, y (k) is Uo(k) As system output, u (k) is used as control input, and a one-step predicted state-of-charge space model is obtained:
Figure BDA0002289435530000071
wherein u (k) is i (k), w (k) is w1(k),w2(k),w3(k)]TIs process noise, v (k) is observation noise, A, B, C is a coefficient matrix, and
Figure BDA0002289435530000072
Figure BDA0002289435530000073
D＝-Re，Tsfor a sampling period, τs、τlAs time constant, η as charge-discharge efficiency, Q0Is the rated capacity, R, of the lithium ion power batterys、RlIs the polarization internal resistance of the lithium ion power battery, S (k) is the charge state of the lithium ion power battery at the moment k, EtF { s (k) } represents the equilibrium electromotive force E of the lithium ion power batterytMapping relation with the state of charge of the lithium ion power battery, ReIs the ohmic internal resistance of the lithium ion power battery.
Specifically, referring to the second order RC equivalent model of the lithium ion power battery shown in fig. 2, EtThe battery is kept in a balanced state by standing for a long time, and the measured terminal voltage of the battery is equal to the balanced electromotive force of the battery in value, ItIs the discharge current of the battery, ReIs the ohmic internal resistance, R, of the batterys、RlIs the internal polarization resistance of the cell, Cs、ClIs the polarization capacitance, U, of the batterycs,t、Ucl,tFor voltage drop, U, produced in the polarization internal resistance of the cello,tIs the observed voltage across the cell.
Further, based on the second order RC equivalent model of the lithium ion power battery shown in fig. 2, it is possible to obtain:
Figure BDA0002289435530000074
Figure BDA0002289435530000075
Et＝ReIt+Ucs,t+Ucl,t+Uo,t(3)
wherein, taus、τlIs a time constant.
Further, the ampere-hour integral of the lithium ion power battery is defined as:
Figure BDA0002289435530000076
wherein the content of the first and second substances,
Figure BDA0002289435530000077
is a lithium ion power battery t0State of charge at time, i.e. initial value of SOC, Q0The rated capacity of the lithium ion power battery is shown, and η is the charge-discharge efficiency.
Thus, according to the formulae (1), (2), (3) and (4), x is expressed1(k)＝[S(k),Ucs(k),Ucl(k)]TAs a state vector, y (k) is Uo(k) As system output, u (k) as control input, w (k), v (k) as system disturbance, established to obtain x1(k) State of charge space model for state vector:
Figure BDA0002289435530000081
wherein x is1(k+1|k)＝Ax1(k) + Bu (k) + w (k) is a prediction equation，y1(k)＝Cx1(k) + Du (k) + v (k) is the measurement equation. Coefficient matrix C in measurement equation and balanced electromotive force E of lithium ion power batterytAnd the state of charge S (k) is related, EtF { s (k) } denotes the equilibrium electromotive force E of the celltAnd the mapping relation between the SOC of the battery is related to factors such as ambient temperature and the aging degree of the lithium ion power battery. Ohmic internal resistance R in the state-of-charge space modeleInternal polarization resistance RlAnd RsAre pre-identified model parameters.
Additionally, establishing a state-of-health spatial model of the lithium-ion power battery may include:
based on a second-order RC equivalent model of the lithium ion power battery, x2(k)＝Re(k) As state variables, a one-step predicted health state spatial model is obtained:
Figure BDA0002289435530000082
where r (k) is process noise, E ═ i (k), F ═ F { s (k) } -Ucs(k)-Ucl(k) I (k) is the discharge current of the lithium ion power battery at the time k, EtF { s (k) } represents the equilibrium electromotive force E of the lithium ion power batterytFunctional relationship with the state of charge of a lithium ion power cell, Ucs(k) And Ucl(k) And q (k) is observation noise.
Specifically, based on the second-order RC equivalent model of the lithium ion power battery shown in fig. 2, the state of charge and the polarization voltage of the lithium ion power battery are taken as known values, and the following can be obtained:
Re,t+1＝Re,t+rt(6)
Et＝Re,tIt+Ucs,t+Ucl,t+Uo,t+qt(7)
wherein r istIs process noise, qtTo observe the noise.
Further, according to the formulae (6) and (7), x is given2(k)＝Re(k) Established to obtain the value of x2(k) A state space model for state variables:
Figure BDA0002289435530000083
wherein x is2(k+1|k)＝x2(k) + r (k) is the prediction equation, y2(k)＝Ex2(k) + F + q (k) is a measurement equation in which the coefficient F and the equilibrium electromotive force E of the lithium ion power batterytAnd state of charge s (k).
S102: performing iterative calculation by using an SR-UKF algorithm based on a state of charge space model to obtain the state of charge and the polarization voltage of the lithium ion power battery, and updating the polarization voltage parameter in the state of health space model according to the polarization voltage; in the process of carrying out iterative computation by using an SR-UKF algorithm based on a state of charge space model, constructing a Sigma point set according to Cholesky factors of the mean value and the state error covariance of a state vector in the state of charge space model, and obtaining a coefficient matrix related to a measurement equation in the state of charge space model by inquiring an OCV-SOC mapping table under the current condition;
specifically, the relevance between the state of charge and the state of health of the lithium ion power battery is fully considered, a combined estimation mode for the state of the lithium ion power battery is provided, and referring to fig. 3, the problem of time variation of model parameters can be well solved by identifying and correcting the model parameters on line in real time, and the accuracy and the effectiveness of the model are guaranteed. The method comprises the steps of obtaining the charge state and the polarization voltage of the lithium ion power battery based on a charge state space model by utilizing an SR-UKF algorithm, and updating polarization voltage parameters in a health state space model. The initial value of the state of charge can be calibrated first to avoid accumulated errors and improve the convergence rate.
The SR-UKF algorithm is a square root Unscented Kalman Filter algorithm (square-root Unscented Kalman Filter). The core of UKF (unscented Kalman Filter) is UT (unscented Kalman Filter), namely the core of UKF is UT (unscented Kalman Filter), namely a Sigma point set is constructed by deterministic sampling to approximate the probability density distribution of a nonlinear function of a system so as to solve the nonlinear filtering problem. However, different from the standard UKF, the SR-UKF filtering provided by the application uses the square root of the state error covariance matrix to perform iterative operation in the UKF filtering process so as to avoid the problem that the filtering result is diverged due to negative determination of the covariance matrix.
Specifically, the UT transforms:
the mean and square root of covariance of the random vector are used to construct a 2N +1 dimensional Sigma point set:
Figure BDA0002289435530000091
wherein the content of the first and second substances,
Figure BDA0002289435530000092
Sxx in the state of charge space model respectively shown in formula (5)1(k) The Cholesky factor of the mean and state error covariance of (a) is the design parameter.
The description of the Cholesky factor is as follows:
1) QR decomposition
If positive definite matrix Q existsm×mAnd the upper triangular matrix Rm×nSo that A ism×n＝Qm×mRm×nThis is called QR decomposition of a, and R is expressed as QR { a }.
2) Cholesky factor
From definition 1, ATQR decomposition of (i.e. R)If the matrix P is AATThen there is
Figure BDA0002289435530000101
So it is called
Figure BDA0002289435530000102
Cholesky factor of P, noted
Figure BDA0002289435530000103
3) First order update of Cholesky factor
From 2) knowing that the Cholesky factor for P is
Figure BDA0002289435530000104
Then call
Figure BDA0002289435530000105
The Cholesky factor ofFirst order update of
Figure BDA0002289435530000107
Further, in order to better approximate the posterior distribution condition of the system state, weight design is performed on the Sigma point set:
α is used to describe the deviation degree of the Sigma point set, and the value range is (10)-41) β is used for describing the distribution of system states, 2 is taken in the case of Gaussian distribution, and lambda is α2(N + k) -N, which affects the approximation accuracy, wherein the parameter k may typically be taken to be 0.
Further, carrying out nonlinear transformation on the Sigma point set, and substituting the constructed Sigma point set into an equation (5) to obtain:
Figure BDA0002289435530000109
in the formula, f {. cndot, h {. cndot, respectively represent the system state transition and input-output relationship, wherein h {. cndot, is obtained by quasi-linearizing the measurement equation by a table look-up method, i.e. obtaining S (k) and E (E) by querying the OCV-SOC mapping table under the current condition on the basis of obtaining a plurality of OCV-SOC mapping tables by pre-fitting under different conditionstAnd obtaining a coefficient matrix C in a measurement equation in the charge state space model.
Further, iteratively calculating: assuming that the process noise w (k) obeys an N (0, Q) distribution, the observation noise v (k) obeys an N (0, R) distribution.
The prediction updating process is as follows:
Figure BDA00022894355300001010
the observation updating process comprises the following steps:
Figure BDA0002289435530000111
wherein Z is an observed value, G is a Kalman gain, used to dynamically adjust the weight distribution between the state prediction and the observed residual,
Figure BDA0002289435530000112
is the optimal estimation result.
In addition, on the basis of obtaining the state of charge and the polarization voltage of the lithium ion power battery through iterative calculation, ohmic internal resistance parameters in the health space model are further updated according to the polarization voltage, so that the accuracy and the effectiveness of the health space model are ensured.
S103: performing iterative calculation by using an SR-UKF algorithm based on a state of health space model to obtain ohmic internal resistance of the lithium ion power battery, calculating actual rated capacity of the lithium ion power battery according to the state of charge of the lithium ion power battery, calculating the state of health of the lithium ion power battery according to the ohmic internal resistance or the actual rated capacity, and updating an ohmic internal resistance parameter and a rated capacity parameter in the state of charge space model according to the ohmic internal resistance and the actual rated capacity; in the process of carrying out iterative computation by using an SR-UKF algorithm based on the health state space model, a Sigma point set is constructed according to Cholesky factors of the mean value and the state error covariance of state variables in the health state space model, and coefficients related to a measurement equation in the health state space model are obtained by inquiring an OCV-SOC mapping table under the current condition.
Specifically, the step aims to calculate by utilizing an SR-UKF algorithm based on a health state space modelAnd obtaining the real-time ohmic internal resistance of the lithium ion power battery, and updating the ohmic internal resistance parameter in the charge state space model according to the ohmic internal resistance obtained by calculation so as to ensure the accuracy and the effectiveness of the charge state space model. Similar to the way of estimating the lithium ion power battery by using the SR-UKF algorithm, the online rolling estimation of the ohmic internal resistance can be realized only by converting the state-of-charge space model shown in the formula (5) into the state-of-health space model shown in the formula (8). I.e. based on the state variable x in the state-of-health space model when UT is transformed2(k) The Cholesky factor of the mean and the state error covariance of (a) constructs a Sigma point set, the Sigma point set is subjected to weight design as described above, nonlinear transformation is performed on the Sigma point set, and the constructed Sigma point set is substituted into formula (8). Wherein, in the step, the OCV-SOC mapping table under the current condition is inquired to obtain S (k) and EtTherefore, the coefficient F in the measurement equation in the health state space model is obtained, and the calculation expense in UT conversion is reduced. Further, assuming that the process noise R (k) follows N (0, Q) distribution, and the observation noise Q (k) follows N (0, R) distribution, a prediction update process and an observation update process similar to those described above are performed, thereby obtaining an optimal estimation result.
Further, the actual rated capacity of the lithium ion power battery is obtained through calculation according to the state of charge of the lithium ion power battery obtained through calculation in the step S102, and the rated capacity parameter in the state of charge space model is updated according to the actual rated capacity.
The calculating the actual rated capacity of the lithium-ion power battery according to the state of charge of the lithium-ion power battery may include:
based on
Figure BDA0002289435530000121
Calculating to obtain the actual rated capacity of the lithium ion power battery;
wherein Q ispresentWhen the rated capacity parameter in the state of charge space equation is updated, namely the actual rated capacity is updated according to the updated rated capacity,
Figure BDA0002289435530000122
is t1At the moment, the state of charge of the lithium ion power battery,
Figure BDA0002289435530000123
charging for constant current to t2The state of charge of the lithium ion power battery is kept still for a period of time,
Figure BDA0002289435530000124
for charging, η is the charging and discharging efficiency, I is the constant current charging current, that is, the actual rated capacity of the lithium ion power battery is calculated by the method of constant current charging-standing in the embodiment.
In addition, the application provides two ways for calculating the state of health of the lithium ion power battery, including a state of health estimation way based on the ohmic internal resistance of the lithium ion power battery and a state of health estimation way based on the battery capacity.
The calculating the health state of the lithium ion power battery according to the ohmic internal resistance may include:
based on
Figure BDA0002289435530000125
Calculating to obtain the health state of the lithium ion power battery;
wherein SOH is a healthy state, REOLIs the impedance value at the end of the service life of the lithium ion power battery, R0Is the impedance value, R, of the lithium ion power battery when leaving the factoryeThe obtained ohmic internal resistance is identified on line.
The method for calculating the health state of the lithium ion power battery according to the actual rated capacity comprises the following steps:
based on
Figure BDA0002289435530000126
Calculating to obtain the health state of the lithium ion power battery;
wherein SOH is healthy state, QpresentIs the actual rated capacity, Q, of the lithium ion power battery0Is a lithium ion power batteryRated capacity at the time of shipment.
Further, on the basis of the above-described embodiment, the lithium-ion power battery state estimation method may further include estimating a variance of the process noise and updating the variance of the process noise based on the estimated value.
Specifically, it is assumed that the process noise and the observation noise are subject to a preset normal distribution, but if the statistical characteristic parameter of the process noise is inaccurate, the filtering result is often diverged, so in order to improve the adaptive fault-tolerant capability, the variance of the process noise is estimated and updated on line in an iteration manner. The Sage-Husa adaptive filter for Maximum A Posteriori (MAP) can better estimate the first moment and the second moment of noise, has clear and simple principle and is widely applied to engineering practice. Therefore, an updated variance value of the process noise can be obtained through Sage-Husa adaptive filter on-line estimation; the variance of the process noise is then updated based on the variance update value. Specifically, a Sage-Husa estimation method is adopted to carry out online estimation on the variance Q of process noise, and meanwhile, a forgetting factor is introduced to reduce the influence of historical data, and the expression is as follows:
Figure BDA0002289435530000131
wherein d (k) is (1-b)/(1-b)k) And b is an adjustable forgetting factor, the value range is (0.95,0.99), if the fluctuation of the statistical characteristic of the process noise is large, the value of b is increased, otherwise, the value of b is decreased.
After one-step iteration is completed according to the formulas (12), (13) and (14), the optimal estimation can be obtained
Figure BDA0002289435530000132
And
Figure BDA0002289435530000133
and updating the statistical characteristics of the mean, covariance and process noise in the formula (1) according to the current estimation result, and repeatedly executing the process in the next iteration.
Further, on the basis of the above embodiment, the method for estimating the state of the lithium ion power battery may further include determining whether the filtering result meets the convergence requirement in the process of performing iterative computation by using an SR-UKF algorithm; and if not, correcting the state error covariance of the state vector in the state-of-charge space model and/or correcting the state error covariance of the state variable in the state-of-health space model.
Specifically, the Sage-Husa adaptive filter is a suboptimal unbiased estimation, and the statistical characteristic of noise is estimated by using one-step prediction, so that the second moment of the noise is easy to lose positive or semi-positive, and filtering divergence is caused. In the inverse equation (14), when estimating the second moment of the noise, the subtraction operation exists in the equation, so that the non-negativity of the second moment is difficult to ensure in the filtering process. Therefore, in order to avoid the risk of divergence of the filter result caused by negative determination of the covariance matrix, the convergence of the filtering result needs to be judged in the iterative process, and the covariance is corrected by the adaptive attenuation factor when the filtering result is in a divergence trend.
Specifically, judgment is made
Figure BDA0002289435530000134
Whether the result is true or not; trace {. is trace calculation, gamma is adjustable coefficient and gamma is greater than or equal to 1. If the above formula does not hold, then according to
Figure BDA0002289435530000135
Correction of Px(state error covariance). Conversely, if the above equation is satisfied, no correction is necessary. In the above formula, λ is an adaptive attenuation factor for weakening the dependency of the state prediction on the historical data to increase the confidence of the current measurement residual, thereby suppressing the filtering divergence, and
Figure BDA0002289435530000136
in conclusion, the lithium ion power battery state estimation method based on the SR-UKF algorithm provided by the application identifies and updates the parameters of the state of charge space model and the state of health space model of the lithium ion power battery in real time, ensures the accuracy and effectiveness of the models, and improves the estimation precision of the state of the lithium ion power battery. In addition, the method obtains the related coefficient matrix and coefficient by a table look-up mode, namely carries out quasi-linearization processing on the measurement equation, and can greatly reduce the calculation overhead during the unscented transformation. In addition, in the iterative computation process, a point set is constructed according to the Cholesky factor of the state error covariance, namely the state error covariance is replaced by the square root of the state error covariance, and the square root is obtained by QR decomposition and first-order updating of the Cholesky factor, so that the problem of filtering result divergence caused by negative determination of the state error covariance matrix due to the calculation accumulated error in the UKF algorithm iterative process can be effectively solved.
The application also provides lithium ion power battery state estimation equipment based on the SR-UKF algorithm, and the equipment described below can be correspondingly referred to with the method described above. The apparatus includes a memory and a processor. Wherein the memory is used for storing a computer program; the processor is used for implementing the steps of the lithium ion power battery state estimation method based on the SR-UKF algorithm when executing the computer program.
The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device, the apparatus and the computer-readable storage medium disclosed by the embodiments correspond to the method disclosed by the embodiments, so that the description is simple, and the relevant points can be referred to the description of the method.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
The method and the device for estimating the state of the lithium ion power battery based on the SR-UKF algorithm provided by the application are described in detail above. The principles and embodiments of the present application are explained herein using specific examples, which are provided only to help understand the method and the core idea of the present application. It should be noted that, for those skilled in the art, it is possible to make several improvements and modifications to the present application without departing from the principle of the present application, and such improvements and modifications also fall within the scope of the claims of the present application.

Claims (10)

1. A lithium ion power battery state estimation method based on SR-UKF algorithm is characterized by comprising the following steps:
establishing a charge state space model and a health state space model of the lithium ion power battery;
performing iterative calculation by using an SR-UKF algorithm based on the state of charge space model to obtain the state of charge and the polarization voltage of the lithium ion power battery, and updating the polarization voltage parameter in the state of health space model according to the polarization voltage; in the process of carrying out iterative computation by using an SR-UKF algorithm based on a state-of-charge space model, constructing a Sigma point set according to a mean value of state vectors in the state-of-charge space model and a Cholesky factor of state error covariance, and obtaining a coefficient matrix related to a measurement equation in the state-of-charge space model by inquiring an OCV-SOC mapping table under the current condition;
performing iterative calculation by using an SR-UKF algorithm based on the state of health space model to obtain ohmic internal resistance of the lithium ion power battery, calculating to obtain actual rated capacity of the lithium ion power battery according to the state of charge of the lithium ion power battery, calculating to obtain the state of health of the lithium ion power battery according to the ohmic internal resistance or the actual rated capacity, and updating ohmic internal resistance parameters and rated capacity parameters in the state of charge space model according to the ohmic internal resistance and the actual rated capacity; and constructing a Sigma point set according to Cholesky factors of the mean value and the state error covariance of state variables in the health state space model in the process of carrying out iterative computation by utilizing an SR-UKF algorithm based on the health state space model, and obtaining coefficients related to a measurement equation in the health state space model by inquiring an OCV-SOC mapping table under the current condition.
2. The method according to claim 1, wherein establishing the state-of-charge space model of the lithium-ion power battery comprises:
based on the definition of the second-order RC equivalent model of the lithium ion power battery and the ampere-hour integral measurement method of the charge state of the lithium ion power battery, x is used1(k)＝[S(k),Ucs(k),Ucl(k)]TAs a state vector, y (k) is Uo(k) As system output, u (k) as control input, obtaining the one-step predicted state-of-charge space model:
Figure FDA0002289435520000011
wherein u (k) is i (k), w (k) is w1(k),w2(k),w3(k)]TIs process noise, v (k) is observation noise, A, B, C is a coefficient matrix, and
Figure FDA0002289435520000021
Figure FDA0002289435520000022
D＝-Re，Tsfor a sampling period, τs、τlAs time constant, η as charge-discharge efficiency, Q0Is the rated capacity, R, of the lithium ion power batterys、RlS (k) is the polarization internal resistance of the lithium ion power battery, S (k) is the charge state of the lithium ion power battery at the moment k, EtF { s (k) } represents the equilibrium electromotive force E of the lithium ion power batterytFunctional relationship with the state of charge, R, of a lithium ion power celleIs the ohmic internal resistance of the lithium ion power battery.
3. The method of claim 1, wherein establishing the state-of-health spatial model of the lithium-ion power battery comprises:
based on the second-order RC equivalent model of the lithium ion power battery, x is used2(k)＝Re(k) As state variables, the state of health spatial model is obtained with a one-step prediction:
Figure FDA0002289435520000023
where r (k) is process noise, E ═ i (k), F ═ F { s (k) } -Ucs(k)-Ucl(k) I (k) is the discharge current of the lithium ion power battery at the moment k, EtF { s (k) } represents the equilibrium electromotive force E of the lithium ion power batterytFunctional relationship with the state of charge of a lithium ion power cell, Ucs(k) And Ucl(k) And q (k) is observation noise, wherein the voltage drop is generated by the polarization internal resistance of the lithium ion power battery at the moment k.
4. The method according to claim 1, wherein the calculating the actual rated capacity of the lithium-ion power battery according to the state of charge of the lithium-ion power battery comprises:
based on
Figure FDA0002289435520000024
Calculating to obtain the actual rated capacity of the lithium ion power battery;
wherein Q ispresentFor the purpose of said actual rated capacity,
Figure FDA0002289435520000025
is t1The state of charge of the lithium ion power battery at the moment,
Figure FDA0002289435520000026
charging for constant current to t2The state of charge of the lithium ion power battery is obtained at the moment and after standing for a period,
Figure FDA0002289435520000027
for charge, η is the charge-discharge efficiency, and I is the constant current charge current.
5. The method of estimating the state of the lithium-ion power battery according to claim 4, wherein calculating the state of health of the lithium-ion power battery from the ohmic internal resistance comprises:
based on
Figure FDA0002289435520000031
Calculating to obtain the health state of the lithium ion power battery;
wherein SOH is the health state, REOLIs the impedance value at the end of the service life of the lithium ion power battery, R0Is the impedance value, R, of the lithium ion power battery when the lithium ion power battery leaves a factoryeThe ohmic internal resistance is obtained by online identification.
6. The method of claim 5, wherein calculating the state of health of the lithium-ion power battery based on the actual rated capacity comprises:
based on
Figure FDA0002289435520000032
Calculating to obtain the health state of the lithium ion power battery;
wherein SOH is the health state, QpresentIs the actual rated capacity, Q, of the lithium ion power battery0And the rated capacity is the rated capacity of the lithium ion power battery when the lithium ion power battery leaves a factory.
7. The lithium-ion power battery state estimation method of claim 6, further comprising:
and calibrating the initial value of the state of charge of the lithium ion power battery.
8. The lithium-ion power battery state estimation method according to claim 7, further comprising:
a variance of the process noise is estimated and updated based on the estimated value.
9. The lithium-ion power battery state estimation method of claim 8, further comprising:
judging whether a filtering result meets the convergence requirement or not in the process of carrying out iterative computation by utilizing the SR-UKF algorithm;
and if not, correcting the state error covariance of the state vector in the state of charge space model and/or correcting the state error covariance of the state variable in the state of health space model.
10. A lithium ion power battery state estimation device based on SR-UKF algorithm is characterized by comprising:
a memory for storing a computer program;
a processor for implementing the steps of the SR-UKF algorithm based lithium ion power battery state estimation method according to any of claims 1 to 9 when executing the computer program.
CN201911173741.9A 2019-11-26 2019-11-26 SR-UKF-based lithium ion power battery state estimation method Active CN110850299B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911173741.9A CN110850299B (en) 2019-11-26 2019-11-26 SR-UKF-based lithium ion power battery state estimation method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911173741.9A CN110850299B (en) 2019-11-26 2019-11-26 SR-UKF-based lithium ion power battery state estimation method

Publications (2)

Publication Number Publication Date
CN110850299A true CN110850299A (en) 2020-02-28
CN110850299B CN110850299B (en) 2021-01-15

Family

ID=69604499

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911173741.9A Active CN110850299B (en) 2019-11-26 2019-11-26 SR-UKF-based lithium ion power battery state estimation method

Country Status (1)

Country Link
CN (1) CN110850299B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113805075A (en) * 2021-09-15 2021-12-17 上海电机学院 BCRLS-UKF-based lithium battery state of charge estimation method

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101251582A (en) * 2007-02-20 2008-08-27 德尔菲技术公司 Method of determining the energy capacity of a battery
CN102645636A (en) * 2012-04-19 2012-08-22 北京优科利尔能源设备有限公司 Battery capacity detection method
CN104730462A (en) * 2013-12-19 2015-06-24 北汽福田汽车股份有限公司 Battery pack capacity online detecting method
CN105223506A (en) * 2014-06-30 2016-01-06 崇实大学校产学协力团 Use battery status method of estimation and the system of two extended Kalman filter
CN105974329A (en) * 2016-07-22 2016-09-28 深圳市沃特玛电池有限公司 Method for estimating SOH of battery pack
CN108445401A (en) * 2018-02-09 2018-08-24 深圳市鹏诚新能源科技有限公司 On-line Estimation method, electronic device and the storage medium of battery charge state SOC
CN108594135A (en) * 2018-06-28 2018-09-28 南京理工大学 A kind of SOC estimation method for the control of lithium battery balance charge/discharge
CN109633473A (en) * 2019-01-23 2019-04-16 刘平 A kind of distributed battery group state-of-charge algorithm for estimating
CN109870651A (en) * 2019-01-22 2019-06-11 重庆邮电大学 A kind of electric automobile power battery system SOC and SOH joint estimation on line method
CN110398691A (en) * 2019-06-26 2019-11-01 重庆大学 Based on the lithium-ion-power cell SoC estimation method for improving adaptive double Unscented kalman filtering devices

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101251582A (en) * 2007-02-20 2008-08-27 德尔菲技术公司 Method of determining the energy capacity of a battery
CN102645636A (en) * 2012-04-19 2012-08-22 北京优科利尔能源设备有限公司 Battery capacity detection method
CN104730462A (en) * 2013-12-19 2015-06-24 北汽福田汽车股份有限公司 Battery pack capacity online detecting method
CN105223506A (en) * 2014-06-30 2016-01-06 崇实大学校产学协力团 Use battery status method of estimation and the system of two extended Kalman filter
CN105974329A (en) * 2016-07-22 2016-09-28 深圳市沃特玛电池有限公司 Method for estimating SOH of battery pack
CN108445401A (en) * 2018-02-09 2018-08-24 深圳市鹏诚新能源科技有限公司 On-line Estimation method, electronic device and the storage medium of battery charge state SOC
CN108594135A (en) * 2018-06-28 2018-09-28 南京理工大学 A kind of SOC estimation method for the control of lithium battery balance charge/discharge
CN109870651A (en) * 2019-01-22 2019-06-11 重庆邮电大学 A kind of electric automobile power battery system SOC and SOH joint estimation on line method
CN109633473A (en) * 2019-01-23 2019-04-16 刘平 A kind of distributed battery group state-of-charge algorithm for estimating
CN110398691A (en) * 2019-06-26 2019-11-01 重庆大学 Based on the lithium-ion-power cell SoC estimation method for improving adaptive double Unscented kalman filtering devices

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
孙幸勉: "基于自适应平方根无迹卡尔曼滤波算法的锂离子电池SOC和SOH估计", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113805075A (en) * 2021-09-15 2021-12-17 上海电机学院 BCRLS-UKF-based lithium battery state of charge estimation method

Also Published As

Publication number Publication date
CN110850299B (en) 2021-01-15

Similar Documents

Publication Publication Date Title
Xu et al. State of charge estimation for lithium-ion batteries based on adaptive dual Kalman filter
CN106918789B (en) A kind of SOC-SOH combines online real-time estimation and on-line amending method
CN111060820B (en) Lithium battery SOC and SOP estimation method based on second-order RC model
CN110441694B (en) Lithium battery state-of-charge estimation method based on multiple fading factors Kalman filtering
CN110596593A (en) Lithium ion battery SOC estimation method based on intelligent adaptive extended Kalman filtering
CN113156321B (en) Estimation method of lithium ion battery state of charge (SOC)
CN111581904A (en) Lithium battery SOC and SOH collaborative estimation method considering influence of cycle number
CN111220920B (en) Retired lithium ion battery state of charge calculation method based on H-infinity unscented Kalman filtering algorithm
CN111198326B (en) Battery monomer short-circuit resistance online detection method with anti-disturbance characteristic
CN112630659A (en) Lithium battery SOC estimation method based on improved BP-EKF algorithm
CN111142025A (en) Battery SOC estimation method and device, storage medium and electric vehicle
CN112557906A (en) SOC and capacity online joint estimation method in full life cycle of power battery
CN112269133B (en) SOC estimation method based on pre-charging circuit model parameter identification
CN113777510A (en) Lithium battery state of charge estimation method and device
Singirikonda et al. Battery modelling and state of charge estimation methods for Energy Management in Electric Vehicle-A review
Takyi-Aninakwa et al. Enhanced multi-state estimation methods for lithium-ion batteries considering temperature uncertainties
CN112946481A (en) Based on federation H∞Filtering sliding-mode observer lithium ion battery SOC estimation method and battery management system
CN114545262A (en) Lithium ion battery parameter identification and SOC estimation method aiming at loss data
CN110554320A (en) SOC estimation method of lithium ion battery
CN113761726A (en) Lithium battery parameter identification method and system
CN114814591A (en) Lithium battery SOE estimation method, device and system
CN110850299B (en) SR-UKF-based lithium ion power battery state estimation method
CN114114019A (en) Lithium ion battery SOC estimation method
CN116819340A (en) Online parameter identification and state estimation method and system for energy storage lithium battery
CN110837049B (en) Lithium ion power battery state estimation method based on UKF algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant